{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Project: Multivariate Analysis - Higgs Challenge\n", "\n", "This sheet is not a regular exercise, but rather a little\n", "\"Project\" to enhance your skills acquired during the exercises.\n", "The project is desiccated specifically to the use of multivariate\n", "analysis techniques. \n", "\n", "A (rather complicated) data set is provided - it is a \n", "subset of data released by the ATLAS experiment at the\n", "Large Hadron Collider LHC at CERN under the title \"Learning\n", "to discover: the Higgs boson machine learning challenge\".\n", "The related document is provided on our web page; the full\n", "information is given at\n", "[http://opendata.cern.ch/collection/ATLAS-Higgs-Challenge-2014](http://opendata.cern.ch/collection/ATLAS-Higgs-Challenge-2014),\n", "a web page hosted by CERN making available to the public\n", "LHC data for educational projects and research. \n", "\n", "The MVA project will be centred around methods to search for \n", "the very rare signal of Higgs-production in the data of an LHC \n", "experiment. The released \"data\" in fact consist of simulated data of \n", "a complex signal, decays of a Higgs boson to two $\\tau$ leptons,\n", "and the sum of all background reactions. Among the many combinations\n", "of possible decay modes of a pair of $\\tau$ leptons the \n", "\"Higgs challenge\" concentrates on the dominant one: one $\\tau$ \n", "decays to hadrons and a neutrino ($\\tau_h$), and the other one to an electron \n", "or muon and two neutrinos ($\\tau_{\\ell}$).\n", "The short document {\\em atlas-higgs-challenge-2014.pdf} describes\n", "the problem and gives background-information on the input data.\n", "Roughly speaking, these fall into two categories: primary or\n", "PRImitive variables (indicated by the prefix PRI) are raw quantities \n", "of the identified objects in the event, and DERived variables\n", "(prefix DER) are quantities derived from these by the ATLAS\n", "physicists to be used in their own multivariate analysis, the final\n", "results of which were published in spring 2015. The challenge is, of\n", "course: try to beat them!\n", "\n", "For the purpose of this data-analysis project, no deep understanding\n", "of the underlying physics is required, simply take the data\n", "as an example of a complex \"multivariate data-set\". \n", "\n", "The original data set consists of 250'000 simulated events\n", "as the \"training sample\" for multivariate algorithms classified\n", "as signal (i.e. a true H$\\to\\tau\\tau \\to \\tau_h \\tau_\\ell$ event)\n", "and background reactions. Weights are provided with each event \n", "such that the sum of weights corresponds to the same total number \n", "of events observed by the ATLAS experiment in the year 2012.\n", "An independent \"validation sample\" with 450'000 events to\n", "measure the performance of any proposed algorithm is also \n", "privided. To avoid \"cheating\", originally there was no classification\n", "of these events, this was added later and now allows everybody\n", "to compare the performance of any newly developed multivariate \n", "classifier.\n", "\n", "To set the scale: a record number of 1'785 teams participated \n", "in the Higgs challenge conducted by Kaggle \n", "(see [https://www.kaggle.com](https://www.kaggle.com)).\n", "The winner achieved a score of 3.81; there were only tiny differences\n", "in the top league, rank four still scored 3.72. The real competitiion \n", "is over now, but we will use the data for the final data classifcation\n", "project of the course \"Moderne Methoden der Datenanalye\", and it\n", "is your chance to combine what you've learned in the course \n", "with your own creativity ...\n", "\n", "\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "This Jupyter Notebook shows some example code of how to train and evaluate a classifier for the Higgs Challenge.\n", "It will work you through the step of extracting the data from the ROOT trees, preparing the data, creating a model, training the model and to finally analyse and summarise the results.\n" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Welcome to JupyROOT 6.18/04\n" ] } ], "source": [ "import numpy as np\n", "\n", "from sklearn.utils import shuffle\n", "from sklearn.utils import shuffle\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.preprocessing import StandardScaler\n", "\n", "import tensorflow\n", "\n", "import ROOT" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We are going to use ROOT [RDataframes](https://root.cern/doc/master/classROOT_1_1RDataFrame.html) as they provide an easy to use interface for transforming ROOT trees into numpy arrays and also allow to easily reconstruct additional variables and to make histograms." ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "RDF = ROOT.ROOT.RDataFrame" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "From the `atlas-higgs-challenge-2014-v2_part.root`we are going to use the 'signal' and 'background' trees for the training of our models and the 'validation' tree for evaluation.\n", "Let's define a RDataFrame for each of those three trees:" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "signal_tree_name = 'signal'\n", "background_tree_name = 'background'\n", "test_tree_name = 'validation'\n", "file_name = 'atlas-higgs-challenge-2014-v2_part.root'\n", "\n", "rdf_signal = RDF(signal_tree_name, file_name)\n", "rdf_bkg = RDF(background_tree_name, file_name)\n", "rdf_test = RDF(test_tree_name, file_name)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "RDataFrames allow to [define](https://root.cern.ch/doc/master/df012__DefinesAndFiltersAsStrings_8py.html) additional variables in python using C++ syntax, which will be compiled at runtime.\n", "For example, if one wanted to reconstruct the transverse mass of the lepton and the leading jet in the event, one could use the following code:" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "reconstruct_transverse_lepton_jet_mass = '''\n", "\n", "float lep_px = PRI_lep_pt * TMath::Cos(PRI_lep_phi);\n", "float lep_py = PRI_lep_pt * TMath::Sin(PRI_lep_phi);\n", "float jet_px = PRI_jet_leading_pt * TMath::Cos(PRI_jet_leading_phi);\n", "float jet_py = PRI_jet_leading_pt * TMath::Sin(PRI_jet_leading_phi);\n", "\n", "//calculate angle between jet and lepton\n", "float cos_theta = (lep_px*jet_px + lep_py*jet_py) / PRI_lep_pt / PRI_jet_leading_pt;\n", "\n", "return PRI_lep_pt * PRI_jet_leading_pt * (1 - cos_theta);\n", "'''" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The variable is then inserted in the RDataFrame using the `Define` method:" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "rdf = rdf_signal.Define('transverse_lepton_jet_mass', reconstruct_transverse_lepton_jet_mass)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "For illustration, let's also create a histogram and fill the newly defined variable with it." ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "model = ROOT.RDF.TH1DModel('h', 'h', 40, 0, 100)\n", "\n", "h1 = rdf.Histo1D(model, 'transverse_lepton_jet_mass')\n", "h2 = rdf.Histo1D(model, 'PRI_jet_leading_pt')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "RDataFrame will create histograms lazily. This means calling the `Histo1D` method will return a ResultPointer object. However, it won't fill the histogram directly. Only once the pointer is accessed via for example the `Draw` method, the histogram will be filled.\n", "This has the advantage, that one can define multiple histograms and all of them are filled simultaneously once the first is accessed.\n", "If you run the code below, it will take a short amount of time because internally ROOT will loop over the tree and fill the histograms." ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [], "source": [ "canvas = ROOT.TCanvas()\n", "h1.Draw('hist')\n", "canvas.SaveAs('hist1.pdf')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The code below will now execute almost instantly as the histogram 'h2' is already filled." ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [], "source": [ "canvas = ROOT.TCanvas()\n", "h2.Draw('hist')\n", "canvas.SaveAs('hist2.pdf')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Let's define this variable for all our RDataFrames, so we can use it in the training later." ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [], "source": [ "rdf_signal = rdf_signal.Define('transverse_lepton_jet_mass', reconstruct_transverse_lepton_jet_mass)\n", "rdf_bkg = rdf_bkg.Define('transverse_lepton_jet_mass', reconstruct_transverse_lepton_jet_mass)\n", "rdf_test = rdf_test.Define('transverse_lepton_jet_mass', reconstruct_transverse_lepton_jet_mass)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "For the training of our models we first need some input variables. We can get a list of all available variables in our RDataFrame with the method [GetColumnNames](https://root.cern/doc/master/classROOT_1_1RDF_1_1RInterface.html#a951fe60b74d3a9fda37df59fd1dac186):" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "['transverse_lepton_jet_mass', 'EventId', 'DER_mass_MMC', 'DER_mass_transverse_met_lep', 'DER_mass_vis', 'DER_pt_h', 'DER_deltaeta_jet_jet', 'DER_mass_jet_jet', 'DER_prodeta_jet_jet', 'DER_deltar_tau_lep', 'DER_pt_tot', 'DER_sum_pt', 'DER_pt_ratio_lep_tau', 'DER_met_phi_centrality', 'DER_lep_eta_centrality', 'PRI_tau_pt', 'PRI_tau_eta', 'PRI_tau_phi', 'PRI_lep_pt', 'PRI_lep_eta', 'PRI_lep_phi', 'PRI_met', 'PRI_met_phi', 'PRI_met_sumet', 'PRI_jet_num', 'PRI_jet_leading_pt', 'PRI_jet_leading_eta', 'PRI_jet_leading_phi', 'PRI_jet_subleading_pt', 'PRI_jet_subleading_eta', 'PRI_jet_subleading_phi', 'PRI_jet_all_pt', 'Weight', 'Label', 'KaggleSet', 'KaggleWeight']\n" ] } ], "source": [ "columns = [col for col in rdf_signal.GetColumnNames()]\n", "print(columns)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In this example we will only use the following list of five variables. \n", "Feel free to experiment with more and different variables in your own code." ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [], "source": [ "input_columns = [\"DER_mass_MMC\", \"DER_mass_transverse_met_lep\", \"DER_mass_vis\", \"DER_pt_h\", \"transverse_lepton_jet_mass\"]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Most machine learning libraries in python expect their input in the form of numpy arrays. RDataFrames have the convenient method [AsNumy](https://indico.cern.ch/event/775679/contributions/3244724/attachments/1767054/2869505/RDataFrame.AsNumpy.pdf) which allows to transform them into numpy arrays. It will return a dictionary of the form `{column_name : numpy.array}`.\n", "Execute the code below to see what the method does:" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "{'DER_mass_MMC': numpy.array([138.47 , 148.754, 154.916, ..., 128.498, 151.113, 104.21 ],\n", " dtype=float32), 'DER_mass_transverse_met_lep': numpy.array([51.655, 28.862, 10.418, ..., 18.588, 70.106, 18.268],\n", " dtype=float32), 'DER_mass_vis': numpy.array([ 97.827, 107.782, 94.714, ..., 69.903, 93.991, 58.438],\n", " dtype=float32), 'DER_pt_h': numpy.array([ 27.98 , 106.13 , 29.169, ..., 54.601, 4.145, 80.275],\n", " dtype=float32), 'transverse_lepton_jet_mass': numpy.array([ 6.8237383e+03, 2.0909959e+04, 6.1837223e+01, ...,\n", " 7.7430044e+03, -9.3656383e+04, 2.3110034e+03],\n", " dtype=float32)}\n" ] } ], "source": [ "res = rdf_signal.AsNumpy(columns=input_columns)\n", "print(res)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Before training any classifier, it is helpful to have a look at the variables and how they are distributed for signal and background. In the following we will make a histogram for signal and background for all our input variables and plot them.\n", "\n", "Feel free to implement you own plotting functions using `matplotlib` or similar python libraries to replace the ROOT plots below." ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "hists = list()\n", "for col in input_columns:\n", " # we just pick the min and max of the variables as the binning range.\n", " # and use an equal-distant binning.\n", " # This might not make sense for all variables. Then you might need to adjust the binning.\n", " try:\n", " xmin = min(res[col])\n", " xmax = max(res[col])\n", " except TypeError:\n", " continue\n", " model = ROOT.RDF.TH1DModel(col, col, 100, xmin, xmax)\n", " h_s = rdf_signal.Histo1D(model, col, 'Weight')\n", " h_b = rdf_bkg.Histo1D(model, col, 'Weight')\n", " hists.append((h_s, h_b))\n", "\n", "# define some dummy list to fill in the canvases and legends, \n", "# because ROOT will otherwise destroy the objects \n", "# after the loop and won't draw anything.\n", "canvases = list()\n", "legends = list()\n", "for idx, hist in enumerate(hists):\n", " canvas = ROOT.TCanvas(f'1D_{idx}')\n", " hist[0].SetLineColor(ROOT.kRed)\n", " \n", " # we are only interested to see how the shapes differ between signal and background\n", " # therefore we normalize the distributions to 1.\n", " hist[0].Scale(1/hist[0].Integral())\n", " hist[1].Scale(1/hist[1].Integral())\n", " \n", " hist[0].Draw('hist')\n", " hist[1].Draw('histsame')\n", " legend = ROOT.TLegend()\n", " legend.AddEntry(hist[0].GetPtr(), 'signal', 'l')\n", " legend.AddEntry(hist[1].GetPtr(), 'background', 'l')\n", " legend.Draw('same')\n", " legends.append(legend)\n", " \n", " canvas.Modified()\n", " canvas.Update()\n", " canvas.Draw()\n", " canvases.append(canvas)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "It is also useful to look at 2D distributions to see how the variables are correlated with each other. Note, to make the plots look nicer you should adjust the binning and the bin ranges for each distribution." ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAArgAAAHYCAIAAAApvgy/AAAABmJLR0QAAAAAAAD5Q7t/AAAgAElEQVR4nO3dQawtyXnQ8Wo74xBkbIfIMSxQNkkmGyRLRKzIre4ZbIkNEnjxBhArEGIRKSLLaGa6+9nyEhsJVihrz2MBEmwAo9fVb2BnCyMilDexZJQFQoxQxoNAjoxpFp/vp3rVXX27z60+Xd3n/1s83dvnnLrd59x36+uqr74qhmEwAAAAUz6x9wkAAIB8ESgAAIAoAgUAABBFoAAAAKIIFAAAQBSBAgAAiCJQAAAAUQQKAAAgikABAABEESgAAIAoAgUAABBFoAAAAKIIFAAAQBSBAgAAiCJQAAAAUQQKAAAgikABAABEESgAAIAoAgUAABBFoAAAAKIIFAAAQBSBAgAAiCJQAAAAUQQKAAAgikABAABEESgAAIAoAgUAABBFoAAAAKIIFAAAQNTP7H0CAICzKYpi71PAw4ZhWPI0AgUAQHoLOyHsZXkwx9QDAACIIlAAAABRBAoAACCKQAEAAOOc27T958+fb9r+dggUAADX0zTN3qcwbeGJXXz+fd9PHi/L8rIGr4ZAAQBwJc65tm39b/UL/+vxEf3629/+9mSzk9/6Lxx/O3OS4wb1SNu286cda9ZaO9lg3/eTrwoanL+Q8duVcICEQAEAcCXa+ZVlWZalfFsUhXOuaRq5Wa+qSh6qqsoYI8edc7Kc78tf/rI09c4777zzzjvSlDwqvWNRFNKONiu37PI0bSemLEv/x+l56mnoVQTt62nH2n/69Km8Vl5YVdV8d+436LccvCHjBuUL+XfJh/KwAQCApGY6F2ut/FvXtX+86zp5SF9b13XXdXVd+898++233377bX2avERfro3IE7quk4f8Z8rTZk7M/+n+k/2f6J+VuS8asbB9fTS43oAeD77wX951nX/JcmL+tzOWBwAUXAIA7EDn5ucn6eUGuigKa61z7qtf/WpRFHd3d3d3d8aYvu9j7XRd1zRN3/d1XQfPjHHOzTQYO39/TuFBfd9rlsNl2QkyjGHukx5kzENGF+SS5e2Sb5MkQDD1AADYjQzda+c3+YSmaYZhkEkBY8zd3d3Xvva1t99+2xhjrZVx+PEwuxwchqFt27IsJc7QWY/JnyXHg6fFBvD9JIPl1ztzwgtJkoRzTgIUOVUZaZCW5du6rlPljTKiAAC4niB3ryxLyUWQu/lx9ym3y9bavu+7rjP3gcKXvvQlY0zTNFVV+Y8qzTqUDnUmPvDVdR08Tdrx29e7djkuIxYLTZ6w5mcsYa3VM5QXBg3G3pDLLZyiAABgoeSdy+Ss/4OPjo/PtzN+2mTCgR5Z2Jo2tfY0YsYnEDS4pP3ln1ExsG8HACCposi6cxmP/MfGG2S+Y+1M/7j9u7u7N998c+Y9GY8obF1wYvlnlPVnCQA4oswDBZg1nxHJjAAAIOpsyYzLN9gGMOlXf+O9vU8Bcz54/629TwG35ZJAQSZsZOZGinGmWqyZBONdwGO8fvds71PAnEP8ieOe7UxWBwplWWo9irZtZVlIVVWH+N0FAFwHscJprM43KYpCxg9kOEFergc3Occ1Zn41CWWAJRhRyNzLF0/2PoWHkcyYv+Wf0SVTD1qsalXdyqvhtxMAgFQuWfUg5Ta1YLWsFs1hOAEAAKS1OlDouq7ve6kQqTuBripgCQAAjuKSZEZ/bD/4FgCAGEmHDw7Kdk0XtCZ3qgfqg+Tu2v92/Bx9gu7wFJRodPf8kpHjXbUuKCgZc7Y6CgCAbPV9L9snrnrVTL58nqlyk2TP65kn+I9KRGWtbdtW1w2Y+9hIrlqG8zUHQJ6/xZkvDRR0L/DYsoIDxXQAgB2lutMNbtCzJWMD4yghOHmJh8x9SKG9alEUur1kEBxUVaVR18UDMw9amqMwDIOcwcw+VAAAXKa5VxRFURQaTMjdqfSIMhQvt6zyrX/vKhtSC+0y5TlBm1cmJQrnb/fLstQdroNFhXVdS21DuSiNDPyp/00DptXJjPLZbHAmyRQRe58XAMAYb5bd32XROSfdoeySrAvr5Ca76zr5VrLp9a5ayVi9bK9sra2qSo7Lk4M2r0zyDGZ+tL+Q0BijVQ3HT5MvpDW/L5bhinGclMTqQEHGOiSaS3geCTHmAQDZkp7e5/cm8rXcf8fW3mvQoKSjHYZB78iNd1srB8uyzKQw4Nj84sHgnCUOkL7Yf0jipK7rJE5KGCusTmbUGFAyLCQtJc+3HgCQm5mp9IW5eOMeJxiT14M6ry/r+fO8v508+ZlnargjsYJMx/g3w/7UTJIzvLDgUtM0MpJjjJEBhiRnAwDA4+nchHZVbdvmORbeNM18LSLt73VoxP92Mhqw1s6vsFjlkkBBJRzZmPzwZKxi/JAuLc08WwIAcAXSZTYeyR809/2Fc24YBs0KzMo4IyEYdNHcxsnBe7m68UMJl0peEijIZ1AUhe4x/cgMAJnIGC8Uadu2LEsJA/0kDslS0VGXx/xoAMCVuZFHNhjcakvvYEbdRIb9heZkBAf7vtez1UhC46Hgtf6uzmaUGvl4q3MUZJbBWpskK0TmkMYjJNKyxB8Sl2j5rbZt9Uf7xwEA+ZNkxuDgg3/GJd1vptPpus6fB5fBfEmK9H+cTEPkw01tryhLJfW0/RJVcpk6LqKXM778hLmDq3cCnRziUPMrQCZbk6DJ7/6NMUVR+KtfdEtrf29rM1W/k71NgUdim+nM3eY20/Ndj/80E892PFze/drLWXWZyz+jxJ/lxXWyZip0mvsFssHSl8kXEigAufnCV76bvM3Pffj95G1u4RCd+hb4U5y/5Z/RAfZ60DIaC59/wRIMfqEBAJj0qFUPW5NanhIlLB8yihVcmrHlRQAAcGD5Bgqy3kFKb/rLRv3MxwxTWAEAOJNMpx6CvbNiDpqfAgDAUWQ6oiDFoccLbWUdhK6GkKftdpYAAJxdpoGCmdo4RI53XSclmCRpkdkHADgKKdY3/rstGWkZ1lfejm5FMTa57jFWqti35DkXyCVQ8BMRTCQhUR6S3S9kjyzyEAHgQCTJbNyTZVhZeWuxXQikrmJwROtLSsHiyVdpuWS5l054qo8NFIIKS1eLB7WONwDgWMYd4U4nsg8ZVhmXJJaiUuOylVKVUqbgh2Hwqzsr3QFLnmOSdserAwUJW+QsJX7xA5wcOu8iYu/zAgD8tLiy39U1TRNkm0lHI/xuRWYoguPaK+lD4340GJCffM7VSIXm2EPjt8K82rf6FZ1922XsrQ4UtOC2nL1MCkwGOHuhWAIA5Czo6sY7GElHI7PMut2RbB8ok85yPBjPloeCzR30Ud16UDrjHW9rZSR+couH8TYIk+c5Ho3w95Ga3GjqMS5ZHiknMbmVBQAA83QTYDN1xywdufZ2st2RPKR7HclzgmEJ3VxxHCjoNkt1XS9Ze58bTWgYZzCIYPurtJtCXRIoyDyKVENKdR4AgBuhIwRyDx3ccwYJj9pBSmqaLpjv+95/4YP9ogwqyJhEuku5hmBnSGvt5JbLWsVYd9lOlaaweupBYhY5Y4lx5OscshMAAIegYwbjeQfz6nS7DAZI/1cUhXaBtzOkPV7rN752f68DeceWLCR57733lpzA6hEFWc3pLwANxn8AAJgnEwST1XWlFxyvp6uq6jHdjTQiExDHmnqQxAV9l/q+TzWc/81vfvOtt9568GmXLI+UQEEjF6IEAMAq0oNI3x885OflmUgvE5uqj9FEyOSJflegHa4ZlSf23xw/tEpbtviSQEHPTBaD3lo5LQDA48lt8bj70Lw86V80aJCVDnJQIozJaYtJul5P2znQLa5ERXLhfo6FHy3J8kNdIJp2pL9YOwLjZ0wURdF1nURqmYzkFMXqKwKwqS985bvJ2/zch99P3uYWXr54svcp7CPJn+LJiQn/oH+rfXqT78bkcx58miiK4lvf+taSqYfVn6V+/JJBKl9LxJDDBzZTWIkAAlji9btnaRv88Ne/mLZBY8znv/O95G0exSGCD+7Z8rf8M7p8m+ls6yjw2wkAQCqXBAqymFUmIMwGRaAAAEAmVgcKUvlBKl1IViOVlwAAOKvVgYJUfoh9CwAAzuSx20yLfXfiAgAAG7lwm2l2cAYArCXr6sc3lrJ/NCV58nTJNtPWWkljrOu66zprbdqNqgAApxRs+KSWbExwJlKSORYYPVgsQV7ux1vllFRne+Hukea+HLdu1JFPJBgb4SCXAgByEFRfvrWZa+mkpHjiuArRfHVq2RlS+l8pNznZ+a6qb/2gy3MUgnAmn096iNj7vAAAP63c7HcZ460g/TluvxOVGYrguDzZf8m4PwoOztzNb01+7jAMslm27qJp7gtNyvaYMbozlnPO3yLSvcrcv89JXLjXg7mv5ZzqPAAAN8LvHc3UTtNyryx7K/d9rwWbZacDPR7sMCkPyT4RMz9RbtnzGQX3e1LZ5yL2TIkA9Mzli8mEj+AdfqTVgYJ+PPLJaTIjOQoAgCVkf0j5eryFgXSWWspPOh15SPPh5DnBsIQ8NNlB+ne2+45/+ycpIYve+ss4x4Md/OTmF762bdNe46PqKMjgiSFKAAAs5m/pNJ53CBIepZeRJ8uUt5C6f0GbMbI5kTSyb5HAsixlykBmDVbd+i/p/ufHJC5zYY6CPxFi9g7QAADHoh3k5CxAEAHIQIJkIci0wji8WPITtdvacd5BQhZ/AiXhnbbET8l75NUjCnKR4+NkCwIAFmqapqqqyTFpiQCC/APj5fFd9hNlLGH37QxlPEMnUGRXhIWvlUuYecIF8dMSq0cU5CJZUwAAuJj0lLrSzycZDBoQBCvsxPwawkkzAxhHoVM244Mi7fiEumTq4bjvMgAgE5IoMO5QZJa9qipJltegQQbq5aBEGGt7fS3/k+YCLmKt9UcFlpy/VFjSl+vXOiyhT1vY4FrF2sEASSfJNlaYqSfNsAewxOt3z9I2+OGvfzFtg8aYz3/ne8nbPIqXL57sfQoPK4rVncvY5MSEf1CSE5c3OHlHfn1BcYHgjZKSSsHmi33f6xG/m/OLNUliwPK3fflntPqzlGswryabmAzeepHktxO4ZQQKmbudQCEt6byCMog7esyawSTrDZd/RquTGSejBACI+cmf+1H6Rr+TvsktHKJTvwVyU26tzSRKMI/r5q98FZfs9bBFnBirMhEbepLaGiL5yQAAzkQWRtJfXObyvR4SksKck5MXun5GyaIacz+OlMmUBwAgZ0QJF7ukhPPkYpXLzOyBUZblZGailKqQ8HCypjcAAEhldaBQVVXf97pwRV18BlJ1a+Fxrf7tfwsAADayOlCY38R57UjDzB4Yk8eDolqZrHUBAOCsEuco5HCLX6y39ykDAJCpLJIZ04qNeczY+5QB4FY0TePfpy28vdRdpGeaKoriKBmLcjlN00yOiM9che6iObnuT5tNe7YHCxSCglZMOgDAgUj1QN0wSDZc9msSx/rImb/23b26rqXGc/rzTqooCqniLGv3gn59fhsL2QhjfFy21pR3qW3btG/CwQKFQJLqVACA65AoQbtG2e1wfkfEB+m9ddM0MkKcc6cg1z4Mg6zdk1BJHppZBuir69p55KBuremckzch4bjCwQIFuXK9/rZtqREJAAcSjA0457quM/cDxv6QgCySXzuhIOMK2vh4ViJYUlfXdXD/La/Sb9MO5k+m5Pvfzndq84MN+vVj9uMeO1igYIzpuk7GVeSDZPYBAI5CevGg69X1a9Zaa62OCvR933WdfLt8U2m/65X77GEYZOdJecha+/TpU33O06dPg7659HZllKH+tIGC3235X88sAwxaWBI/rd2Ge0YugcIwDJPXPAxD8K6VZSmfuv4CAQAOoWkaGT/Q+71YvyhRwsVr4HVrZp2hlljBTO1V9O677wZHhmGQ+EBK/K396QtpxsbaF0oPaF7dSVLnLOazHC5wyV4POch5CgoAEKO1faUvb9t2vDnyOP9s7SyzDlqMUwXfeOMNY4wkB0gnLUcC8oSN9pHSfZjX7mbpv1EytCCDEMMw+Kv9rbX5jijQfwMAYoIZB01QSDiJPNOU3ru/++67Mvvw9OnT8XDC1mTHIln68fhOU683GGtPmMC3LlAIpk9krkVWsuoTUp3ZxaiqBAB5Gu//N9lTjmcclt8fa5K7JAY2Hm1ZFho8f/5cv55sRzIq0mbC6XTGBd2lzqf4NJdCSywYYzQhI4mlUw86TmLut2XKdjcmEhcAIFtVVfnj7X7mYEDrEc131fqoc07rE0ibEpfojwjSEt98883YbXdRFBJkyAkn7FZ01UOQxjj/Eg0R5AL9WRVNxdAZnMk5l8dYGijIale5sOZ+o2e6ZADAcjLYHtxnalciXbtOScjug/LQzEC635q/LFCWQfqP+mmJMvswOe/gZ0HKAECZbs9kGRqJvQOTdOWFXJFkdchDekW6FCI4nkSxsLMvisKvklEUxdoUjOsoiqVXBGDS63fP0jb43//ar6Vt0BjzZ/7F7ydvcwsvXzzZ+xT28eCf4li5vOB4kqp65yvNt/Ddm7e8u1wRKPiRAYECcFbJA4Wf/Z0fp23QGPPHX38teZtbIFBAtpZ/RrnUUQAAABkiUAAAAFErCi4FKZTjKhYJTgcAAORkRaAQLGNNWyESAABkaGmgcKC0lFhtpQNdAgAAmViaozA/s5BDQUY1ROx9XgAAI8V81cLuQ6oIzDe1dkNqLLQ0UKiqyq9+5d+1azEsAABm6H6Jcv8m5YM0VtAKxGMzN6vdPam4fJSC/eN6zFqDWc28cFwDKnhtwsRBVj0AAK5EogTtIKXm7yNvNbVrlE0UzUFqK2mBav/Ig8l/UhlZXlhVlV+ZarvEQQIFAMD1BL2jbiApezH4QwJlWV4woSDjCtr4eFZCCiH7zw8GIXSLajE56/EYcl3jfr3v+7qunWf8Wom05NGu64JG/NcmPGcCBQDAlUgvHvTEulektdZaq6MCfd/rpsnLb5f9DrKqKmlQ+lR5yFore0yLp0+fBhtJyJYT+q3s9bDuOh86Qz9SiZ38zMv9L3TfrIT7SgcIFAAAV9I0jYwftG07n8woUYLfF66iGQC6/YHef4871GBrKP+FujXU2hOYIXteT/brmp4ZixistTL14O+Kae6nHtamiC60oo4CAACPpHl20pfLXojBwrTx5kZrb5d10GKcMPjGG28YYySPUu7s5YhP9piOjf9vQX+QBFKyc+Z4vZ5Mpujmk7pLpMRA8q2sMJhJh1xrxYiCnLeeokYuwXaZuysi9j4vALh1wYyDJigk7I9nmtIBf9lj2hgT22laJj7MfU5AqnObUZblMAy68EFChPG1FEUhkymyZkT7XzmiSZ0y8JDq3JaOKKTd3HpTlEwAgDyN73Qn73p1OF0f7ft+4aBC27byTHntZFpi27ZPnz59/vy5fD1uRIY6JsckrinISfTnRIx3kuMz1EAniaWBwiFWmwAAMldVlSYfmPvOZbKL0QWE8+MNfo0f6fV15qJtW38uv+97v0998803Z4IPWbe5XYZgQKY5Lg5K0hZOCKxLZpR33F9tcrX5GwDA0Q3DIKPi2o/0fa/DwNKXS78uuYc6wT3TYVf3pF/3W5PBef1B/tC4zDhMzjvoy811hxMkrJGv/cEMDSCCEQ4/k0PfN3MfMCWcMZnIlYjRoQxrrYQI+m0+4cJk9geA5V6/e5a2wZ/9nR+nbdAY88dffy15m1t4+eLJ3qewjwf/FI/TFSePx562ymWNyMD+dh3KeEGH/ET9VsddpPPVrAU/+UCrVwXHl/TLy7vLpc+TC/DHi4ScnF9pa18ECsAjESgkRKBwXJI2eP3b4CVhzcIwa176QKEoilg0sHXYtcoJfjuBfX3hK99N2+Bf+vv/IW2Dxpjf++oXkre5BQKFI9K78+NewhLLP6MVOQqxIIU8RwDAaUh1pnNHCatQcAkAkB7Va07jhIFC7LeT8BAAruOW/97qkP5/+vbf3PQHffHL37rO+7wiUMgkXfFBt/wLCgBAWksDBWvtdntdAwCAPC1NZnTODbMeeR6x9RTBbqR6MnI8n/oNAACcUhbbTEsZqaDXL4pCSk/IbqR6vGkaWbgiK1iIFQAA2M7OgYKMDYw3uZJRBNlKS4YrdFxB6j5JWey0G2QBAIDA/iMKUo47OBhsxaHlsSRc0MoNR8mvBADgoHYOFCTVYLK/D/YhlVRKGUUInsPsAwAAG8m3jsLFBR8vqPLBikoAACblGyhcjF4fAIBU9s9RiJmcUNA5iJnnAACAVDINFILNPYPUBP+4YVcqAAA2k2mg0DSNjhw45/q+l4RH/18zWhwBAADSyjRHoSxLa62mJdZ1rcMGXddVVSW1mAyzDwAAbCmXQGGcgSgllcxoZqEsSynENH4IAACklUugMGkmDiBEAADgCrIOFABc349/6dNpG3zjZ/5r2gaNMb9nvpC8TSCtP/xE4v9KezlhoBAruER9BQAA1jphoEBAAABAKpkujwQAADkgUAAAAFEECgAAIIpAAQAARBEoAACAKAIFAAAQdcLlkdRRAAAglRMGCgQEAACkwtQDAACIIlAAAABRJ5x6AADg1rh7ZVk2TTP5hOC4c25Jy4woAABwbE3TVFUlHX/btpNJ/c65vu8vaJwRBQAAjq1t27qudcCgKIqmacbjCtbahaMIPkYUAAA4PD8smAwIZFbigpYJFAAAOLagLkDf9+OYoO9751xxb/nQwgmnHii4BAA4lr/6l/9pknacc1VVmVcHGFTf913Xmfuchq7rlowxnDBQICAAHuP//sX/kbbBHwyfTdsgcAh/+IlPL3/yP37+D9a2/5tvfCM4UpZl3/exRAS/c5ShhaZplowrMPUAAMCxScdvjOm6buGcgrV2YeMECgAAHFtVVTKQEJtKuDiT0RAoAABwaDKEUJale5Xx4gOZldDEBampMJnHMHbCHAUAAG6H1llq21YPygCDX2Sp67qqqvQ5dV0vHGMoTpb6VxRnuyLgyj7z3vtpG/x7v/g8bYPGmH/V/FryNrfw8sWTvU8B16bd0D/pfnvTH/Sbb3zjgv5Ohx+Wv4QRBQAAbsUFmQonDBSoowAAQConDBQICAAASIVVDwAAIIpAAQAARGUdKMgCUFkbGjzUNE1ZlgvXgAIAgMvkGyjIlhXydVVVfqJmURSyErRt21jqIgAAeLx8A4W2baVmtXOu6zotGSGjCMMwOOckb5FxBQAANpJpoBBUhJAvJCBo29bfyiK2TRYAAHi8TAOFSRoQ+NMQUr96l/MBAOD0Mg0U/CEEM5pcmC8sVay3zUUAAHB4+RZcCravWI6CSwAApJLpiIIxpizLYRi6ruu6Tvp+HUggKQEAgOvId0ShaRoplqBH5Osge9E55+c2AnikL/3Cf0zb4D/7+C+kbdAY83Pmfydvk50ekdYPhs/ufQpp5Dui0LZtsOpBExc0e1F22mZ5JAAAG8l3REFyFDTTUDMPyrK01urxuq4v2DQTAAAskW+gIDkKQUEFIVWYxscBAEBa+QYKIhYKECIAAHAF+eYoAACA3eU+onCBWAEl6isAALDWCQMFAgIAAFJh6gEAAEQRKAAAgCgCBQAAEEWgAAAAoggUAABAFIECAACIOuHySOooAACQygkDBQICAABSYeoBAABEESgAAIAoAgUAABB1whwFAI/xjdf+TdoG//wPvp62QWPMz5nfT97m63fPkrf58sWT5G3iKH7wk5/f+xTSYEQBAABEESgAAIAoAgUAABB1whwFCi4BAJDKCQMFAgIAwK1x98qybJomYctMPQAAcGxN01RV5ZwzxrRtGxtZvwyBAgAAx9a2bV3XMqIgw+oJBxUIFAAAODw/MrDWyuhCEgQKAAAcW5Cc1/d9WZapGj9hMiMAAMfyz7/8W0nacc5VVWWSTj0QKAAAsLO//m//0dqXjGOLsiz7vk8772BOOfVQROx9XgAAbMI5J91c13VpowRzyhEF6igAAG5KVVXJBxLUCQMFAABuh8QHZVkGgUKqfEYCBQAADkzrLLVtqwcTDjBknaPgnGuaZhwlGWPkeNoqlQAAHE7TNMPITdRR8AtSVlXlD6EURSFxU/JClQAAwJdvoOAXpOy6ru97CRpkFEHCpeSFKgEAgC/rHAUdRfCHE9q2tdbqt9vleQL5e/3uWfI2//O/+/dpG/zkv/wTaRsEDuEPfvKn9z6FNPIdUbDWytSDbJppInGD1JfY4fwAALgB+Y4oSPkIKUVpjOm6Th+aX/JxQdYCpRcAAJiU74hCURTWWsnerOtaExsfNE7+fNDGlwIAwFFlGihITKCRgaQratIiSQkAAFxHpoHCmCYwBtmLzjk/txEAACSUaaAgWQj+EELf9zquoNmL/nEAAJBcvsmMXddVVaUFKeu61rUP1lrNWNTjAAAguXwDhbIstQhlEArImsnxcQAAkFa+gYKIhQKECAAAXEHugcIFYnUUWAYJAMBaJwwUCAgAAEgl01UPAAAgBwQKAAAgikABAABEESgAAIAoAgUAABBFoAAAAKJOuDwSQFaGjz+59ykAO/joR5/b+xTSOGGgQMElAABSOWGgQEAAAEAq5CgAAIAoAgUAABBFoAAAAKIIFAAAQBSBAgAAiCJQAAAAUSdcHkkdBQAAUjlhoEBAAABAKkw9AACAKAIFAAAQRaAAAACiCBQAAEAUgQIAAIgiUAAAAFEECgAAnETTNLGHnHPlqxa2ecJAoYjY+7wAANiQc65tW+dc7NG+7y9oloJLAF5R/M/PpG3wU3/0w7QNAofw0Y8+e7Wf5ZxrmubBOMBaGwsjZpxwRAEAgFtTlmVd1zNPkKmHC1o+4YgCAAA3RXMO2raNPUfGG3Qivuu6hXFDpoGCjKIEB8uy1INN00hwNJO4AQDAIfyvv3F3hZ/S933XdcaYpmmqqloYK2QaKIz5Uy8SEFlr27Zt25akBADAof2pb71Y+5K1sYXfVzrniqKQW+4HX5hpoFCWZXD2RVHIERlC0AuWS2VcAQCA5ay1C595jIElZKwAAA+BSURBVGRGP0ejbVv/8i7L4QQA4HZcnMloDhEoyJIPf8zAv9qyLC9bGAoAwLlpfCB9pfakUlNh4WB8plMPvrZtJflCzYdFF9RWIssBAHA+fpGlruuqqtJlEXVdH3vVg5J4Z9WACb0+AOA2BT2gn8NXluUwDDJZv6pXzX3qoW3bcQUJkhIAALjAql0eRNaBgr/MQQXZi8655ambAABglaynHiZHDqRMhD5By0cAAIDksg4UgpWQoixLa61mLC5PxwAAAGtlHSjE0hKdcxekYwAAgLWyDhRmECIAAHAFRw0UZsTqKLBsEgCAtU4YKBAQAACQStbLIwEAwL5OOKIA3I6PPv/Lydv81//lryRvE7hBP/4/f3LvU0iDEQUAABBFoAAAAKIIFAAAQBSBAgAAiCJQAAAAUSdc9UDBJQAAUjlhoEBAAABAKkw9AACAKAIFAAAQRaAAAACiCBQAAEAUgQIAAIgiUAAAAFEnXB5JHQUAAFI5YaBAQAAAQCpMPQAAgCgCBQAAEEWgAAAAok6YowDcjh//0qeTt/m3PvzbaRv8sx9+P22DG3n54knyNl+/e5a8zS3OE1v41A+37WF/tGnrHkYUAABAFIECAACIIlAAAABRJ8xRoOASAACpnDBQICAAACAVph4AAEBU7oFC0zRlWTZNs/A4AABIKOtAoSgK55wxpm3bsiz9423byvFYRgIAAHi8fAOFsiyttc4559wwDH3fS9AgowjDMMhxPQIAAJLLN1Do+96PAIZhkEGFtm2ttXpcgolrnxwAAPnZ4s4500BB+n7JQhjnIvjTEGVZ9n1/1ZMDACA/zrm2bZPfPGcaKAg/R8HPRfADhclXrbXxdQAAsCHnXFmWVVVt0XjWgUJd15qjYBaPqAzrbXsZAABsrCzLuq63aDnrgkt+ZODnIkjotMcZAQCQnbIsNY0veeOZBgpywX5A0Pe9xEpB9qJzzs9tBADgcD7+7S/ufQpRmQYKxhhrbdM0EhNobqMxpmkanYZxzvV933XdXicJAMDjfeYffm/tS64WW+QbKDjn/EzDuq4lUJD6CuPjAAAguXwDBXNfVcmMljlIhuP4OAAASCvrQMHEQwFCBGAjn3r5C6mb/G+pGwQO4LWPX9v7FNLIenkkAADYV+4jCheIFVCiXgIA4Ny26OlOGCgQEAAAkApTDwAAIIpAAQAARBEoAACAKAIFAAAQRaAAAACiCBQAAEDUCZdHUkcBAIBUThgoEBAAAJAKUw8AACCKQAEAAEQRKAAAgCgCBQAAEEWgAAAAoggUAABAFIECAACIOmEdBQou4XYMH38yeZs/2aDN5F6+eLL3KSxylPPEFv7fDw/wX2mJEwYKBAQAAKTC1AMAAIgiUAAAAFEECgAAIIpAAQAARBEoAACAKAIFAAAQRaAAAACiTlhHgYJLAACkcsJAgYAAAIBUmHoAAABR+Y4oOOeapgmO6NdN0zjnyrIMngMAABLKd0TBOdf3/eRDRVG0bWuMads2lpEAAAAeL+sRBWutP4ogZAhBExGKomiahnEFAAC2kO+IQt/3ZVmOj7dta63VbyeDCQAAkES+gYIxxjlXFEVRFGVZ+tGAH0CUZRmboQAAAI+UdaBgjOm6rus6Y0xVVXpwcqRBFettfRUAABxUvjkKfjkEGVpYmItAHQUAAFLJN1AI+LkIsjByz7MB8vCLH3w3eZsfff6Xk7cJ3KDh40/ufQppZDr1MA4FNBEhyF6UxRFXPDUAAG5IpoGCpCjqRIN8of9q0CC1FlgbCQBA0zRB7r9P7sB9C5vNNFAwxtR1LfWUpLxSXddyVWVZWmvleFVVehwAgNskmXwSIlRVNXn/PFPGcF6ReeqfXPY4FIgdL4rcrwhI6PW7Z8nbTJ6j8LkPv5+2QWPMyxdPkrcJJKHd0M//3fS/+b6PfvdX/NqDOi/fNE3btuOuULY+uKDyUL4jCiI2PLJq2AQAgHObnKz3XbwO4DCrHgAAwFhsiD0g8w5aOqjruoVxwwkDhVgBJaYkAAB5+uh3fyVtg5NTDH3fSw3DpmmqqloYK5wwUCAgAAAcy+f+zh+sfcna2GKyjOGSlIXccxQAAMBaDw4VLC9BRKAAAMCBSUwQjA0EgcJjKhoTKAAAcGzW2mDVg0YPWoLIr0+4qlzhCXMUAAC4KZJz4K9o0ONaZKnruqqq2raVb5eXKzxbeSIKLuGmUHAJyM0uBZfEknWSC9dS+hhRAADgDJZ0/xdkKpwwUKCOAgAAqZwwUCAgAAAgFVY9AACAKAIFAAAQdcKpBwCPscUiBeAGfeqPfrj3KaTBiAIAAIgiUAAAAFEECgAAIIpAAQAARJ0wmZGCSwAApHLCQIGAAACAVJh6AAAAUQQKAAAgikABAABEESgAAIAoAgUAABBFoAAAAKJOuDySOgoAAKRywkCBgAAAgFSYegAAAFEECgAAIOoYgUJZlsGRpmnKsmyaZoezAQDgZhT5z+iXZdn3fdd1Gi5IuqK1tu9782pSQlEc4Ipwm16/e7b3KZzHyxdP9j4FYJp2Q1v/l//g/beu09/lPqLgnJNoQMkowjAMzjl5jxhXAABgI7kHClVV1XXtH2nb1lqr31prnXPXPi0AAG5D1oFCWZZ1XY8HDPyUBZmYuOZZAQBwO/Kto9A0Td/3k6MF49xGX6zg0gzSGgAAmJRpoOCca9v2sv6bXh8AgFQyDRRkusEfOaiqStMRnHPzgwoAACCJfAMFf9Kh73trrUQPQfaic87PbQQAAAkdo+pAURRaR8E5V1WVnLZ8HZRYWHJFJyu3wOVky7+WE9RR+OD9t371N97b+yyMSVRH4Uy/aYbLycb56ihkOqIwoyxLa61mLNZ1zTQEAAAbOUagEARNzjmZfSBEAABgU8cIFMYIEQAAuIKjBgozYnUUDjrdBQDAjk4YKBAQAACQStYlnHOwsM7j8nKQWzxzxx/N5aTywftvHeKZO/7ohc9M/tGsanPHvxjL3eblHOLvwCM1TVOWZfL9jwgUAAA4NudcURQSIlRVlXZTZQIFAACOTYsXO+fqum7bNmHjBAoAAByejiLIFwkHFQgUAAA4sM0LCw3nstXbBADAMlfrvOS1XdcFP9QYY619fJcqzrY88uK3GwCAveTceTH1AADA2SSciSBQAADgwHRr5fHBJAgUAAA4NmttsOqBQGEF51ysWJUcH68hcc7J8eT1rR5PL2d82ke8HDG5jOe4l6M2qpK2tdN8HCf7z3KyP2Vq3J8d+nL24pzr+74oiqIo2raV9MZkUmVF5slaa4yx1uoX+pBcvhz334e6rv2XdF13/dOOCc7NP+0jXo6QX+jgxI57OUIuSs+zruu9z2ip03wcJ/vPcrI/ZWp8boe+nN11XbfFO3PyQMH/ffIXkMjvnP80/VPuv0R+I691sg8Luhz99qCX03Wd/jnwf7kPejk+/095cDnZOtnHcbL/LCf7Uyb0rlfP89CXc2Jnfq8nl5bKr1oQkkuUOox+Tcct7Cs4GT3tg15O13V1XctJBrcUR7wc3/iK8h9UONnHcab/LOf7Uybk/4X/+3boyzmxG3qv/SA0+MOtv4X6e6lyHuPSqzj65cx3q4e7nMk/6/kHCupkH4c4zX+W4Sx/yqy1+on4gcJBL+fczp/MaIwpy7Ioir7v/fyODatdbk82CjNe3tmhL2fsZJdjRiuXjuXQH8eZ/rOc5k9Z0zR930+mzR7xck7vDJUZZb+s8XH9LXT3qqrqui7zX8QHL6csy77vrbWH6HsevBxgO8f6z/KgY/0pi3HOtW07ZFyIEIGbGFEwxuhiG/17MfmHQ/6s6Le5/XHRe6Ou64JzO+LlzDjZ5ZiD3ycd8eM463+WE/wp01X+whhTVZX+Bznc5dyCM4wo6G9boGmaWNwa3GE45zTf2+e23pJrSuxyjLfjeHD8oJcTk/PlLKFV0vxzy/A8Fzrox3HE/ywxB/1TFhMUQpAhH4kejng5N2HnHImNGS81xl8g7qebBQvHzasLcmy6DbgeSc6zruvuVcMxL8dnXs1LOvrlDK/mXh1leaQ6wcdxvv8sZ/pTFjCRlZ8HvZxTOtLfrwvI32jl59P6gap/3M8Syurve3AtQv+3HO5yfGaUwHzoyxH+eR4rPfsEH8f5/rOc6U9ZIPh9O/rlnFIx3EBGSWycamb86ohDW1xObo5ynkuc4OPwHfRy+FN2xMs5gZsIFAAAwGVuZdUDAAC4AIECAACIIlAAAABRBAoAcFukFLTPr5QqtarGJJFwfLwsy0NUQGqaRs42OK7XK9/KmzMuHSsvD94o/508cYrlGQouAQCW82scmfuCTlIcWp8TLEc03loD/7VSj7mqqqPkxfsVHsVklDM+2Latf9w5J0W9pJa2fHuUmGm1nZdnAgCuy4wKFsWKOC157XCQ/VG1FkVwqkFvqIUcgpfLQb32+ffwZJh6AIBbJ6MFm27VJpMX/kSGjtv74xPjyQ59+XiEX6YDxk+eUdd1UCVaDvrPkVjBfzeappFyq/6rgp9YluVkpa8TIFAAABhrbbDxUiD2QulQlwQZTdPIHaq53whqGIa6rmVU39zv0CHPsdZWVSXHi6LQ433fa1jTtq3cwdd1rU+eN95fanI7iSCeaNvWD1Bi70bTNOfMVLj6GAYAYE9mavpANyUZZyf4ncXkQ0vG2/2nSd88fmhcs1mf4G91IU8OGnnwHPQCzavbRnRd52/Iohu1BGfoPxT86NNjRAEAEN4lj3sLfUgy+JQxZvndvN/I+AmysaTcl/ttWmvlnl4e8idKZOYi2Kx1no4WPFgQWgdLgrM957BBHIECAOCnSyEWPrn0SAyRJNu/KIqqqqTX9+f7nXMSkbRtq2kKOnMhKw50feODmqaR2YfYvIPQeKJt24XZGxrEnAyBAgDcOukRN01mXHgOwzDooIL/kKw8HIah6zrt5uWZctysPH9JcZh5icQTk6MOsbQMzZ84GeooAMAt8ksCtG1rrZ3P15vvAlcN/i/hd8NVVdV1LUf0xKQjHy6q36AZlA+es+RXzrSg53mFlSO7uWZCBABgd+OOYJxFGHuOidRReLA38Z+gWYH6kKQiBj9Omw2WHfqpiGp8VgE/YzG4kMlkRj3u52D6P2W8GPKURRSGYWCbaQBALvyh/mDYf3IWQAcY9hrz3/0EroBAAQAARJGjAABIQPIKYw9d4YZbchsnH5KllVufwFkxogAAAKJYHgkAAKIIFAAAQBSBAgAAiCJQAAAAUQQKAAAgikABAABEESgAAIAoAgUAABBFoAAAAKIIFAAAQBSBAgAAiCJQAAAAUf8fqcwDHBuMgfEAAAAASUVORK5CYII=\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAArgAAAHYCAIAAAApvgy/AAAABmJLR0QAAAAAAAD5Q7t/AAAgAElEQVR4nO3dT6z9xnXY8aGdOE3s2rKBoC2KIEHRRtkUCNBdAT8OpdqLAl1L6brLLlx3K4nkz94W9qabAt10E2nRbTdGxaGUXVfeRYqBoDbQojFQWQ4a2E2V28XxPZjfcIaX977h5ZD3+4EgvMfLO/eQ7/3eHM7f6nQ6GQAAgJjPbR0AAAAoF4kCAABIIlEAAABJJAoAACCJRAEAACSRKAAAgCQSBQAAkESiAAAAkkgUAABAEokCAABIIlEAAABJJAoAACCJRAEAACSRKAAAgCQSBQAAkESiAAAAkkgUAABAEokCAABIIlEAAABJJAoAACCJRAEAACSRKAAAgCQSBQAAkESiAAAAkkgUAABAEokCAABIIlEAAABJJAoAACCJRAEAACT92tYBAACOpqqqrUPAZafTaclpJAoAgPwWVkLYyvJkjq4HAACQRKIAAACSSBQAAEASiQIAAMY5t2r577///qrlr4dEAQBwP13XbR1C3MLAbo5/HMfocWvtbQXeDYkCAOBOnHN93/vf6hf+19Mj+vUPfvCDaLHRb/03Tr+dCXJaoB7p+34+7FSxdV1HCxzHMfquoMD5C5nerowNJCQKAIA70crPWmutlW+rqnLOdV0nD+tN08hLTdMYY+S4c06m833zm9+Uot5+++23335bipJXpXasqkrK0WLlkV1O03JSrLX+x2mcGoZeRVC+hp0q/8WLF/JeeWPTNPPVuV+gX3JwQ6YFyhfy/yU/lMtOAABkNVO51HUt/2/b1j8+DIO8pO9t23YYhrZt/TPfeuutt956S0+Tt+jbtRA5YRgGeck/U06bCcz/dP9k/xP9qMx50YiF5eurwfUG9Hjwhf/2YRj8S5bA/G9nLE8AWHAJALAB7Zuf76SXB+iqquq6ds595zvfqarq6enp6enJGDOOY6qcYRi6rhvHsW3b4MwU59xMgan4/T6Fi8Zx1FEOt41OkGYMcx70IG0e0roglyy3S77NMgCCrgcAwGak6V4rv+gJXdedTifpFDDGPD09ffe7333rrbeMMXVdSzv8tJldDp5Op77vrbWSZ2ivR/Sz5HhwWqoB3x9ksPx6ZwJeSAZJOOckQZFQpaVBSpZv27bNNW6UFgUAwP0EY/estTIWQZ7mp9WnPC7XdT2O4zAM5pwofOMb3zDGdF3XNI3/qtJRh1KhzuQHvrZtg9OkHL98fWqX49JisVA0YB2fsURd1xqhvDEoMHVDbrewiwIAgIWyVy7RXv+Lr06Pz5czPS064ECPLCxNi7o2jJRpAEGBS8pf/jOqTuzbAQDIqqqKrlymLf+p9gbp77i2p39a/tPT0+uvvz5zT6YtCmsvOLH8Z1T0zxIAsEeFJwow1/yMGMwIAACSkoMZpdnEnyWiI05LtnyDbQD79ftffzd7mR9/+Gb2MoEDiCcK1loZyel3zPR93/d9rnmZ66G9Czi8V5/ey14mfzoy4pntSCKJgmQJQUKgK1Y2TcM/JwDAPHKFw4gkCtMsQZ1OJ5k5WnijAgBgWzxSFm55JpfsesgWy92lLp7fWgAArhWf9ZBxe8r7m1lcAgAAXCWSKMhK1NGz5fiu2xsAAMBykURBttuaZgNd1/V9f9Wi1gAAKNm4IXDzw6dzbl9DJoMrtTH6qqwIOX1ulw20ZCct/2BQTsaegfish2EYmqaZ/gAy7kYFAHg04zjONFqnVFWVGmJ/1RbP25KH8JkT/Fdl+mFd17IwgfaeO+dkzydjTNM0WinL9lQr3Y3kYMbT6eTO5MdDigAAeL5c/dd5n5vXI7s/T7OEIHjJh8w5pdDkoKoq3V4ySA6aptGqWfbRXiP+uSWcpdFDWjnIEgAA6+nOgv4IaduWGlEeXKXHQb71W779fg2tMuWcZ/ZxPJP0Bcw/7ltrdZFD55x/ctu2fd+bc2Kh1bE80svXqyZMkRYFufsz79lFBgcAKNN050Zzbjxv21basyUz6LpOusKl60HO0adqvxx/qUBrra4N6J+sZd7xWn8VmwTfNE30BLkuvRy5D9HT5AudWKCpj/Q7aNqUdw3lSKIgmYvZVd8PAGAXpKb3j/ij37QK1Ib0aYU3rQWlovUfr6XJQU7T/0vDfoEkm0m9aq3VetkYU1WV1M593/vdDZInGWO6rtPUav5z33333TffvLzFSSRR0FYOc+59uFgKAABLzHSlL3w6ndZ/QZu8HtR+/bqui63OosHPnKkZgDRRSHeMv1aQ3zUzX+D3v//9JYlCfHqkrFAkWYz06+you2E690ZsHRcA4E60leJ0OslztlRnBeYKXdfNrzug9a/fQGK8LpvpW+q6np9hcZW5wYySMUjyIrMld5ExsDIjADwUnZqntP9exz+eTie/vbwc04WLgkYXHdsY7UqQq5u+lHHwwFyioEFoxjDtWwIA4Cpu4pkFBo/aOmxQG+f1+DM/KLvoescyPdIffOBf4HQ8hwzk9C9/HMclbSff+ta3lgQZX0ch4M8BZWVGAMDNog+cFxt9ZbjfzOi8YJ1AqapkUKT/caWNZwxmQgqZKqlh+0tUyWVqu4hezvTyl8x6WDJAwRhTzfx4gvygwK6dqaqauyIAx/Dq03vZy/zogzeyl/mwsv8pjrauR08z6dGOW62jcLNrL+eqy1z+M4qct8f8QK2RKDzsn6SHvfAHl/3nzg/9AfHMVr7lP6ML6yhEk7gCu3kAAMAaIokC6ywBAAARSRRoMAAAAGLRrId9Sa2tRIcZAADXiiQKFxcxLLzGLTw8AAB2ZK5FgSUTAAAZ+Xs8+se7ruv7fnfz7G6mW2b7O0Ca8/4U0+M+3cNCd72aFpv3NkZWZhyGQVKEvu/1I30ZPx4A8Dhk4v20HilwZeX1VFUl16s7X8tx2S1hetwnO0PK1/45shGUDDGUXS0yBhxJFPw1m/e1ywMAoHzBfkUPVbnoVlWydrW//YS0tUyPK+dc3/d6zjAMeo7skCnHpf8941P93F4PqYwh12cDAB6NtFj7yUHXdcG0fHk+Fn6l03Xd9Lg+TOtL08wjaI2PnnMfwZrN/lX4p0Wr2mDhxejWD2JmL+8bXN4Uypxvsfx0M+5cCQB4QP7mBcaY6Q5GsrODbkao+0DK87Qe99/VdZ28FGzuoK9KZ7o5bwCx1UNvsAmWe3kLaRlkIF0Pz1zT6E7bTAsZLiF9KvKTy/XZAIAHJLsjytfT7QmkItetESUnkJd0ryM5J2iWCJ6wfbrNUrAx47Z0CKd8q1s060iF6flm0oSgp2l6pJsw5JKc9RDs+FDIbQUA7J22EEhzdfDoHAx4lNpH5wi4s3Eco234KVIrS5tEvku5kTtvhO1PAJF2FM2QonsxyNgFGZogly9vP51O0u0ip9V1vW6LgrYfSCp3Op32NdOhStg6LgDAr2ibwbTfwby8k4A0BkiKUFWV1K/T9GJHZOaCtNBrlhC0Fkh6NL0zOnBQhjT6L+lxSS8y3p9Ii4KmIZq2BArvfSg8PACAVJbRbZGlhgvGHxhvYP/Nn2iMkQ6IDasJadiYriSx/O3ayeLfPV19QU4bxzHjSkiRRIF1lgAAq5IqLTpkT3MIHY4wffu062GeDoS01kp7+eazHvwApI7v+14XK9IOCHMe/yhH5Ok9ek7f9/7EyIxdAZFEYWHpLL4EALiZdLdP6xEZqOjPXJD6bxgGmaUvB6UbvpusThglTf1yppSjicidSZt9MC9D+iD88QfGG7kpAzDlRknweo42jejsUPk27ziMyFiJhcpcgik6+uOZXn16L2+BxpiPPngje5nZPeyFP7jsP3d+6A8oy5/iaMeEf3Cryn5V0ateeM6S96rlP6OCEoVpgV1i2WqXXg17jUQBGT1y8vGw1/6wF/7I+FNcvuU/o0ULLt2BtTaY26qrYQfLVutK1zK9pMBWDQAADqOIRGG6OoS/GnawbLW/0nV0BS4AAJBLEYmCDDPxj/R97w9n1Skx/iBPk3VUJwAAmNo+UZChntGBr/7X0uQQ3U6D3gcAAFaSXML5PmaW3b55LOsNizAy6AYAgKgtWxT8BSIyOl0vbwAAgCjZIiC63VFVVfQml+mKFoXovlXPEQw4MC+v0BmdICtrV+m3dDoAwI7ohk/BX+/odgHH5q+3GJguFqBbYU3XC9C3RD8iQ6DziYLMP5Q1L7U9X2JNhXWV4HdF1uOUWxAs6B0MTfCPZ4kEAHA3wTS3x3zkSz1sT6cBSl0slWDf91oLz8i7zfRc14OueSkxSSt9xp28JTNS5uUNxf3dynX1yqBVI5gcAQAonMxx8+uR6VaQuiBxVVX+o6D0UATH5WT/LdNKKjiYei6/D3n2ntbl0mAwnfMvdbHkB6la2L3MZN226cIYBbmVqQf69cha3/Ij99foNsYMwyBLMEkjx2OmogCwX9p4LKY7TesuzMMwjOPo73sk2yjL8WCHSXkpur6O/4n+w+cmZK5f6qVpzmRebjgP7t6UJF7rbgrlkwSn7/s7bCkZDCpM9XFYa2UhpulLAIDy6QK7JlYRSmWpg9hkGyR5SR8a5ZygWUIbpKeJgj++bfPHS30CD47LpgTS0eAfnJYw37OQfZbAXKLg52UynkC+vVv1PPNBpAgAsFP+lk7Tfgcd8Cjf+sPbpR4VwTbT85VC13XSu3+3R9+8dEjfdARDYNom8XxzXQ/OubZt67rWDSvrumYyIQDgmbTNINoLEGQAOlpOOqNNbFjDkk/Uhup9zcOUfhbti5+5cEkjsjeZXOh68O9m9u0iV5JacIkUBwAKIR0E0U5kqQiD8Qfm5fnzN5C2hPsPuXu+oMN9ps3ghvxpiQuDGTU5kFGau1gQg1WVAKBwkhxEn4+DFXujz6gXW+CnZhowCqfDNeSm6ejOqZmXnmMuUdAdFuTjh2Fo2/YBl8UAAGQnAwWiG/3ICDl5OtWkYdoCf22tL0nJ7oa4OW/5wWCNAD+LWq9XZa7rYRxHf4tnSWd0PEj2UAAAx+Y37uoKOtOXdOCe8er16ZQ3rYz898ppqQB21+9gvFUl5FsdNRg0q6w3NqCauaFV9atX9UciB2WtxpUCeiaNGWV69em97GV+9MEb2ctcw8Ne+8Ne+CMr8E+x89Ya3jqWW2RfFGD5z+jCYEZpRZB+BxPbnQG4yiP/fd/LtWev19e4cJIPXEUqsrqu91t/bRj5XKIgy1zozZVVsXY3/RQA8OBkYuR+s4RtzSUKQU/PfMcPAADFIku42YXpkVPR/TYAAMAhXViZsZq4W2Q3m8a8l8gBACjNhW2mdf3mtm2HYajr2t/IsUwsuAQAQC6Xd4803iIV0sawu2WtAADAbRaNUQhW0GSMAgDgNl3X+Z3CC588dRfpmaKqqiq8wVvJ5eiekIHUVegWmr5gU4zgSBaX93owL6/lDADAbWS757ZtpUdYtgXwFySeqSNTZQ5nbdvKGs/5486qqirZDEHWgArq9au2sdD1raW9X4rt+z7vTZhbmEmuoW1bydr0eMn9/QUuBwbsCwsu4flSf4qrqpI6RY/Ig+h0FeBA9CVJO4IPqqrqOZtMri2I2f/WOScbYpll9azU0Xrr9Daa2H2eWl5dXtgU6nQ6ySedTidJ2aiGAQA3C6pw55wMmZeqzm8S0F2Lr+pQkHYFLXzaK2Gt9VcObNs2eP6Wd+m3eRvzg32ug0uz6S2kp2RFav12vW0sLo9RcGf67UqhAACOTWrxoOrV1oK6ruu69h+R9el0eWu8X/XK3D150NUtmOu6fvHihZ7z4sWLoIrVPndz3vM6b6KQGvMnt2X5oA3jXay/N3f2zRbmZj1Ik8j0OI0KAIAb6Gi7vu+lfkm1kEuWoDnEtZ3uWh/rXkqyKYGJPXm/8847wZHT6SQDLfu+95/a89IRGze8NwjMnvfmlm/zLmQw16LgDzlhQQIAwPPJNDp5yq/rOjrybrpT4rXt6jpZrzvTvOG1114zxkj1LP+XIwEZaLnSPlI69nAYhhuaK6ZtBn4DzDAM/hDR57vQ9bDHJRNYmREAyhT0OOgAhYyd2jNF6bP7O++8I70PL168mDYnrK3rOpkocDqdbstCpu0QfgOMDMKIdgjcZi5RqOt6j4kCKzMCQJn6vg8q8mhNOZ3jsHyMgjQDmHNrfOfRkqUSff/99/XraDkyoiLvyDztzri5evUHItzHXKIg1yODRX33ig0AcDRN0wQj+EwiXdC6cL6q1hH3OpNfey78at5aG+QEr7/+eqpHQ+ZYdl3Xtq12/Gehsx6c5+JbglWVoqf5xzVbymJuMGNq3AcAADeQxvag6tUWX6nLtUuiaRrtNZ6pifzS/BUUpAXef9Uf/Se9D9F+B6lx9cFdQ7riOtOkaSR1B6KCmRfjOE7HP8roS/92ZWwImVtvYY+LF+0xZqAoLLiE57v4p3g6XDF6PHXaVbIUsgt+knTx5OXV5QEThd//+rtbR3HZw/5J4u97+XaRKDyyXfwj2mP18WjyrMwoQyhZYQkAgId1eYzCdBwHeSIAAA9irkVhfp4hLQ0AABzeXIvCvNRG2pv7+MM3o8d3MXYBAICi3J4oFIuEAACAXC7vHgkAQC6yLJJavlnidMpfUNS1G1JjIRIFAMCd6H6JMuJNtiTQXGFm8d+Znu7hTFZc3sXOPrqLZnBdsgjj/HxD51z0vX7heQcGbJ8ozNwXvZXTt8zfJgBAgSRL0L/qXdfJBpLPKVO3F+i6TobbF96uIPtGGmOcc03T+HdD17dumia1C4ZMRZT3Rlesmm6o8UwbJwpyzdH7orcy2IRUbqU53yZyBQDYkekztKysLFsz+E0C1tobOhSkXUELn/ZKyNLO/vlBI0TwgBrt9biZlHw6nWSXB3+bR9ksSo4PwxDdj0p2idT3TvfKyrszhdg4UfCvWe6LHPdvpWSI+mPzb2Vd12vcFADAGqRuC2pi3Suyruu6rrVVQCoI+Xb57pF+pd40jRQo9Yu8VNe17DEtXrx4EWwkEWwfJXs9XHedabopVBBtsM70dP/MJecEOVAuWyYKM9cc7Hyl+1v4+4Sa++6zCQB4pq7rpP1AmopnBjNKlhCtL5fQHRe1otFn0en+UsHWUP4b/aoni2C7SH93hulihkGCEpwzrRD9vaMyuj1ReH6GFVxzsNmoX76kliaRi9H7AAB7ISPS5ClfBihMhx9Ot3G6dh9jHffWnWnF/9prrxlj5Mlb/i9HfLLHtFm5ftGhncFx6TGp6zpVz0qnjORS+pa+71daNzm5joLco2A7cD+UjGmL5gH+HqA3JyKpBZdmsPQCANyB1NnytdTlOtosV/P+TNWuVbLsMd33fWqnae19iFbkWYKUfnNtNfE/WjaSnqlk5b7JiD2pkZumWSNOEUkU9ALMeUDA2uMA3FnTNNO7di1qfQAok/T3Bw3G09O0tVhfHcdxYaOC9lzLe6PDEiVFeP/99+XraSHyrB88MOeiDQlByVIJ6iiNeZJmSd/N9HbJxIdczSGRrgeZr6KTXCVLmO71kJeObdELi16htj3MnAMAKFYwWy3odPYFTdop+qgp6y8Zr+fCnzgQDFE0xrz++uszyYd0i1zb5XGRZAnDMEzzD2kVmLlYvcCAjGG0Z8aYmW6LG0RaFKTRQ8OSS8r1eT4pPJp56OhFEQxN8I+b4qfMAgDE6XSy55UA/IPyhdTl8qwsjdlaL85U2H5pft0h1af/ql+XSe9DtN9B377G2ECtzoJsSUcmTrMoTYPk/mjLgT+Y0Y9T7mHGyKtpPV1Vld/+H3ybV1VV2vzidz3I1xJb0CXhv0WGe/i3taqqXXQ9fPTBG1uHsI1Xn97LXubD3syVZP8Z8QPKaxf/iKoqUrn4Uo95wfEsT4O3FTLzKPsc0SYByZ+mU0ClspOXdHSg3y4SHcrg15LzkSy8uo0ThZlr9u+af9wfQmG8VFSjJVEo2S7+xj04EoXC7eIf0fJKqFjTB9FyZMmfdpMoiIWp5ZKXSBQyWuPv0RrWuJl7uXZktIt/lSvJ/gv/8Ydv7jdR0MfR/V7CEssThfj0yKDJYjoy87awUmamil77FgAAnsNau/bj8b7EE4Wgp2T52pkAAJhEZzz2KJIoHLuxBQCwtjXqkV0MzjBek/4Pf/Avsxfu+8Nv/vF96uv4gkszTS7+ulplSq3MuIuxCwAAFCWSKPhzEf1piua8hHPhiQIJAQAAuWy8zTQAACgZiQIAAEgiUQAAAEkkCgAAIIlEAQAAJMUXXAq29mLdDAAAHlMkUVhpU2kAALA7kURh7wtcs+ASAAC5xLsejDHOua7rdJeHuq5lV+w7xfUMJAQAAOQSTxSstZIi1HVtrXXOjePYNE2xm3MDAIA1RBIFaUiYbrIpyzmXv9cDAADIJTI9su/7tm2nvQzW2rZt+76/R1wAAKAA8XUUUmMRdjFGAQAA5JIczIhVrbGx+sPiZiILfpGQ148/96WtQ8iDlRkBAEBSvEWB4YoAAMBEE4W6rnX5BAAA8MgiicLeV0pgZUYAAHI54GBGEgIAAHJhMCMAAEgiUQAAAEkkCgAAIIlEAQAAJJEoAACAJBIFAACQRKIAAACSDriOAgsuAQCQywETBRICAAByoesBAAAkkSgAAICk7RMF51zXddba6d7WqePOOTm+9/2rAAAo3MaJQtd1TdNIfd/3fVVV+lJVVX3fT4/LW4wxzjl9LwAAWMPGgxn7vm/bVtsMqqrqzowxp9MpOC5vGYbBWmuMsdY2TaOnAQBK8OrTe1uHsMiqcf74c19ar/B72r7rwe9ZqOtaWxfqup4el5MlSwjeCwAAsts4UQgaA8Zx1CRAv5Cvx3E0xjjn/ARCzqH3AQCAlZSyjoIMODBeI4GfKFwlteDSDJZeAAAgqohEQRoMtH/hmaj1AQDIZeOuB+eczGgYhiHIEqJJg/ZBzJwDAABy2bhFoWmaaENCcDAYmuAfN8/opAAAAPO2bFHQat69zBjTdZ22HDjnxnGUsQv+/81kcgQAAMhryxYFnQkpCysJaUuw1tZ1resstW2rzQbDMDRNo2+h9wEAgPVUJa9WNNOzkHqpqioGMwIAtvLxh29Kxfrvh2+v+kH/+rXv3acGL2LWQ8rM4APGJQAAcAfbr8wIAACKRaIAAACSiu56uE1qZUbGLgAAcK0DJgokBAAA5ELXAwAASCJRAAAASSQKAAAgiUQBAAAkHXAw4w9e/Ke8Bf7ws7+ft0BjzI8/96XsZb7//34vb4H/7f/8g7wFGmN+9ouvZC/zr//qt7KXuYYv/uQ3tw5hkb/59PN5C/zcVz7LW6Ax5rOffCF7mdWX88e5htPPM/+AjDFf+OTT7GXiz0/5/9xtghYFAACQRKIAAMBB6O7KU7Lhom9hmQfsevjd5r9Ej//34Z/fORIAAO7GOdf3fSoJcM6N41jX9bXFHjBRICEAADwU51zXdeM4zp9W17XsvXwVuh4AANg9a23btjMnSNfDDSUfsEUBAICHot0Nfd+nzpH2hqqq5NthGBbmDSQKAABs7N+93t3hU8ZxHIbBGNN1XdM0C3MFEgUAADb2b/9rd+1brs0tTqeTfu2cq6qq67olQxYYowAAwMNZPv2BRAEAgIO7eSSjIVEAAOCoND+w1o7jqMsxyZoKM6sz+Q44RoEFlwAAMOeEQL4ehqFpGp0W0bbt4856ICEAADwmf8SiMabrOm02sNaeTicZvXhVN8QBEwUAABB1w0gFxigAAIAkEgUAAJBEogAAAJJIFAAAQNIBBzP+3X/6J5kLzFucMcaYz//p72Qv81/879/NW+AnX/xK3gKNMeaL+Yv8k7+zxo8ov/d/5/eyl/lnn30te5k/+8UruQtc4Rfp1fxFVv/zy9nL/PxP/lb2Mtfwf7+a+Wf0hU8+zVugMeaVn/4oe5lr+IvzF3/+2Ve3jCMfWhQAAEASiQIAAEg6YNfDb/zGz6PHf/nL/O2KAAAc2wETBRICAAByoesBAAAkkSgAAICkUhKF6GaXXddZa6cvyb6ZXdfJ5hYAAGAlRSQKzrm+74Nav6oq2Q2z7/uqqvR413VN08i7mqYhVwAAYD0bJwrSNiAVv09aEWRDTNk0U9sV+r4fhsE555yr63r6XgAAkMv2LQrW2rZtg4N939d1rd/WdS0tB5Iu6C6Z0Q4LAACQy8aJggw1iNb3/p7Z1tpxHI0x0ooQnEPvAwAAKyl3HQU/UbhKasGlGSy9AABAVLmJws2o9QEAyGX7MQop0Q4F7YOYOQcAAORSaKKgoxdFMDTBP26e0UkBAADmFZoodF2nLQfOuXEcZcCj/38zmRwBAADyKnSMgrW2rmtdZ6ltW202GIahaRpZi8nQ+wAAwJpKSRRkVSWfLKlkJj0L1lpZiGn6EgAAyKuURCFqJg+Yeekf/7P/sEYwef3st//h1iHgCF756Y+yl/nL3/57eQv8zbzFGWPWuXBj/scKZeJx/dlnX9s6hDwKHaMAAABKUHSLwm0+/vDN6PHf//q7d44EAIC9O2CiQEIAAEAudD0AAIAkEgUAAJBEogAAAJJIFAAAQBKJAgAASCJRAAAASSQKAAAgiUQBAAAkHXDBJVZmBAAglwMmCiQEAADkQtcDAABIIlEAAABJJAoAACCJRAEAACQdcDDjLvyv//xPspf56tN7eQv86IM38hZoVghyJY987a/89Ed5C1znZmYO0jz2D32Na89uLzdT/ewXr2wdQh60KAAAgCQSBQAAkHTArgcWXAIAIJcDJgokBAAA5ELXAwAASCJRAAAASSQKAAAgiUQBAAAkkSgAAIAkEgUAAJBEogAAAJIOuI4CCy4BAJDLARMFEgIAAHKh6wEAACSRKAAAgCQSBQAAkLTXRKHrOmutc27rQAAAOLL9JQrOuaqqJEVomqbruo0DAgDguKrT6bR1DNepqqqua0kUuq7r+96/hKra3xUBAA5Dq6Evv/vhqh/0l3/0NK3vuq7L/vy8vxYFY4zeBfmCRgUAAJxzfd9n75TfWaIg12+t3TgOAACK4Zyz1jZNs5oYhRQAABFKSURBVEbhO0sUooLsqbreRoEDAJCHtbZt2zVKPuDKjIxRAAA8FGuttLX3fZ+98CMkCvREAAB27S//6GnrEJJ2lihITiCdMcFBAAB26m//8QfXvuVuucX+xijUdR3MeiBRAABgJTtrUTDnBZd0BOIwDNvGAwDAge0vUTDGnE4n5kkCAHAHu0wUDCkCAAB3sddEAQAABNZYIOCAiUJqASXWVwAA4FoHTBRICAAAyGV/0yMBAMDdkCgAAIAkEgUAAJBEogAAAJJIFAAAQNIBZz0AALC5v/6r39o6hDxoUQAAAEkHbFFgwSUAAHI5YKJAQgAAQC50PQAAgCQSBQAAkESiAAAAkkgUAABAEokCAABIIlEAAABJJAoAACCJRAEAACQdcMElVmYEACCXAyYKJAQAAORC1wMAAEgiUQAAAEkkCgAAIIlEAQAAJJEoAACApAPOegAAYHNf+HTdGvYXq5buoUUBAAAkHbBFgQWXAADI5YCJAgkBAAC50PUAAACSSBQAAEBSKYlC13XRg9ba6UvOOTnunFs/NAAAHlcRiYJzru/7oNavqqrve2NM3/f++MSu65qmkXc1TUOuAADAejZOFKRtQCp+n7QinE4n55wMTtR2hb7vh2Fwzjnn6rqevhcAAOSyfYuCtbZt2+Bg3/d1Xeu3dV1Ly4GkC9ZaOR7tsAAAALlsnCjIUINofa/ZgHw9jqMxRloRgnPofQAAYCXlrqPgJwpXSS24NIOlFwAAiFo9UZDBBNPj6/UaUOsDAJDL9mMUUqLphfZBzJwDAAByWb1FwVp7QyeCjl4UwdAE/7h5RicFAACYV2iLQtd12nLgnBvHUboq/P+byeQIAACQV6GDGa21dV3rsMS2bbXZYBiGpmlkLSZD7wMAAGuqSh76N9OzkHqpqoq+IgDAsWk19JXv/XDVD/r5t//wPvXd0apVEgUAwIaOlygU2vUAAMCu/frPf33rEPIodDAjAAAowQFbFFIrM9IlAQDAtQ6YKJAQAACQC10PAAAgiUQBAAAkkSgAAIAkEgUAAJBEogAAAJJIFAAAQBKJAgAASDrgOgosuAQAQC4HTBRICAAAyIWuBwAAkESiAAAAkkgUAABAEokCAABH0HWdtdY5F33VOWdftrBYEgUAAPbNOVdVlaQITdN0XRc9ZxzHGwo/4KwHAAAeStM0dV1LotB1Xd/30VxBz7kKLQoAAOT3N59+ftX/go/TzEC+mCYK0vVww4UcsEWBBZcAAI9DGgkuJgHS76BV5DAMC/OGAyYKJAQAgH355Ht/kLfAaBfDOI7DMBhjuq5rmmZhrnDARAEAgH356r/502vfcm1u4T9Fy+DHruuWDFlgjAIAAEdzsamgruuFRZEoAACwY5ITBG0DQaJw80hGQ6IAAMDe1XUdzHrQ7EG+sNaO46jnyJoK0SmUU4xRAABg32TMgT+jQY/rIkvDMDRN0/e9fNu27cI2hupgcwSq6mhXBADYEa2Gvvbtj1b9oE++9wdBfbdknuTCuZS+o1WrJAoAgA1tmCishDEKAAAg6YBjFFiZEQCAXA6YKJAQAACQC10PAAAgiUQBAAAkbZ8oOOe6rrPWTld+SB2XFSQWLlINAAButnGiIBtYSX3f970/DrGqKlkXIjgubzHGOOf0vQAAYA0bJwp937dt65xzzskgRGk/kP+fTqfguLxlGAZ5S13XkjQAAIA1bLw8UbA+kq5NXVVVXdfaWqDHu67r+17fIo0KfgksuAQA2BALLmUWXOQ4jrqupL/ApOxmYYyRVgT/uJlsmQUAwOZOP//8qv/d7UK2H8wopBXBeF0MN2+IWV0v00UAAHA0qy+4JIMJpsf9uQzSYOD3NTwHXQ8AAOSycYuCNiTI+MTgpen52gcxcw4AAMhl9RYFa+1MJ0LTNNGGhOBgMDTBP26e0UkBAADmbTlHQOYstG0b1PTWWn86g3w9DIOcVlVV27bScxFMjjDMegAAbEqroa/+qx+t+kE/+4//6D713ZbVqsx1DA5qxe/3MmhmYM55g54fxE+iAADYEInCXc30LKReIlEAAGyIRKF0JAoAgA0dL1EoZR0FAABQoNVnPdxfagElWhoAALjWARMFEgIAAHKh6wEAACSRKAAAgCQSBQAAkESiAAAAkkgUAABAEokCAABIIlEAAABJB1xHAQCAzX3hk0+3DiGPAyYKrMwIAEAuB0wUSAgAAMiFMQoAACCJRAEAACSRKAAAgCQSBQAAkESiAAAAkkgUAABAEokCAABIOuA6Ciy4BABALgdMFEgIAADIha4HAACQRKIAAACSSBQAAEASiQIAAEgiUQAAAEkkCgAAIOmA0yMBANjcKz/90arl/8WqpXsOmCiw4BIAALkcMFEgIQAAIBfGKAAAgCQSBQAAkESiAAAAkrZPFJxzXddZa51zwUtyvOu66Vvk+PQtAAAgo40TBWtt0zRS3zdNY63Vl6qq6vveGNP3vT+Roeu6pmmMMc45fS8AAFjDxrMexnEchkHyA6n45bi0Iuj8haqquq6Tg33f61skz9j1NIeqqvYVPwGvbXcBmx3GTMBrI+Aj2bJFQRoDtBVB0wVjTN/3dV3rmXVdy3HJFfQt014JAACQ0ZaJgrXWz+C0kcD/Vr8ex9EY45zzEwg/twAAANkVseCS5gHDMPgHbysttTLjDFqcAACIWj1RcM5Fn/j9XgN31jSNjj+4GbU+AAC5bD89Uug0SM0qoumFtj3MnAMAAHJZvUXBWptqIei6ru/7aAOAjl4UwdAE/7i5vpNiyejWLOfkGkZLwM885yICfuY5FxHwM89ZIss/cAJ+/mc9/4NuJssLyRJEGYvdskVBmhC0D8Kv9buu05YD59w4jv7J+pZgcgQAAA/IOVdVlS5KlHdK4MYzR6VRQb9t21Yvz+9l8I/7yy2YyYiEop4DigpmyTlFBZPrnKKCWXJOUcHkOqeoYJacU1Qwuc4pKpgl5xQVzPJzzLlievXpvfmTn+njD9/0Vxvy1xFItdbfpoglJlI9CDM9C6mXivqlKSqYJecUFUyuc4oKZsk5RQWT65yigllyTlHB5DqnqGCWnFNUMMvPMRslCv5UgKqq/AfsZypleuRVx+dfAgDgcdw2XG+5IhKFvJaMJbnbOUUFs+ScooLJdU5RwSw5p6hgcp1TVDBLzikqmFznFBXMknOKCmbhOeqjD95YfvJVJS+RcVbg0RKFEnpSAAC4SsmVVynrKAAAgFwy9kSQKAAAsGPRbY9IFAAAwK/Uda1zHIJtlp+PRGED0SkrspZWaRtn6yJf08AKD3g6kKfMgNX0X3WZATvn7Mv8V8uM2aQDKy3g6e0NwistYLG7f3d6n/cS8EWyMmFVVVVV9X3v77CYwQn3JT+/YRj8g/Kz0FUmNwot1LatRDUNrMyAJR4NuK5rfanMgJUE5v9WFBuw/1sh9KViY97Rb8UwDPXL/JgLDPg0+UNR+B0+7TDg5YZhCCqXLPZ3I/ZL/gTI75//s5TfWv3WGNO27f3Dmwoi0W9LDlhvrCRk8nWxAQvN/TX4kgMOkgNVbMzTbEbuc7EBB8r/NfYj8R+ESg54j38oNkSicD/DMLRtK7+LwbOj/4cs9Yf4/oKEWgMrM2D/H7zQ+1xmwEr+Hvm/FSUHnPrrWWzMZtKAp8fLDNhX17Xe7WIDnv5BK/nfXfQPhdzkMgMuAWMU7kf6vaJdX35Hb7CV9oZOL8/rHcdR4ywwYGutH7BEWHLAwlobXWm12IDNefuZqqqCLt4CY9YV66IdzwUG7JO98fyYywy4ruumaZxz0vFv9vDvLqC/xnsJ+M5IFIqQcXjqGqRiMN4wzJIDttZWVTWOoz+cp8yApzWBKjNgJV2hxhh/h7ZiY67Ou+r1fe+vf1dswGI6JK3MgHXHwqZpyv93Z88bFMu3M7kjFIkCLrDWNk2jnQ7lc85JL4885WwdTpJzLu8Ob/dxOp2CEePljw9v21aed+Vulx+wWWGG23pk30Jpoy7/350xZhgGSRllgsDW4ewAiUIRyvx3pQ0JwzAEEZYZsNJGZo2zwIC1JhDGmKZptGIoMOCo+ryzrSk4Zj8z2EXAxpi+72Xkiq/AgCUkDUxutd7wAgM2525KnSBgvISszIA3R6KwPf8vlzHGOaeTI7YlDQna76jKDLjrutSuKsUG3LatnyjU5yVTygzYnGef+0e0E7fMmO1kxTodalNmwCLaVFNywD6NqtiA/RxdjpT/K7GxDQZQPjzz8iBhfxRudJWFTUgkbdsOLzuVGvApPU2r2IB9JjFlq7SA/ZvsT9YoNmZ/7PpefiuCeXqi2IB39+/OeLMbJBWQr4sNeHMkChuY/v75eWshM3enzZ7Tf11FBXyaxOwHVmbAvuC3otiA93iT9xhwdGJemQEHIy7Lv8NBwP5LZQa8ueq0t7FUR6XzuDaOY7FiA04FVmzAKSUHvLubvLuAU4oNeHd3eHcBb4hEAQAAJDGYEQAAJJEoAACAJBIFAMcka3T6/AmHuhB1QLqop8dtbEtiHJj+MuDXtg4AAFYxjqMuTWGM6bqu73tZolHPCQbAG28Um/9eWUazaZryB3U553YR57aqivF5VyBRAHBk/qJ7Uon6y0bND273T5MVvVL7ugEHRtcDgEcRbAi0Bmmv9jsytAfEb5+Ydnbo27WnQw9KjjI9eUoyIQ1D98zUN/pF+SmUnhZ8in9Ql2cOlkDV8/3r0sKnMfhdQv7lpI5H2fOOoBqbH2r01un9lxOql7ui5kVvThDDkWdUbrqKAwCsxcSWLdKV+HTpvenao9H3RldLnP9Q+Rsr6/b4bzeJlQGD98rX/vKXS2LwlxTU1bH8b/1lEyU2+ToIO1qU7hqqN0pDmhboX3g0huk9keP+PUmRc/zzJf5UGNOvL666OBPzcN4kIlr+8ZAoADimaKIQVGypZ6foS0sW9PVPCyo8v7LRg/6ywebltZCjtebFGPwCg8TCz4T8j/Pf4r9r+vZpnLpCtr9U9kwhwWnTxZ71I+avNLUy95KQpPyZwoMYzGTVVH833Wj5B8MYBQAPJGjTPqVHtPmDGY0xTdMsHCToN0FHdxXqus6d6a5acrIMt9TtT+XkpmmqqtItxC4GMB+YfvTMabL5sgz/rKpKboV+dNu2fd9LhOM4Sj0dDB1NlR+9nOgmWNMdyKKXk/p2eUjzNDYt3/+R+Sua6307HsYoAHggUnksPNl6JEXIMl+uqiodU+nXNM45qXSlepaaST66bVsZf1AltkhdQjraZRDDwu750/lBWT7ar9FlDIRJDAgNLs0/vuRyJI1YEuFyqZAWvle/bts2dWnmoBtV06IA4FFEn1w3iUFbJmTipX6tqzXosMQg5uoZMy/6vm/bdvl7ox+tFaFErlWmpF9+4dEPipYpz+LBe7MkCn7bzG1VuKZr/thPfXX69SGHNNKiAODItKVdGr3ruvb/lLuJi6XlDS/o3ZjWahJ23g81yy5EkxUV9D4YL35r7TiO/oyA6EdMsxytg4MVLxZdxiwNUvi39ypB/4V/T/xLlt+uW4Mt24bjIwBgPdM/d9NRhKlzTGwg5JK/mSY9cM+8PFpeP06LTW3h7R+Mbj8djXMYhmB4nV++xGa8iQzBnYnew+BTgmCCatL/0KDw6eUExy+O2azr2h/4mbrtfkjBz+Lip/gnRGM2kzEo8zHvF6tTAcC9+c3UQZN1tAXbf1K/qvCLH32xwKs++uKnz5+zRut9Kv4l1z4tx3+LDMnU7phDdjoIEgUAAK6micLWgayOwYwAcAV/QN/0pTs8Vup0gyl/7N7e3eEyN/9R7gUtCgAAXO3azov9IlEAAABJTI8EAABJJAoAACCJRAEAACSRKAAAgCQSBQAAkESiAAAAkkgUAABAEokCAABIIlEAAABJJAoAACCJRAEAACSRKAAAgKT/D50nZr56rabzAAAAAElFTkSuQmCC\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAArgAAAHYCAIAAAApvgy/AAAABmJLR0QAAAAAAAD5Q7t/AAAgAElEQVR4nO3dT8gmyX3Y8SqNNRLWWjsbZQUOEb6s7YNiYoLBi42numcjgQ8JxBevAj4EOXeBr6vt7rFzNOic4JMOssHYEB8MVjxdrcVGBwn8BwU0GjD2hhw0mJ0ZS5a061Xn8PNbqanu6rf7eau76+nn+2EZ3qeffuqp588+9euqX1Xpvu8VAADAmA/sXQEAAJAvAgUAABBFoAAAAKIIFAAAQBSBAgAAiCJQAAAAUQQKAAAgikABAABEESgAAIAoAgUAABBFoAAAAKIIFAAAQBSBAgAAiCJQAAAAUQQKAAAgikABAABEESgAAIAoAgUAABBFoAAAAKIIFAAAQBSBAgAAiCJQAAAAUQQKAAAgikABAABEESgAAIAoAgUAABBFoAAAAKIIFAAAQBSBAgAAiPqRvSsAADgarfXeVcD1+r6fcxqBAgAgvZmNEPYyP5hj6AEAAEQRKAAAgCgCBQAAEEWgAACAstauWv6DBw9WLX89BAoAgO3Udb13FcbNrNjJ9e+6bvR4URSnFbgZAgUAwEastU3T+DfdH/7fwyPu7y9/+cujxY7e9B84vDlRyWGB7kjTNNPVjhVrjBktsOu60UcFBU6/kOHblbCDhEABALAR1/gVRVEUhdzUWltr67qWi/WyLOWusiyVUnLcWivT+T796U9LUZ///Oc///nPS1Fyr7SOWmspxxUrl+xymisnpigK/+lcPV013KsIynfVjpV///59eaw8sCzL6ebcL9AvOXhDhgXKH/LvnA/lej0AAElNNC7GGPm3qir/eNu2cpd7bFVVbdtWVeWf+cYbb7zxxhvuNHmIe7grRE5o21bu8s+U0yYq5j+7f7L/jH6t1NWiETPLd/cGrzfgjgd/+A9v29Z/yVIx/+aE+QEACy4BAHbgxuanB+nlAlprbYyx1v7mb/6m1vru3bt3795VSnVdFyunbdu6rruuq6oqODPGWjtRYKz+/pjCtbquc1kOp2UnSDeGukp6kD4P6V2Qlyxvl9xMkgDB0AMAYDfSde8av9ET6rru+14GBZRSd+/e/a3f+q033nhDKWWMkX74YTe7HOz7vmmaoigkznCjHqPPJceD02Id+H6SwfzXO1HhmSRJwlorAYpUVXoapGS5WVVVqrxRehQAANsJcveKopBcBLmaHzafcrlsjOm6rm1bdRUofOpTn1JK1XVdlqV/r+OyDqVBnYgPfFVVBadJOX757qpdjkuPxUyjFXb5GXMYY1wN5YFBgbE35HQzhygAAJgpeeMyOup/7b3D49PlDE8bTThwR2aW5opaWo2YYQWCAueUP/8z0j37dgAAktI668Zl2PMf62+Q8Y6lI/3D8u/evfvaa69NvCfDHoW1F5yY/xll/VkCAM5R5oEC1JLPiGRGAAAQdbRkRq31T/3S7+5dCwDn5+Fbrycvk58jHMApPQpudSpJ+5TMz7TVAoCNzUzsWmTv1wQksLhHoSgKtx5F0zQyLaQsS/6XAAA40ysl44wszjfRWstiT3VdN00jD3cHV6njEhNfTfoAAUz45ld+NXmZP33395KXeRYevvU6V4+Zm5/MeEqOglusatG6lZshIAAAIJVTchRkuU23YLUkKOTQnQAAANJaHCi0bdt1nawQ6XYCXbSAJQAAOBenJDP6oxrBTQAAYiQdPjgo2zWdUJpcqZ5RGyRX1/7N4TnuBLfDU7BEo73iLxk53FXrhAUlY462jgIAIFtd18n2iYseNZEvn2eq3CjZ83riBP9eiaiMMU3TuHkD6io2klct3fkuB0DOX6PmcwMFtxd4bFrBGcV0AIAdpbrSDS7QsyV9A8MoIai8xEPqKqRwrarW2m0vGQQHZVm6qOvkjplrzc1R6PteasC6IgCA5Oorso6fCybk6lRaROmKl0tWuelfu8qG1MEygHJOUObGZMep6cv9oijcDtfBpMKqqpqmUVeBhYsM/KH/VQOmxcmMblnGbD186/XR//auFwBAKW+U3d9l0VorzWHf95I1Ly2iXGS3bSs3JZveXVU70lcv2ysbY8qylONyclDmxiTPYOKp/YmESim3quHwNPlDSvPbYumuGMZJSSzOUXCf5fBzygTrKABAtqSl94/4rYn8Idffsbn3w3wFaWj9y2vpcpDT3L8Sc2RoevJgURTS7ArJBFBKNU3jDzdInKSUquu6LMuEqyCeEigIybCQtBQWUQAAzDExlD4zF2/Y4gR98u6gG9eX+fx5XtyOVn7iTBcBSI6CDMf4CQD+0Mx0gZ/73Oe+8IUvXPu8Jy64VNe19OQopcqyZE1vAEA+XC+Fa6qappHmc9+KDdV1Pb0WkWvv/Q4S5a2SPHyIMWZ6hoWYEyWo0wIFJ+EoyOiHJ30Vw7vc1NLMsyUAABuQJrP2SP6gumovrLV937uswKwMMxKCTheX2zjaeS+vbnhXwqmSJ24zLcml8o5L8shNKiEDGcOJIk3TyNiMn5ohoy/K63W5yVMDADZmB25YYHCpLa2DGjQTGbYXLicjONh1nZ984L/AYT6Hv6uzGqRG3tziHAUZZTDGJEmUkDGkYQ+JlCzxh8QlbvmtpmncU/vHAQD5GyYzqhkr8Ui630Sj07atPw4unfmSFOk/XW75jHZse0WZKumq7S9RJS/T9Yu4lzN8+XMa6Jk5Cou3mR7t4nCmZ4CMliZBk9/8K6W01kEerCxN5e9trcbW79RaM+sBwAnYZjqh5NtMTzc9/mkqnu14dnn3S1/Oopeptf75n//5r371q9efmfazPHmdrIkVOtXVBNm+74e5G8EDtdYf/5WvnVCBCXceP0pbIABkJXmQpHXixgXJzQ8UzmCvB7eMxszzv/0HP7f0KZLHFgAAZO5zn/vcnNOyDhRkoEGNLa8xgVYfAIBrvf76rDWLbzQ9clUy30GW3vSnjfqZjxmmsAIAcCSZ9igEe2fFnGl+CgAA5yLTHgW3hHUw0VbmQbjZEHLabrUEAODoMg0U1NVcW58cb9tWlmCSCaOMPgDAuZDF+oa/27K7dIbrK6/K7UYRmOgml2mivuE5yZctziVQ8BMR5OaQ3CW7X7Rte/MVIQEAW5Iks2HrmOHKyhsYXVxY1lWMPWR0icLg4cOVjm/opjkKwVjAZvEgeQkAcKaCpu4yO4aHmynGlioO+KsRDg1Xvby5xT0Ksn+lfK6yHYPsyCD35tB+f/sPfm70v73rBQD458WV/eCgrusg20waGuE3KzJCERx3rZK7axh5BPsLjp6zJXnJo4s3TyfeTYcRsvZzgvo9b3Gg4BbclndZBgX87St29/Ff+drof3vXCwCg1PObFyilhjsYSUMjo8xuuyPpVJdBZznuP6qua7kr2NzB3es65KUx3vGyNjY6INHMtR3zLiQa3UpqjX79E3ePVJGtLAAAmOZ3sA9nuUtD7rZGlJhA7nJ7Hck5QbdEsLmiz22zFGzMuIuyLG+4N5Vk6Slv/EKCj5Xy9k7JUZCsS1kNKXmFAADH5noI5Bo6uOYMEh6lUXfZ/m7CfNd1/gOv7SGQTgXpk0j3Uha7YX+GHwpI14J0QkgfTJoqDizuUZBeHYliZA6G/J1DdgIA4Cy4PoPR3vIgAnDj3VprGVYYhhfnous6GUyRhYbl75v0cLh1hlwspZTycwdvbnGPgqRl+rM/g/4fAACmyUXw6Oq6EgEE+QdKqbIsb9LcSCEyALHj1Hr/ul/6TmLLIQxJTBDEVfLwoNi0SRinDD34tbxhKAQAuEDSjEnbH9zlYgiXjjB8+HDoYZpLhJRx8x1bLr8B9VccnuDHB7LghL8wQbBIgbrqTkiY1XhKMqN7i2WNrQtcTgsAcENyETxsPmQUX8a4tdYu91CyGuWgRBjzk/zdfD1Xzhld4rqNlKXnwK1NvFm+hV7aAyPDKhKXaa3btl012XIprXXymZB3Hj9KWyAAZOWbX/nVtAVqvbhxGRodmPAPul6HSzP6ziw1/zNa/Fm6oiXGkb8lYsjhAxuudeXc+ey3Tivz4w+/fmp1op68/EryMgloAGTi4VuvZ3L1iJj5gcLpSzhnu47CyQEBAAAInBIouHkdMjrilsVIWjEAALC/xYFC27ZlWbrZF5KgwMpLAAAc0uJAQXZ5jt0EAABHcsr0yKHdd+ICAABrOHGb6cAaNQMAHIzMqx9eWMr+0SzJk6dTtpk2xkgaY1VVbdsaY9xCFgAAxAQbPjmyoNDlkL0uY4HRdHvqNtMajbfSrskoTl9HQeY+SIWSrK2RBOsoAMDuYusouJ/oYBdE2eqpqqpL6FSQN0HWF3ALGLp75d2INaky5dA91m1+4b+HEnVdu7jR/Ib79ByFYK3sfHIU7nz2W6P/7V0vAMA/r9zsNxnDrSD9MW6/tZMRiuC4nOw/ZNgeBQfXuOyeSZ6373vZwcHtoqmuugqkvY+RwEIeW1WV9ND4xdZ13fe9X+zNnbjXg7qKa1LVAwBwIYJmbLhlg2zN0Pe97MvgFmyW3Q3c8WCHSblL9omYeEbZUCqfrgu/JZV9LmJnBis3u7dFCvHXKUjbQC8OFNzHI1V0yYzkKAAA5qjr2jVjw20LpLF0S/lJoyN3uXw4OSfolpC7RiMAv+Hct//br6SELK6Bl36OiQgmWI/AX+1Q+hLcXWmXTr7ROgrSeaKIEgAAs/lbOg3HHYKERzcGL4lx9kqwzfR0MySbE0kh+y4S6HaAlEyC08YIXNwzunukv3RyEifmKLiPyt1MVSEAwOG5BnJ0FCCIAKQjQbIQZFhhGF7MeUbXbO047iAhiz+AcsKVtrVWphwGgyySwzFMkLyhxT0KbmPsQCazHgAA+avruizL0T5piQCC/AN1NTn/5OtS6UvYfTtD6c9wAyiyK8IJ5bjIyY1WuI6E5H38i3sU5EX2A2mrBQA4MGnMpO0P7pIMBhcQBDPshAw9LHrGiQ6McyEdBqN3SZTQ9/0amQCnDD2c77sMAMiEJAoMGxRJVCzLUpLlXdAgHfVyUCKMpa2+BCX7JtUZY/xe+Tn1t9b6eYsubPIPujUVfKnqvHihJH+dpQyx4BIA7C624NIiowMT/kFJTpxfYNDQ7iWYuxi8UcMFl1xvgRqM/rslqkbbvumPYP6CS4sDBbf803B9jEXlrERrnXxtJQIFAFgkSaCQljReawzhn+YmcwaTzDdcMVDw15705RMofKz+Rtoyb/3V99IWqJR696UXk5d5+52naQsk8gBwmtwCBbfycSZNVQ7mBwqLZz2odSY4xFaZiHU9ydoaInllAABHImP2tBenOX2vh4RkYc7RQM/Nn3FkUo266kciPAQAXIso4WSnLOE8OlnlNBN7YBRFMZqdIUtVSHg4uqY3AABIZXGgUJZl13Vu4opzcg1k1a2Zx/3ZIIqJmgAArGxxoDBcaslfcGlpT8PEHhijx4NFtTKZ6wIAwFGdksw4oa7r3Zvtv68/ufQhySdKAABwDIkDhRzQ6gNAtmJLBl37qOFqg8O9h85rAqS8In+fzOCtGF27euIceZfmL4r46quvfvWrX732tCxmPcwXLGh1Rl8IAIA07W7DINlw2bVqEzPeJ37t2ytVVckaz+nrvY5g4p5scjH9kIlzZNcrpVTTNDPfhFdffXXOaefdo5BkdSoAwDYkSnCRgVwB+7HCCVwTIFfSWuuEU/PWM2zLu66b078yeo4EEC5f0N9VcsIXvvCFOVU9sx4FednuxTdNs++GoQCARYIm3Frbtq266jD2uwRkkrw0/PPLl34FV7ibmufHE/6UuqqqgjZbHuVurrG9UV3Xxphh+3XtK411JwStYd/3Cet8ZoGCUqptW+lXkQ8y/7ARACCkFQ+aXjd/TdpOuSyWuKFtW7k5f1Npv62VTSb7vpedJ+UuY8z9+/fdOffv3w8abDlN/pUr9bSBwsQag9IjMh0buegnOEf6UdYIa3IJFGK7aA/DoqIo5FN3XyAAwFmo61r6D9z1XqxVkyjh5DnwLunPjVBLrKDG9ip68803gyN930t8IEv8LX32abI31bDC8oe0btPJFnKOen78omkaWYFQ3tuEV9HnmqNAXgIAnCOXQCBtedM0TdMMt1pWz//OLx1ldhfWwzkC9+7dU0pJHqWMQciRgJxgjEnb3BRFMVqmXAO7m33fS2MfnOmfI10LLhfBn/EhLz9VrJC4R4H2GwAQE4w4uASFhJe/E0W51IQ333xTRh/u378/7E5YleRhyPwO93esznPeFj84cAeDGYI3tKxHIQjQXFTo4sEc1lSOLbjE+goAsK+maYI5kKOXl27Ewd0rm0TPfAo50yUZuLtk22E55/79+w8ePJC/Y+VIp0LabSf9PEppy+UNCdZUGDV6zgbX53MDBdmqUf6WMZtsd2MiIACAbMkIvT8HQUVaO9d5Pn1h7e6Vq1bljVz4zbxcZPut7GuvvRYLPrTWxhg5uSzLhPlwfgWknv48PhdF+ZflfnwgL9C/V/4NVqRwoypJzB16kLkcbokMiRKGez0AABDT973s+utmLfqz/6Utl5bSJfRprWXyQqzM8or0JfilSWvlnshPIZQRh9FxBz8Lsn5+J8L1BLX1kyjdApRyjssDDc4xxrh31UU5SeiZbbzW2l/kQWvth4T50Fon71G49VffS1ugUurdl15MXubtd56mLfDO40dpCwRwIR6+9fp04xJbLi84nmRVvfNamm9ObaffvWsfLrSeHQDMDxT8yCDnQOETf/hHacv8wRd/PG2BSqn3fuKF5GX2z26lLTB55LESAhogN9cGCtjd/EAhl3UUAABAhggUAABA1ILpkUFmxLW7YQIAgHO3IFAIVm9IuJgDAADI09xA4YzSUt7+T/9h9HjyJEcAAA5vbo7C9MhCDgsyOp/4wz8a/W/vegEA/v8GidObQg0fNZxnFxS1dENqzDQ3UCjL0l/9yt+xyi2GBQDABFk7qKoqt3yfv55gsLqzb+Jitb0ie1hPbLqYoeGSzLKUpG/4kGJgWEjCxEFmPQAANiJRgmvVZD3BG15q+o2ljJKfUb/CcI/Huq6XpgB2XRcU4l/b3xyBAgBgO0ED5jaQdLspui6BoihOGFCQfgVX+HBUQhZC9s8POiGCC/TRUY+bk1c3jAm6rquqynqGD/Tv9RecdsWmrSqBAgBgI9KKBy2x2yvSGOM2a5C4oW1buTn/Ittv1GWTiL7vZecIucsYI3tMi/v37wcbSchuUu6m7NW07HXOq2ds36ZFTyebbM0p9mQECgCAjdR1La2a29kolswoUYKLIZY+UXCpXRSFxApKqeH+UsHWUP4Dgw0eE6rrWkZeRu+a2ZUy3LNKik1XTaUIFAAAW5Kec7nKlwSFYVf5cNOjid0jJ55FXTWc/o7V9+7dU0rJZbf8K0d8bvfFjdcSdE8n6ZnX5mb6G0iuZ8GCS7K1tJNtZinrKABAnvzrXTfWLpl3qbr3J5p21yf/5ptv3r9/v2ma+/fvj+407UYfJPsyScXmKIrCX7Wo73utdezN2WwL7LmBwgYxSyoEBACQJxnv99u20XbOjTi4e7uum9mp0DSNnCmPHU1LlBDhwYMH8vewEJnGKY/dfaGgWKCwWRAzN1A4o9kmAIBsSfKdPwdBRZoYf7xgokB/jR9p9d3IRdM0rpWV7Ei/1X/ttdcmgg8ZFlk65HFD/kSGa89UWwUxy3IU5B33Z5uwFxQAYKa+740xZVm6dqTrOtfZLm25tOtuhF5rLZMXYmWWV6Rd90urqso9l2RHukfJiMPouIN7uNqjO0GCG/nb79IIAogtK6bnb+IgH6FSyhgjIYK7mU+4oLVOPvTwgy/+eNoClVLv/cQLycvsn91KW+Dtd56mLXAldx4/2rsKAJ7z8K3XpxuXYbri6PHYaYucVoiMPqy9z9FwWoc8r7vpel+kCXb10Vr7S1cFpu9158x8dXPPk6r7/UVC8lCurdBmCBQSIlAAcJprA4X8aa13vAxOEiFNSx8oTIQn24RdM2mtf+aPv5i2zLd//9W0BSqlfvjJ7yYv8wPf+EjyMpMj+AAuwVkHCnIBrM5q2+QTzA8UFuQoxEIb8hwBAIchqzMdO0pYZME6CgAAzJTtWjtY6oCBwl//8q+NHk8+JAEAGPVTv/S7e1dhN27Y5S+//J9XfaKf/fSXtun2WBAoZJKueC0CAgAAUpkbKBhjlu6QDQAAzt3cZEbZw2PCDesRm08R7EbqKiPH81m/AQCAQ8pi90hZdzNo9bXWbk8OPymmrmuZuOK2Etm0rgAA5Go6SUCWdxy9Ap+wc6AgfQPBvpTq6qX2fS89Gcp78bLukyyILUuBblpjAACyNHrV7d/rrq5Hd/eO2b9HQZbjDg4GW3G45bGCXTXPJb8SAID1xK66fbJlhlxmyxX4zGWQdg4UpANktL0P9iGVVErpRQjOYfQBAHDhRq+6A35rO39jzHzXUTh5wcfYOgoTmFEJADhfRVFIo+lvKBUIph10XXdtYCHyDRRORqsPADgvP/vpL235dIs20c43UJARl+BgURR+uMSgAwDgAP7iTz6z9CGnxRayj6MxZv66BpkGCsHmnkFqgn9csSsVAAAzyEyHtm0XtZv7z3oYVde1WwjSWtt1nfSQ+P+qweQIAAAwSmstHQlLr64z7VEoisIY42Z5VlXlXljbtmVZugEIRh8AABglKyy5hrIoiqDRnBM05BIoDAdLZK6nGryMoihkIabhXQAAwJEueXV1Ud00jZ/nF4zyx+htNqncjNY6+ayHt3//1bQFKqV++MnvJi/zA9/4SPIyk7v9ztO9qzDLnceP9q4CgLN00dtMn4v/9sGvpC3wv3zi36UtUCn1w+QlKqU/+n7aAvtnt9IWqJR696UXk5e5RvDx5OVXkpdJ8AFclL/7wAt7VyGNAwYK//Hf/4/R4//zf/3XjWsCAMC5O2CgQEAAAEAqmU6PBAAAOSBQAAAAUQQKAAAgikABAABEESgAAIAoAgUAABB1wOmRrKMAAEAqBwwUCAgAAEiFoQcAABBFoAAAAKIIFAAAQBSBAgAAiCJQAAAAUQQKAAAgikABAABEHXAdBRZcAgAglQMGCt/5s1+I3PONTesx6d0X/yl5mR959n7aAt9/dittgSt596UXk5f58YdfT17mk5dfSV7mncePkpcJIIm/+8ALe1chDYYeAABAFIECAACIIlAAAABRBAoAACCKQAEAAEQRKAAAgKgDTo984Rf/fPR4fNokAAAYd8BAgYAAAIBUGHoAAABRBAoAACAq60DBWlsURVEU1trgrrqui6Ko63qHagEAcDHyDRTqui7LUv4uy7IoCneX1rppGqVU0zRa612qBwDAJcg3UGiapm1ba621tm3bruvkuPQi9H1vre373h0BAADJZRooyFiD60WQPyQgaJrGGOPONMYMByYAAEASmQYKo1xA4A9DFEXhOhsAALhka3SxZ7qOgutCkNccvHI/UBiKLbg0gaUXAADnzlrbNI1MAkhYbKaBglKqbduyLCVpcRFafQDARbHW1nW9Uv96vkMPRVH0fd+2bdu2krToQiSSEgAA8BVFUVXVGiXn26Mg4w5BOoIaZC9aa/3cRqXUh194uk0Nb+L20/Tv/A+f3kpe5sV68vIre1dhljXqeefxo+RlAhfob/oXN3suN9xwQjf8tfINFJqmkbmR6ipEcIkLbn0Fa23XdW3b7lVJAABu7rdfq/euQlS+gYLkKLj1lGT0QSlVFIUxxh2vqipt1gYAABv7jT+tlz5ks9gi30BBchSCBRVE0NMAAABWkm+gIGKhACECAAAbyHfWAwAA2F3uPQon+JF/+43R4//0l5/cuCYAAJy7AwYKBAQAgMvkEv8TYugBAABEESgAAIAoAgUAABBFoAAAAKIIFAAAQBSBAgAAiDrg9EjWUQAAIJUDBgoEBAAApMLQAwAAiCJQAAAAUQQKAAAg6oA5Cj98+eneVbhe/+PP0hf69ofTl3kObr9zBp+4UurO40fJy3zy8iv5l7nGCwfy9zfvv7R3FdKgRwEAAEQRKAAAgCgCBQAAEHXAHIXb/+r/jB5/9//+641rAgDAuTtgoEBAAABAKgw9AACAKAIFAAAQRaAAAACiCBQAAEAUgQIAAIgiUAAAAFEHnB7JOgoAAKRywECBgAAAgFQYegAAAFEECgAAICrrQMFaW9d1URTW2uAuOV7X9Q7VAgDgYuQbKNR1XZalhAhlWRZF4e7SWjdNo5RqmkZrvVMFAQA4vnwDhaZpqqqy1lpr27btuk6CBulF6PveWtv3vTsCAACSy3rWg+tF8LsTmqYxxribxphgYKL/2LO01Xjvo++lLXAl73/i+2kLvPX2h9MWqJTqn91KXua7L72YvMw1PFGv7F0FANv51vv/Yu8qpJFvj4IxRoYerLUSKIzGDUVRdF23Q/0AALgA+fYoWGu11mVZys22bd1dfqAw9KEPLe5R+MEPPrr0IQAAXIJ8exS01saYvu/7vq+qyiU2XusHP/jo0v9WfikAAKwuNk/QcXMJF+X2ZRooyOt0r1ZeknthMyMGAAAugfTBu3mCo3GAtdZdcjdNM90378s0UBhyCYxB9qK11s9tBADg0pRlKY2jtbaqKllBYOKcvu/dXMJrZRooSKTjdyF0Xef6FVz2on8cAICL5ZrCoA/eF/QizAwU8k1mbNu2LEsXFlVV5eY+GGPcOkvuOAAAF0ja+2ubQulp8K/DZ15m5xsoFEUhqyqpsSBo5vsCAED+/vqXfy1tgaNbHzRNMzqXcFq+gYKIhQKECACAw/iZP/7i0ocsjS1kLqFb47gsy7Zt5zSmmeYo3MSHPvRs9L+96wUAwEZG0xH8uYTGmLMfejgZ6yIAAC6HxARuFWP/YBIH7FEAAOCi+N0D8kcQPQQ7IcicwZnBxAF7FAAAuCiy4JKbD+gSFSUgcAfLsnTnXPTQAwAAl2Z0nmBd1y4amJhLOI1AAQCAI5jT/J+Qu0COAgAAiCJQAAAAUQccenjva/9m7ypc771//NHkZShHI88AABK/SURBVN5OXuIK9EffT17mB//2O8nLvGR3Hj/auwrAETz5/p29q5DGAQOFF37xz0ePf+fPfmHjmgAAcO4OGCgQEAAAkAo5CgAAIIpAAQAARBEoAACAKAIFAAAQRaAAAACiCBQAAEDUAadHso4CAACpHDBQICAAACAVhh4AAEAUgQIAAIgiUAAAAFEECgAAIIpAAQAARBEoAACAKAIFAAAQdcB1FFhwCQCAVA4YKPzFn3xm9Pij725cka3devvDe1fhev2zW3tX4VDuPH60dxUAjHvy/Rf3rkIaDD0AAIAoAgUAABCV6dCDtbau6+BgURTuYF3X1lr/CAAASC7TQGGo6zr3t9ZaKWWMaZqmaZq+7/erFwAAR5ZpoFAUhbXWP6K1liPSheCCA611Xdf0KwAAsIbzyFEoiqKqKvm7aRpjjLvLGBOEFAAAIJUzCBTquu66zu8zKIrC/9sflQAAAAllOvTga5qmbVv/iB8oDP3sp7+09CliSy8AAHDhcg8UpCNhOjII0OoDAJBK7kMPTdO47ASHpAQAALaRdaDgT3NwguxFa62f2wgAwGWq63o4Z3D0nEVTBc8gUAhIbqM7IchzBADg0lhr3SICZVnGmkWtddM0SqmmaWRFojmyDhSCmZCiKApjjNZaa12WZVVVizIYAAA4mLIspbvdWltVlUQDAWkr+7631spaRDMvs7NOZowtuSjvhVqY5AgAwFH5Wxw0TTNcirDrOj/nb/6ixln3KEwoioIoAQCA+VfOkp2wNEch6x6F08TWUWDaJAAgT//wmbtpCwyS/Fz6glq+U9IBAwUCAgDAefmxL31l6UNOiC38aYPzd0o616EHAAAQMzoS4YcF83dKOmCPwhdv/XTaAm+9/eG0BSqlPvLs/eRlvv/sVtoC9UfTV/L2O0+Tl3nJnrz8SvIy7zx+lLxM4AK9948/us0TSUxgrQ02Qhqecxp6FAAAOG/GGH/WgxqLHowxkqOgFq5CdMAeBQAALoosuOTWUHI7KUpAMHrO/FWICBQAADh7spKSen6UIUhXHD3nWgQKAAAcwcylFJYWS44CAACIOmCPwm+/Vo8e/40/HT8OAABiDhgoEBAAAJAKQw8AACCKQAEAAEQRKAAAgCgCBQAAEEWgAAAAoggUAABA1AGnR7KOAgAAqRwwUCAgAAAgFYYeAABAFIECAACIIlAAAABRB8xR+O/fvpe2wFvPbqUtUCmlXnw/fZmpffBvv7N3FXZz5/Gj5GU+efmV5GWuUU8ASdx+um4L+/1VS/fQowAAAKIIFAAAQBSBAgAAiDpgjsI/fObu6PEf+9JXNq4JAADn7oCBAgEBAACpMPQAAACicg8U6rouiqKu65nHAQBAQlkHClpra61Sqmmaoij8403TyHGt9U61AwDg+PINFIqiMMZYa621fd93XSdBg/Qi9H0vx90RAACQXL7JjF3XtW3rbkpMoJRqmsYY445LMLFx3QAAuBCZ9ihI2y9ZCMNcBH8YoiiKrus2rRwAABcj3x4FpZTWWjoPmqZpmsZ1KviBwlBsHYUJzKgEAGBU1oFCVVWuL0FrXdf1nHQEWn0AAFLJdOhB+GGBn4tAUgIAANvINFCQwQU/IOi6Tg4G2YvWWj+3EQAAJJRpoKCUMsa4HgWX26iUquvaZS9aa7uuY3okAAAryTdHwVqrtXbrKVVVJYGCrK8wPA4AAJLLN1BQV6sqqcE0B1mFaXgcAICLVde1tVaWFbj2zKIoZrahWQcKKh4KTLy829/8WNo63HrnadoClVLqnfRFvv/S7fSFIp07jx/tXQUA2/ngsw9u9lzW2rIsJWOvLEt/zuBQXdfBxgjT8s1RAAAAc0iUIN3tVVXJdkijrLUT947KvUfhBH9ff3L0+Mfqb2xcEwAAtuG6EKTDILbyUFmWbduWZTm/5AMGCgQEAIDLMT9pT2YDLE3vO2CgAADAeYn1hZ9suDKhLCjgNkOYj0ABAICdndAXvjS2kEGHpc+iCBQAADieYHxB8hX8rAWX/3htUQQKAACcMbfpgR8cBIFCcLPruvnJCvqE4Yqcaa2TJzPe+qvvpS1wJe++9GLaAm+vsYDEmWDNAwCnefjW69Kw/svmf6/6RH9ff9K14P4GSTLrQe6S9ZeG3QZa67ZtD7LgEgAAmBZseuByESSB8YaFH7BHIXbXyT0N9ChcIHoUAJxmlx4FsdLmBgfsUWAdBQDABVpp/yOWcAYAAFEECgAAIIpAAQAARBEoAACAKAIFAAAQRaAAAACiCBQAAEDUAddRiG2odeez3zqtwI89/usbVGfck5dfSV7mxx9+PXmZAIDT/PDprb2rkMYBA4WTAwIAABBg6AEAAEQRKAAAgCgCBQAAEEWgAAAAoggUAABAFIECAACIIlAAAABRB1xH4cnv/OTocdZXAABgqQMGCgQEAACkwtADAACIyrdHwVpb13VwxP1d17W1tiiK4BwAAJBQvj0K1tqu60bv0lo3TaOUappGa71tvQAAuCBZ9ygYY/xeBCFdCH3fy02tdV3X9CsAALCGfHsUuq4rimJ4vGkaY4y7ORpMAACAJPINFJRS1lqttda6KAo/GvADiKIoYiMUAADghvIdehBt2yql6rouy9INN4z2NDixdRQmMKMSAIBR+QYKLixQV10LM3MRaPUBAEgl30Ah4OciyMTI2Jkff/j1bap0E3ceP9q7CgCAFfXPbu1dhTQyzVEYhgIuESHIXpTJERtWDQCAC5JpoCApim6gQf5w/7qgQdZaYG4kAAAryXfooaqqpmlkYSW5KX0MRVEYY9w6S+44AABITvs5gxmSUYZhKBA7rrX+qV/63fXrBQDAiIdvvS4N60u/vm4u2pPf+cltWvB8exRErLeAXgQAADaQaY4CAABYpK7rYH3CgOy2uHQ/xdyHHpaa2COKIQkAwNp2GXqw1pZlKXMAu66rqmoYCgTnqOfXK5pwwECBgAAAsJddAgWttVs7oK7rpmmGjbt/zvDmBIYeAAA4e6MLCsTOUUrNX4Io92RGAAAwITYNMBD0McgIxZzyCRQAANjZCdsZTpseU5CoYmZKI4ECAAA7O2E7w9NiC8lgMMbMz1AkUAAA4GhGRyJkYmDbtovWIiKZEQCAMyatfjDWMLpysXQkLF2x8IA9Cg/fen30ONMmAQCHZIyp69pNj1Re9OCOy8Fr44mhAwYKBAQAgItirdVauyUH27Z1x2VtJYkP/K0WlVIz11FgwSUAAJLZcVOomfMklzpgjwIAABdope0SCRQAAEjv9jtP965CGsx6AAAAUQQKAAAgikABAABEESgAAICoAyYzsuASAACpHDBQICAAACAVhh4AAEAUgQIAAIgiUAAAAFEECgAAIIpAAQAARBEoAACAqANOj2QdBQAAUjlgoEBAAABAKgw9AACAKAIFAAAQdR6BQlEUwZG6rouiqOt6h9oAAHAxziBHoSiKruustS5c0ForpYwxTdM0TdP3/Z71O8k3v/Krycv86bu/l7bAs6jkSi75tQNI4s7jR6uW/+1VS/fk3qNgre26zj8ivQh931trJUSgXwEAgJXkHiiUZVlVlX+kaRpjjLtpjLHWbl0tAAAuQ9aBQlEUVVUNOwz8lAUZmNiyVgAAXI58cxTqupbUhOFdw9xGX2zBpQksvQAAwKhMAwVr7clZirT6AACkkmmgIMMNfs9BWZYuHcGfAQEAANaTb6DgDzp0XWeMkeghyF601vq5jQAAIKFMkxllMSVHXa2wpK5yF+Q0mTyZ+fTIE3Im1pZblXKrj7paqyMrGb5LuVUpt/ooqjRPblXKrT67y7RHYUJRFMYY9zteVRXDEAAArOQ8AoUgq9FaK6MPhAgAAKzqPAKFIUIEAAA2cK6BwoTY8BLTJgEAWOqAgQIBAQAAqWQ662FVczJaZ2a9pkqOnZNjP+echMm6c4pKVe2ZUn1wuX1qKl2VcnuLEj7dgauU4S9Sbm9RwqfbuEobk+mByfc/usRAAQCAI7HWaq0lRCjLMu2qAQQKAACcN7d4sbW2qqqmaRIWTqAAAMDZc70IbpXCVCUTKAAAcMZWX1ioP5a13iYAAObZrPGSx7ZtGzypUsoYc/MmVRxteuTJbzcAAHvJufFi6AEAgKNJOBJBoAAAwBmTmCBYPoFAAQAA/DNjTDDrgUBh3EqLUi1irZVqDKemxI5vZvi92atKub1F7lMbfnk2rtLoE8XqYK2V46t+53Or0sRnMXzey3yLcqvSCb+Kq1Ypt/okYa3tuk5rrbVumkbSG5NJlRW5L3lTjDHGGKVUVVW7VKOqKr8a/tsrN4fHNyNP3bbt7lXy3yI/L3ev+gSf2o5Vkq+x/xlN1CGodvCoDaokFfDv3aBKo/Xx7/I/vh3fIvddkj9yqNJeX6QTfhVXrdK19dnli51K27ZrVO8ggYL/AyEf6l7V8GMUdzOo0vahjIsug6/+9lWS/9n855Uq7fgW+c/l/8JuWaW2bd3Plv//+UQd/DOD1mjVKgXP5d9ctUqx+jiu1fGP7PIWDT81V6tMPrUtv0gn/CquWqVYffb6Yp+Fg7xgNbhW3qVTIfgCuUYx+P0KGsttKib/W7p3aa8qTfzK7/UWDb88cnPLKrVtW1VV8BlN1CH4kR3Ool61Sv7/XK4ma1cpVh9hPEHFVqrPRJViX5V8qrTlF2npr+LaVZqozy5f7LNwhBc8utbEXqMPo9WIfQW3YYxx1fADhe2r5D6pqqpcrXasj3Ddie5qbMcqzWyVh01RLAJLXqWAe8c2q9Kw2NG3Zce3SD41+Tr5H9+OVZL/9fxLebl3yyq58qd/FTeuUqyx2P6LnbNDJTP6dk9plP2FXVLMiotrTqrruuu60cymvarktjhrmsbfhXmv+rj91sqy7LrOTwLaq0q+HOoQUxRF8I7tIn3q1o01TVOWpbr6ku+eAVcUhTFGKtM0jTFm++9VPr+Ksfo4mXyx83HYQGFHRVHIRl793ittWWubptm9GoGqqmSLM6nYjtNAhNba768uy3L3n/X81XWttZYf031/7rXWVVVlGE71fS9fcmOMBA07ci1f3/dt28YuHlatQCa/iiJWn3y+2Fk5bKCwy2fsQtS2bYfTtLavj5tNK5RSZVm6d2av5tD/hTLGuGrsUh95UvfUwa5rOUQMo3WQ3/3pc9ZTFEXTNNJh675Ou1TJfXzyDe+6rus6qdK+b5HLnFfeb9GOVeq6zoVTRVFUV9sQb1Clpb+Ka1dpoj75fLFzc4S9HtyiVH5wsEug4HYED44HB6215iozeVXBrN+u68zVohy7VGn4Scnv1171GeWeN4cqzayDXXvvOI8MZl17XbhNlaTNczflGz76pFu+RTO/J1tWaeYTrVGlG/4qJq9SrD5ZfbGzs2E+xIomsp0349KF2uf1z6dbtvGJ4Gvzn3evKvmflP+8O75FKjI9cpcqqbEctNE6qOenk42m2a9RJXmu4Zd8sypNfBBB3tleb9HEDKO9qmQG0yO3qdJpv4rrVWmiPrt/sXN2kEChfz4M3KUZ9q9sHD8H2x3ca0bG6G/H9lXy358gJ3yX+rTPpyztW6XhtzdWh6Dam1Vp+CUPfu7XrtLE/+BBoLDXW9Q//6llUiX/eTer0mm/iutVaaI+u3+xc6b7PFJLksi8UyjD6u1Vpdjz7vgWZVilOXXIoXqB3Kq075d89Kmp0pzn5YuUj0MFCgAAIK3DznoAAAA3R6AAAACiCBQAAPsrikI/z19zRdY/GJLUgeHxYmzX+LTckgyHd4R1FAAA585f6EUpVdd10zTWWr+9bwfLKrvsQv+x1lpZRfvkJDxr7U0efjAECgCAXPhLx0pr7S/RNj3pwD9NFmOu63r3ReIPgKEHAECOpOFftaXXWrdtGwx2SICivB3sJvhjIkdd4JkeBQBApowx01stxPoYgn1bJty7d+/BgwdlWcpgh2ziJdthzBl6KMuybVvZbeSwoxV7rvYEAEDf95HVkd2S/MPsBL8JG71rzhK96vlFIV0d/BWmY4KVp+c85EzRowAAyFTQhdDHr9f9ZEalVFmWM6/vg5WkZV/N+S5hrUYCBQBApmQqxMyT/Ta773tJGri2IZd0BEwgmREAkCPpTlh72oI/qLHjBvc5o0cBAJALN9YgayEYY/wugfnJjO78a3sU7t2711/tt9l13YMHDxZW+fgIFAAAWei6zh8IqKoq6E4YDhMMz/E1TXNth4Qxxi2wKFMe3F0yefISshCmsXskAOBCaa1dd4J6PhC55H2lAwQKAIAL5QIFTGDoAQBwWHVdxxZMnJMmOfHwo67DOEQwBQC4UG3bMj3yWgQKAAAginUUAABAFIECAACIIlAAAABRBAoAACCKQAEAAEQRKAAAgCgCBQAAEEWgAAAAoggUAABAFIECAACIIlAAAABRBAoAACDq/wFglnuTTczqLgAAAABJRU5ErkJggg==\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import itertools as it\n", "\n", "perms = it.permutations(input_columns, 2)\n", "\n", "hists = list()\n", "for xvar, yvar in perms:\n", " try:\n", " xs = res[xvar]\n", " xmin = max(np.mean(xs) - 2*np.std(xs), min(xs))\n", " xmax = min(np.mean(xs) + 2*np.std(xs), max(xs))\n", " ys = res[yvar]\n", " ymin = max(np.mean(ys) - 2*np.std(ys), min(ys))\n", " ymax = min(np.mean(ys) + 2*np.std(ys), max(ys))\n", " except TypeError as e:\n", " continue\n", " \n", " model = ROOT.RDF.TH2DModel(col, col, 20, xmin, xmax, 20, ymin, ymax)\n", " h = rdf_signal.Histo2D(model, xvar, yvar, 'Weight')\n", " h.GetXaxis().SetTitle(xvar)\n", " h.GetYaxis().SetTitle(yvar)\n", " hists.append(h)\n", "\n", "# don't draw the titles\n", "ROOT.gStyle.SetOptTitle(0)\n", "\n", "canvases = list()\n", "for idx, hist in enumerate(hists):\n", " canvas = ROOT.TCanvas(f'2D_{idx}')\n", " canvas.cd(idx)\n", " hist.Draw('colz')\n", " \n", " canvas.Modified()\n", " canvas.Update()\n", " canvas.Draw()\n", " canvases.append(canvas)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The function `RDF.AsNumpy()` returns a dictionary.\n", "The function below can be used to get a numpy array instead of a dictionary:" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [], "source": [ "def to_array(rdf : RDF, columns : list) -> np.array:\n", " dic = rdf.AsNumpy(columns=columns)\n", " ntot = dic[columns[0]].shape[0]\n", " out = np.zeros((ntot, len(columns)))\n", " for icol, col in enumerate(columns):\n", " out[:, icol] = dic[col]\n", " return out" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We will now transform our signal, background and testing RDataFrames into numpy arrays:" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [], "source": [ "vars_signal = to_array(rdf_signal, input_columns)\n", "vars_background = to_array(rdf_bkg, input_columns)\n", "vars_test = to_array(rdf_test, input_columns)\n", "\n", "inputs = np.concatenate([vars_signal, vars_background])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We also need an array for the event weights:" ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [], "source": [ "weights_signal = to_array(rdf_signal, ['Weight'])\n", "weights_background = to_array(rdf_bkg, ['Weight'])\n", "weights = np.concatenate([weights_signal, weights_background])\n", "weights = weights.reshape((weights.shape[0],))\n", "\n", "weights_test = to_array(rdf_test, columns=['Weight'])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "For the targets we will use an array of '1's for the signals and an array of '0's for the background." ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [], "source": [ "y_signal = np.ones((vars_signal.shape[0], ))\n", "y_bkg = np.zeros((vars_background.shape[0], ))\n", "\n", "targets = np.concatenate([y_signal, y_bkg])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In the 'validation' tree (which is our rdf_test) the 'Label' branch gives the the true label for each class. In the branch 's' is used for signal events and 'b' to denote background events.\n", "To compute comparison metrics we first turn these labels into integers: 1=signal and 0=background." ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [], "source": [ "rdf_test = rdf_test.Define('IntLabel', '''\n", "const char ch = Label[0];\n", "const char s = 's';\n", "if(ch == s){\n", " return 1;\n", "}\n", "else{\n", " return 0;\n", "}\n", "''')\n", "\n", "truths_test = to_array(rdf_test, columns=['IntLabel'])\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Our arrays are currently ordered, such that first N events are all signal and the remaining all background. Let's fix this by randomly shuffling them:" ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [], "source": [ "inputs, targets, weights = shuffle(inputs, targets, weights)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "When training machine learning algorithms it is usually good practice to use three different datasets. The 'training', the 'validation' and the 'testing' dataset. \n", "The training dataset is the one which is used to fit the parameters (for example the weights in a neural network).\n", "The validation dataset is the one which is used to optimize the hyperparameters of the model.\n", "The testing dataset is not looked at during the training and optimization and only used to validate the model on an independent dataset.\n", "\n", "In our case the testing dataset is the data in the 'rdf_test' RDataFrame. \n", "Let's also create a training and validation dataset by randomly splitting our data.\n", "Here we will use 20% of the data for validation and 80% for training." ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [], "source": [ "x_train, x_val, y_train, y_val, w_train, w_val = train_test_split(inputs, targets, weights, test_size=0.2)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We will normalize our input features to have zero mean and unit variance using the [StandardScaler](https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.StandardScaler.html). This normally helps the models as they don't have to learn the mean and the widths of the input distributions." ] }, { "cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "StandardScaler()" ] }, "execution_count": 22, "metadata": {}, "output_type": "execute_result" } ], "source": [ "scaler = StandardScaler()\n", "scaler.fit(x_train)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "It is important to only fit the scaler on the training dataset, as otherwise information from the validation dataset will be used. This can bias the results. The same transformation is then applied on the training, validation and testing dataset." ] }, { "cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [], "source": [ "x_train = scaler.transform(x_train)\n", "x_val = scaler.transform(x_val)\n", "x_test = scaler.transform(vars_test)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As a first simple model we will train a boosted classifier.\n", "[Boosting](https://en.wikipedia.org/wiki/Boosting_(machine_learning)) is a technique in which many weak individual classifiers are combined into a single strong classifier.\n", "We will use the [AdaBoost](https://en.wikipedia.org/wiki/AdaBoost) algorithm from [sklearn](https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.AdaBoostClassifier.html).\n", "For the individual classifiers we will use a [DecisionTreeClassifier](https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html).\n", "An example is also given [here](https://scikit-learn.org/stable/auto_examples/ensemble/plot_adaboost_twoclass.html).\n", "\n", "Classifiers which combine boosting algorithms with decision trees are also known as Boosted Decision Trees (BDTs).\n", "In the following a BDT is setup and trained." ] }, { "cell_type": "code", "execution_count": 24, "metadata": {}, "outputs": [], "source": [ "from sklearn.ensemble import AdaBoostClassifier\n", "from sklearn.tree import DecisionTreeClassifier" ] }, { "cell_type": "code", "execution_count": 25, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "AdaBoostClassifier(algorithm='SAMME',\n", " base_estimator=DecisionTreeClassifier(max_depth=1),\n", " n_estimators=100)" ] }, "execution_count": 25, "metadata": {}, "output_type": "execute_result" } ], "source": [ "bdt = AdaBoostClassifier(DecisionTreeClassifier(max_depth=1),\n", " algorithm=\"SAMME\",\n", " n_estimators=100)\n", "\n", "bdt.fit(x_train, y_train)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Using the [predict](https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.AdaBoostClassifier.html#sklearn.ensemble.AdaBoostClassifier.predict) or [predict_proba](https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.AdaBoostClassifier.html#sklearn.ensemble.AdaBoostClassifier.predict_proba) methods the predicted labels or probabilities, respectively, for the validation dataset can be obtained:" ] }, { "cell_type": "code", "execution_count": 26, "metadata": {}, "outputs": [], "source": [ "predictions_bdt = bdt.predict_proba(x_val)[:, 1]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "To judge how well a classifier performs it is useful to look at the [ROC](https://developers.google.com/machine-learning/crash-course/classification/roc-and-auc) curve .\n", "We can use sklearn to compute the ROC curve." ] }, { "cell_type": "code", "execution_count": 28, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Text(0, 0.5, 'background rejection')" ] }, "execution_count": 28, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from sklearn.metrics import roc_auc_score, roc_curve\n", "import matplotlib.pyplot as plt\n", "\n", "fpr, tpr, _ = roc_curve(y_val, predictions_bdt, sample_weight=w_val)\n", "\n", "plt.plot(tpr, 1-fpr)\n", "plt.xlabel('signal efficiency')\n", "plt.ylabel('background rejection')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The area under the ROC-curve (AUC) value is often used to express the ROC curve in a single metric.\n", "The larger |AUC-0.5| is, the better is the classifier. An AUC score of 0.5 corresponds to a classifier which is randomly guessing.\n", "In the following, we compute the AUC value for the trained BDT." ] }, { "cell_type": "code", "execution_count": 28, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "AUC: 0.882973840745154\n" ] } ], "source": [ "auc = roc_auc_score(y_val, predictions_bdt, sample_weight=w_val)\n", "\n", "print('AUC:', auc)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In the Higgs Challenge, the evaluation metric was not the AUC value but the approximate median significance (AMS).\n", "It is defined as:\n", "\n", "$$\n", "AMS = \\sqrt{2((s+b+b_r) \\log(1+\\frac{s}{s+b_r})-s)}\n", "$$\n", "\n", "Here:\n", "- $s, b$: unnormalised true positive and false positive rates, respectively\n", "- $b_r = 10$ is the constant regularisation term\n", "\n", "We will use here an approximation of this metric." ] }, { "cell_type": "code", "execution_count": 29, "metadata": {}, "outputs": [], "source": [ "def ams_score(x, y, w, cut):\n", "# Calculate Average Mean Significane as defined in ATLAS paper\n", "# - approximative formula for large statistics with regularisation\n", "# x: array of truth values (1 if signal)\n", "# y: array of classifier result\n", "# w: array of event weights\n", "# cut\n", " t = y > cut \n", " s = np.sum((x[t] == 1)*w[t])\n", " b = np.sum((x[t] == 0)*w[t])\n", " return s/np.sqrt(b+10.0)\n", "\n", "def find_best_ams_score(x, y, w):\n", "# find best value of AMS by scanning cut values; \n", "# x: array of truth values (1 if signal)\n", "# y: array of classifier results\n", "# w: array of event weights\n", "# returns \n", "# ntuple of best value of AMS and the corresponding cut value\n", "# list with corresponding pairs (ams, cut) \n", "# ----------------------------------------------------------\n", " ymin=min(y) # classifiers may not be in range [0.,1.]\n", " ymax=max(y)\n", " nprobe=200 # number of (equally spaced) scan points to probe classifier \n", " amsvec= [(ams_score(x, y, w, cut), cut) for cut in np.linspace(ymin, ymax, nprobe)] \n", " maxams=sorted(amsvec, key=lambda lst: lst[0] )[-1]\n", " return maxams, amsvec" ] }, { "cell_type": "code", "execution_count": 30, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "AMS: 0.2626097248309577\n" ] } ], "source": [ "prob_bdt = bdt.predict_proba(x_val)[:, 1]\n", "bs = find_best_ams_score(y_val, prob_bdt, w_val)\n", "print('AMS:', bs[0][0])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The score is only around 0.25. Far away from the record of 3.81 which was obtained in the Higgs Challenge.\n", "The main reason for the low score is that the dataset used here corresponds only to 10% of the original data. Additionally, the validation dataset is only about 20% of that.\n", "Therefore, the score can be safely scaled up by a factor $\\sqrt{50}$." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "To give an example of another classifier we will train and evaluate here a [GradientBoostingClassifier](https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.GradientBoostingClassifier.html)." ] }, { "cell_type": "code", "execution_count": 31, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "AUC: 0.9087184621174675\n", "AMS: 0.30634630269636487\n" ] } ], "source": [ "from sklearn.ensemble import GradientBoostingClassifier as GBC\n", "\n", "gbc = GBC(n_estimators=50, max_depth=5,min_samples_leaf=200,verbose=0)\n", "gbc.fit(x_train, y_train)\n", "\n", "pred_gbc_val = gbc.predict_proba(x_val)[:, 1]\n", "\n", "auc = roc_auc_score(y_val, pred_gbc_val, sample_weight=w_val)\n", "print('AUC:', auc)\n", "\n", "bs = find_best_ams_score(y_val, pred_gbc_val, w_val)\n", "print('AMS:', bs[0][0])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Let's also take a look at the output distribution of the classifier for signal and background." ] }, { "cell_type": "code", "execution_count": 32, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 32, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEGCAYAAACJnEVTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8GearUAAAZgUlEQVR4nO3de5RV5Z3m8e8jlDIKeAMJAqEwgyM0yMUSGWWEqKG9BSGDtrSMmBghKJPpsWVUelZAne5lR3p6Oqsj3UZZmiyiGB3FC/GGItJq5KrgbUIMYgmjJTJEQmCC/OaPs6vmAHU5dTuHOu/zWatWnfOed+/9npfiOfu8e+93KyIwM7M0HFHqBpiZWfE49M3MEuLQNzNLiEPfzCwhDn0zs4R0LnUDGtOjR4+orKwsdTPMzDqUNWvWfBYRPet77bAO/crKSlavXl3qZpiZdSiSPmzoNQ/vmJklxKFvZpYQh76ZWUIO6zF9M+sY/vjHP1JdXc2ePXtK3ZSkdOnShb59+1JRUVHwMg59M2u16upqunXrRmVlJZJK3ZwkRATbt2+nurqaAQMGFLych3fMrNX27NnDiSee6MAvIkmceOKJzf525dA3szbhwC++lvS5Q9/MLCEe0zezNld5y9Ntur7Nd17SdJ3Nm7n00kvZuHFji7ezfPly5s+fz1NPPdXidbSX2otVe/To0ar1lHXot+YPr5A/MjOzWhFBRHDEEYf3AMrh3Tozs2bYt28f06ZN4/TTT2fy5Mns3r2b22+/nTPPPJMhQ4Ywffp0au8WuGnTJi644AKGDRvGyJEj+c1vfnPAulatWsWIESP44IMPqKmp4Rvf+AYjR45kxowZ9O/fn88++4zNmzczaNAgrr/+ekaOHMlHH33E7NmzGTJkCEOHDmXx4sVA7hvEpZdeWrfuWbNmcf/99wO5Pfi5c+cycuRIhg4dynvvvQfA9u3bGT9+PCNGjGDGjBm01V0OHfpmVjbef/99pk+fzltvvUX37t25++67mTVrFqtWrWLjxo384Q9/qBu6ueqqq7jhhht48803efXVV+ndu3fdel599VW+973vsWTJEk455RRuu+02zjvvPNauXcukSZPYsmXLAdu8+uqrWbduHatXr2b9+vW8+eabvPDCC8yePZtt27Y12e4ePXqwdu1aZs6cyfz58wG47bbbGDNmDOvWrWPChAkHbLM1HPpmVjb69evHOeecA8DUqVNZuXIlL730EmeddRZDhw7lxRdf5O233+aLL77g448/ZtKkSUDuIqejjz4agHfffZfp06fz5JNP8tWvfhWAlStXcuWVVwJw4YUXcvzxx9dts3///owePbqu3pQpU+jUqRO9evVi7NixrFq1qsl2f+tb3wLgjDPOYPPmzQCsWLGCqVOnAnDJJZccsM3WcOibWdk4+BRGSVx//fU88sgjbNiwgeuuu449e/Y0OlTSu3dvunTpwrp16+rKGqt/zDHHNFmvc+fO7N+/v+75wefWH3XUUQB06tSJffv2Nfh+2oJD38zKxpYtW3jttdcAePDBBxkzZgyQGz7ZtWsXjzzyCADdu3enb9++PP744wDs3buX3bt3A3Dcccfx9NNPM2fOHJYvXw7AmDFjePjhhwF47rnn2LFjR73bP/fcc1m8eDFffvklNTU1rFixglGjRtG/f3/eeecd9u7dy86dO1m2bFmT7+Xcc89l0aJFAPzyl79scJvNVdZn75hZaZTq7LdBgwbxwAMPMGPGDAYOHMjMmTPZsWMHQ4cOpbKykjPPPLOu7s9+9jNmzJjBD37wAyoqKvjFL35R91qvXr148sknueiii1i4cCFz585lypQpLF68mLFjx9K7d2+6devGrl27Dtj+pEmTeO211xg2bBiS+OEPf8hXvvIVAK644gpOP/10Bg4cyIgRI5p8L7XbHDlyJGPHjq0bamottdUR4fZQVVUVrbmJik/ZNCuOd999l0GDBpW6Ge1m7969dOrUic6dO/Paa68xc+ZM1q9fX+pmAfX3vaQ1EVFVX33v6ZuZNWHLli1cccUV7N+/nyOPPJKf/OQnpW5Sizn0zcyaMHDgwAMO7HZkPpBrZpYQh76ZWUIc+mZmCXHom5klxAdyzaztzTu2jde3s0WLffe73+XGG29k8ODBbdqcrl27HnKOfkfh0DezsnXvvfeWugmHHQ/vmFlZ+P3vf88ll1zCsGHDGDJkCIsXL2bcuHHUXuB53333ceqppzJu3Diuu+46Zs2aBcA111zD97//fc4++2xOOeWUuqkadu3axfnnn1835fGSJUtK9t7akvf0zawsPPPMM5x88sk8/XTuSvydO3eyYMECALZu3codd9zB2rVr6datG+eddx7Dhg2rW3bbtm2sXLmS9957jwkTJjB58mS6dOnCY489Rvfu3fnss88YPXo0EyZM6PD3AvaevpmVhaFDh/LCCy9w880388orr3Dssf//uMIbb7zB2LFjOeGEE6ioqODyyy8/YNmJEydyxBFHMHjwYD755BMgN2PmnDlzOP3007ngggv4+OOP617ryLynb2Zl4dRTT2XNmjUsXbqUW2+9lfHjx9e91tQcY7VTG+fXXbRoETU1NaxZs4aKigoqKysPmRK5I/KevpmVha1bt3L00UczdepUbrrpJtauXVv32qhRo3j55ZfZsWMH+/bt49FHH21yfTt37uSkk06ioqKCl156iQ8//LA9m1803tM3s7bXwlMsW2PDhg3Mnj2bI444goqKChYsWMBNN90EQJ8+fZgzZw5nnXUWJ598MoMHDz5g+Kc+V111Fd/85jepqqpi+PDhnHbaacV4G+2uyamVJfUDfgp8BdgP3BMR/yDpBGAxUAlsBq6IiB3KHeX4B+BiYDdwTUSszdY1Dfiv2ar/W0Q80Ni2PbWyWcfQEaZW3rVrF127dmXfvn1MmjSJ73znO3W3S+zImju1ciHDO/uAv4yIQcBo4AZJg4FbgGURMRBYlj0HuAgYmP1MBxZkjTgBmAucBYwC5kpqm5s+mpk1Yd68eQwfPpwhQ4YwYMAAJk6cWOomlUSTwzsRsQ3Ylj3+QtK7QB/gMmBcVu0BYDlwc1b+08h9hXhd0nGSemd1n4+IzwEkPQ9cCDzYhu/HzKxe8+fPL3UTDgvNOpArqRIYAfwK6JV9INR+MJyUVesDfJS3WHVW1lD5wduYLmm1pNU1NTXNaZ6ZldDhfBe+ctWSPi849CV1BR4F/iIiftdY1XrKopHyAwsi7omIqoio6tmzZ6HNM7MS6tKlC9u3b3fwF1FEsH37drp06dKs5Qo6e0dSBbnAXxQR/zMr/kRS74jYlg3ffJqVVwP98hbvC2zNyscdVL68Wa01s8NS3759qa6uxt/Oi6tLly707du3Wcs0GfrZ2Tj3Ae9GxH/Pe+kJYBpwZ/Z7SV75LEkPkTtouzP7YHgW+Ju8g7fjgVub1VozOyxVVFQwYMCAUjfDClDInv45wH8ANkiqvf37HHJh/7Cka4EtQO11zUvJna65idwpm98GiIjPJd0BrMrq3V57UNfMzIqjkLN3VlL/eDzA+fXUD+CGBta1EFjYnAaamVnb8TQMZmYJceibmSXEoW9mlhCHvplZQhz6ZmYJceibmSXEoW9mlhCHvplZQhz6ZmYJceibmSXEoW9mlhCHvplZQhz6ZmYJceibmSXEoW9mlhCHvplZQhz6ZmYJceibmSXEoW9mlhCHvplZQhz6ZmYJceibmSXEoW9mlhCHvplZQhz6ZmYJceibmSXEoW9mlhCHvplZQhz6ZmYJceibmSXEoW9mlhCHvplZQhz6ZmYJceibmSXEoW9mlhCHvplZQpoMfUkLJX0qaWNe2TxJH0tan/1cnPfarZI2SXpf0p/mlV+YlW2SdEvbvxUzM2tKIXv69wMX1lP+9xExPPtZCiBpMHAl8CfZMndL6iSpE/Bj4CJgMDAlq2tmZkXUuakKEbFCUmWB67sMeCgi9gK/lbQJGJW9tikiPgCQ9FBW951mt9jMzFqsNWP6syS9lQ3/HJ+V9QE+yqtTnZU1VH4ISdMlrZa0uqamphXNMzOzg7U09BcAXwOGA9uAv8vKVU/daKT80MKIeyKiKiKqevbs2cLmmZlZfZoc3qlPRHxS+1jST4CnsqfVQL+8qn2BrdnjhsrNzKxIWrSnL6l33tNJQO2ZPU8AV0o6StIAYCDwBrAKGChpgKQjyR3sfaLlzTYzs5Zock9f0oPAOKCHpGpgLjBO0nByQzSbgRkAEfG2pIfJHaDdB9wQEV9m65kFPAt0AhZGxNtt/m7MzKxRhZy9M6We4vsaqf/XwF/XU74UWNqs1pmZWZvyFblmZglx6JuZJcShb2aWEIe+mVlCHPpmZglx6JuZJcShb2aWEIe+mVlCHPpmZglx6JuZJcShb2aWEIe+mVlCHPpmZglx6JuZJcShb2aWEIe+mVlCHPpmZglx6JuZJcShb2aWEIe+mVlCHPpmZglx6JuZJcShb2aWEIe+mVlCHPpmZglx6JuZJcShb2aWEIe+mVlCHPpmZglx6JuZJcShb2aWEIe+mVlCHPpmZglx6JuZJcShb2aWEIe+mVlCHPpmZglpMvQlLZT0qaSNeWUnSHpe0q+z38dn5ZL0I0mbJL0laWTeMtOy+r+WNK193o6ZmTWmkD39+4ELDyq7BVgWEQOBZdlzgIuAgdnPdGAB5D4kgLnAWcAoYG7tB4WZmRVPk6EfESuAzw8qvgx4IHv8ADAxr/ynkfM6cJyk3sCfAs9HxOcRsQN4nkM/SMzMrJ21dEy/V0RsA8h+n5SV9wE+yqtXnZU1VH4ISdMlrZa0uqampoXNMzOz+rT1gVzVUxaNlB9aGHFPRFRFRFXPnj3btHFmZqlraeh/kg3bkP3+NCuvBvrl1esLbG2k3MzMiqilof8EUHsGzjRgSV751dlZPKOBndnwz7PAeEnHZwdwx2dlZmZWRJ2bqiDpQWAc0ENSNbmzcO4EHpZ0LbAFuDyrvhS4GNgE7Aa+DRARn0u6A1iV1bs9Ig4+OGxmZu2sydCPiCkNvHR+PXUDuKGB9SwEFjardWZm1qZ8Ra6ZWUIc+mZmCXHom5klpMkx/VRV3vJ0i5fdfOclbdgSM7O24z19M7OEOPTNzBLi0DczS4hD38wsIQ59M7OEOPTNzBLi0DczS4hD38wsIQ59M7OEOPTNzBLi0DczS4hD38wsIQ59M7OEOPTNzBLi0DczS4hD38wsIQ59M7OEOPTNzBLi0DczS4hD38wsIQ59M7OEOPTNzBLi0DczS4hD38wsIZ1L3YByVHnL0y1edvOdl7RhS8zMDuQ9fTOzhDj0zcwS4tA3M0uIQ9/MLCEOfTOzhDj0zcwS4tA3M0tIq0Jf0mZJGyStl7Q6KztB0vOSfp39Pj4rl6QfSdok6S1JI9viDZiZWeHaYk//6xExPCKqsue3AMsiYiCwLHsOcBEwMPuZDixog22bmVkztMcVuZcB47LHDwDLgZuz8p9GRACvSzpOUu+I2NYObTAzaxvzjm3Fsjvbrh1tpLV7+gE8J2mNpOlZWa/aIM9+n5SV9wE+ylu2Ois7gKTpklZLWl1TU9PK5pmZWb7W7umfExFbJZ0EPC/pvUbqqp6yOKQg4h7gHoCqqqpDXjczs5Zr1Z5+RGzNfn8KPAaMAj6R1Bsg+/1pVr0a6Je3eF9ga2u2b2ZmzdPi0Jd0jKRutY+B8cBG4AlgWlZtGrAke/wEcHV2Fs9oYKfH883Miqs1wzu9gMck1a7n5xHxjKRVwMOSrgW2AJdn9ZcCFwObgN3At1ux7bLlaZnNrD21OPQj4gNgWD3l24Hz6ykP4IaWbs/MzFrPN1Exs/LWmlMuy5CnYTAzS4hD38wsIQ59M7OEOPTNzBLi0DczS4hD38wsIT5ls4z4wi4za4r39M3MEuLQNzNLiEPfzCwhDn0zs4Q49M3MEuLQNzNLiEPfzCwhDn0zs4T44iwDfGGXWSoc+lZS/rCxgvhGKG3GoW9m1l5a82E1b2fbtSOPQ99arTV762ZWXD6Qa2aWEO/pm1lxeFz+sOA9fTOzhDj0zcwS4uEdS5JPFW0hD9F0eA5967Da4qyhzV3+vPkLzav93T6n1LU7B3fSHPpmLdSSD52kvyXYYcGhb1YKrd3b7qjfMqzkHPpmLdSqoaHW8hCNtZBD39pEiwIwT+Wen7dRS8ysMQ79w1BrArSjhmdrPzTMrDAO/XZSqhBzeJpZY8o69FPcYzYza0xZh35reI/ZzMqRp2EwM0uIQ9/MLCEOfTOzhBQ99CVdKOl9SZsk3VLs7ZuZpayooS+pE/Bj4CJgMDBF0uBitsHMLGXF3tMfBWyKiA8i4v8CDwGXFbkNZmbJKvYpm32Aj/KeVwNn5VeQNB2Ynj3dJen9AtbbA/isTVrYsbkf3Ae13A85HbcfblNrlu7f0AvFDv363kUc8CTiHuCeZq1UWh0RVa1pWDlwP7gParkfctwPhyr28E410C/veV9ga5HbYGaWrGKH/ipgoKQBko4ErgSeKHIbzMySVdThnYjYJ2kW8CzQCVgYEW+3waqbNRxUxtwP7oNa7occ98NBFBFN1zIzs7LgK3LNzBLi0DczS0iHCv2mpnCQdJSkxdnrv5JUWfxWtq8C+uBGSe9IekvSMkkNnq/bkRU6nYekyZJCUlmetldIP0i6IvubeFtS2d0oooD/E1+V9JKkddn/i4tL0c7DRkR0iB9yB35/A5wCHAm8CQw+qM71wD9lj68EFpe63SXog68DR2ePZ5ZbHxTaD1m9bsAK4HWgqtTtLtHfw0BgHXB89vykUre7BH1wDzAzezwY2FzqdpfypyPt6RcyhcNlwAPZ40eA8yW16rK2w0yTfRARL0XE7uzp6+SuhSg3hU7ncQfwQ2BPMRtXRIX0w3XAjyNiB0BEfFrkNra3QvoggO7Z42NJ/NqgjhT69U3h0KehOhGxD9gJnFiU1hVHIX2Q71rgl+3aotJosh8kjQD6RcRTxWxYkRXy93AqcKqkf5H0uqQLi9a64iikD+YBUyVVA0uB/1icph2eOtLtEpucwqHAOh1Zwe9P0lSgChjbri0qjUb7QdIRwN8D1xSrQSVSyN9DZ3JDPOPIfet7RdKQiPg/7dy2YimkD6YA90fE30n6t8DPsj7Y3/7NO/x0pD39QqZwqKsjqTO5r3KfF6V1xVHQNBaSLgD+CpgQEXuL1LZiaqofugFDgOWSNgOjgSfK8GBuof8nlkTEHyPit8D75D4EykUhfXAt8DBARLwGdCE3EVuSOlLoFzKFwxPAtOzxZODFyI7elIkm+yAb1vhncoFfbuO3tRrth4jYGRE9IqIyIirJHduYEBGrS9PcdlPI/4nHyR3cR1IPcsM9HxS1le2rkD7YApwPIGkQudCvKWorDyMdJvSzMfraKRzeBR6OiLcl3S5pQlbtPuBESZuAG4GyujNXgX1wF9AV+IWk9ZLKbm6jAvuh7BXYD88C2yW9A7wEzI6I7aVpcdsrsA/+ErhO0pvAg8A1ZbYz2CyehsHMLCEdZk/fzMxaz6FvZpYQh76ZWUIc+mZmCXHom5klxKFvJSVpnqSbsse3ZxeWtXadSyUd14z6ExqbqbOJZY+TdH1Llm0NSRMlDW7F8iVpt5WeT9m0kpI0D9gVEfPbYF0i9zddtMvrs+m7n4qIIc1YplNEfNnK7d6fbfeRFi5fSTPbbeXBe/pWdJL+Kpv//AXg3+SV3y9pcvb4zrz7AszPynpJekzSm9nP2ZIqJb0r6W5gLdBP0mZJPbLX3pN0r6SNkhZJuiCbfOzXkkZl671G0j/mteFHkl6V9EFee7oqd3+CtZI2SKqdyfFO4GvZhXB3KeeubHsbJP1Ztvy4bE73nwMb6umTKVn9jZL+Nq98V97jyVn7zgYmAHdl2/2apOWS/kfW7o15763um1T2fGMW+Ae0u3X/otaRdKQJ16wMSDqD3KXyI8j9/a0F1hxU5wRgEnBaRETeUM2PgJcjYpKkTuSuPD6e3AfHtyPi+mz5/NX9a+ByYDq5S/b/HBhDLjTnABPraWbvrM5p5C7pf4Tc9MyTIuJ32XQGr2dXO98CDImI4dm2/z0wHBhGbn6XVZJWZOsdldX97UHv92Tgb4EzgB3Ac5ImRsTj9fVhRLyabbtuTz97z8dExNmSzgUWkpt/qCEHtNvS4T19K7Z/BzwWEbsj4nccOk8KwO/Ihey9kr4F1N4f4DxgAUBEfBkRO7PyDyPi9Qa299uI2JAN+bwNLMsuwd8AVDawzOMRsT8i3gF6ZWUC/kbSW8AL5Kbv7VXPsmOAB7P2fQK8DJyZvfbGwYGfORNYHhE12bQCi4BzG2hbYx4EiIgVQPfmHNewdDj0rRQaPZCUBd8o4FFye+LPNLG+3zfyWv4so/vznu+n4W+6+cvUfm24CugJnJHtHX9CbuKugzV2056G2tnYMvl9Vd/2Gqpb+3wfB/4/b2odVuYc+lZsK4BJkv6VpG7ANw+uIKkrcGxELAX+gtxwCcAycreARFInSd0PXrYdHQt8GhF/lPR1oPbew1+Qm8q51grgz7L29SS3x/5GE+v+FTA2Ow7Ridz87y9nr30iaZBy9wiYlLfMwdsFqD1+MAbYmX0T2gyMzMpHAgMaWd4S4NC3ooqItcBiYD25PflX6qnWDXgqG0p5GfjPWfl/Ar4uaQO54wB/0v4trrMIqJK0mtxe/3sA2YyV/5IdIL0LeAx4i9y9Wl8E/ktE/O/GVhwR24Bbyc2C+SawNiKWZC/fAjyVrWtb3mIPAbOVu9n317KyHZJeBf6J3BzykOvjEyStJ/eB+b8aaLclwqdsmpUBScuBm8rwngHWxrynb2aWEO/pm5klxHv6ZmYJceibmSXEoW9mlhCHvplZQhz6ZmYJ+X9AdoD9ro3neQAAAABJRU5ErkJggg==\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "h_b = plt.hist(pred_gbc_val[y_val==0], bins=20, label='background')\n", "h_s = plt.hist(pred_gbc_val[y_val==1], bins=20, label='signal')\n", "plt.xlabel('discriminator output')\n", "plt.legend()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The next model we will train is a neural network using [tensorflow](https://www.tensorflow.org/) and [keras](https://keras.io/)." ] }, { "cell_type": "code", "execution_count": 37, "metadata": {}, "outputs": [], "source": [ "from tensorflow.keras.layers import Dense, Input\n", "from tensorflow.keras.models import Model, load_model\n", "from tensorflow.keras.optimizers import Adam\n", "from tensorflow.keras.callbacks import CSVLogger, ModelCheckpoint, EarlyStopping\n", "from tensorflow.keras.metrics import AUC" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Using the keras API, models can be defined via the [Sequential](https://keras.io/api/models/sequential/) or [functional](https://keras.io/guides/functional_api/) API.\n", "The following gives an example of a model defined via the functional API." ] }, { "cell_type": "code", "execution_count": 38, "metadata": {}, "outputs": [], "source": [ "def create_model():\n", " inputs = Input((x_train.shape[1],))\n", " x = Dense(len(input_columns),activation='tanh')(inputs)\n", " x = Dense(32, activation='tanh')(x) \n", " out = Dense(1, activation='sigmoid')(x)\n", " return Model(inputs, out)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "To build more sophisticated models you can make yourself familiar with the different layers and activation [available](https://keras.io/api/layers/) in keras." ] }, { "cell_type": "code", "execution_count": 39, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "2020-07-02 07:04:17.450259: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory\n", "2020-07-02 07:04:17.450358: E tensorflow/stream_executor/cuda/cuda_driver.cc:313] failed call to cuInit: UNKNOWN ERROR (303)\n", "2020-07-02 07:04:17.450426: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:156] kernel driver does not appear to be running on this host (8adb3809ddaf): /proc/driver/nvidia/version does not exist\n", "2020-07-02 07:04:17.451282: I tensorflow/core/platform/cpu_feature_guard.cc:143] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 AVX512F FMA\n", "2020-07-02 07:04:17.484329: I tensorflow/core/platform/profile_utils/cpu_utils.cc:102] CPU Frequency: 2200000000 Hz\n", "2020-07-02 07:04:17.490615: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x7f4e1c000b20 initialized for platform Host (this does not guarantee that XLA will be used). Devices:\n", "2020-07-02 07:04:17.490681: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version\n" ] } ], "source": [ "model = create_model()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The model is compiled using the [binary_crossentropy](https://keras.io/api/losses/probabilistic_losses/#binarycrossentropy-class) loss function and the [Adam](https://keras.io/api/optimizers/adam/) optimizer." ] }, { "cell_type": "code", "execution_count": 40, "metadata": {}, "outputs": [], "source": [ "model.compile(loss='binary_crossentropy', optimizer=Adam(), metrics=['accuracy', AUC()])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "During the training it is also useful to use some [callbacks](https://keras.io/api/callbacks/). The [CSVLogger](https://keras.io/api/callbacks/csv_logger/) will write the training and validation losses as well as the metrics specified in the compile statement into a file at each epoch. The [ModelCheckpoint](https://keras.io/api/callbacks/model_checkpoint/) callback will save the model to a file so it can be used later for inference. The [EarlyStopping](https://keras.io/api/callbacks/early_stopping/) callback will stop the training if the validation loss doesn't improve for 10 epochs." ] }, { "cell_type": "code", "execution_count": 41, "metadata": {}, "outputs": [], "source": [ "callbacks = [\n", " ModelCheckpoint(filepath='model.h5', save_best_only=True),\n", " CSVLogger('training.log'),\n", " EarlyStopping(patience=10)\n", "]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The model is trained using the [fit](https://keras.io/api/models/model_training_apis/#fit-method) method. If you want to watch the training progress, you can set verbose=1." ] }, { "cell_type": "code", "execution_count": 42, "metadata": {}, "outputs": [], "source": [ "history = model.fit(x_train, y_train, batch_size=1000, epochs=20, validation_split=0.2, sample_weight=w_train, \n", " callbacks=callbacks, verbose=0)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "After training a neural network, it is a good idea to inspect the training and validation loss values as a function of the training epochs. This can help to quickly identify problems during the training. For example an overtraining effect would show itself by an increasing validation loss and a decreasing training loss." ] }, { "cell_type": "code", "execution_count": 43, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 43, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "loss = history.history['loss']\n", "val_loss = history.history['val_loss']\n", "\n", "plt.plot(loss, label='training loss')\n", "plt.plot(val_loss, label='validation loss')\n", "plt.legend()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Using the [predict](https://keras.io/api/models/model_training_apis/#predict-method) method we can apply the trained model to the validation dataset and compute the AUC value and AMS score." ] }, { "cell_type": "code", "execution_count": 44, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "AUC: 0.7896948376943995\n", "AMS: 0.17838003952894713\n" ] } ], "source": [ "model = load_model('model.h5') #load the best model\n", "\n", "pred_val_nn = model.predict(x_val)\n", "auc = roc_auc_score(y_val, pred_val_nn, sample_weight=w_val)\n", "\n", "print('AUC:', auc)\n", "\n", "y_pred = pred_val_nn.reshape((pred_val_nn.shape[0],))\n", "\n", "bs = find_best_ams_score(y_val, y_pred, w_val)\n", "print('AMS:', bs[0][0])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**Your turn** (obligatory)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now it is time to work on improvements.\n", "Inspect all the variables in the tree and their correlations.\n", "Which of them show the most prominent differences between signal and background?\n", "Which of them are not so relevant? \n", "Summarise your ideas and collect some of the graphs for your documentation. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "You should implement code to produce some figures of the distribution of one or more highly-discriminating variables\n", "(e.g. the reconstructed Higgs boson mass `DER_mass_MMC`),\n", "after applying a cut on the classifier corresponding to the best performance.\n", "A Higgs-signal may then become visible in a single distribution." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**Be creative!** (obligatory)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "This is the most challenging, but also most rewarding\n", "part of the project. Develop, discuss, test and improve\n", "ideas to increase the performance of your classification.\n", "Consult the documentations of the different ML-libraries (https://scikit-learn.org/ or https://keras.io/) to get inspiration\n", "on options to try out, or clever combinations of\n", "variables to play with or implement your own variables. It`s really a creative process,\n", "without any further prescription ..." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**Run it on the full dataset Higgs Challenge** (voluntary)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "If you are convinced of your improved classifier,\n", "you may run it on the full data set of the challenge,\n", "prepared for you in a compatible `.root` file\n", "\n", " atlas-higgs-challenge-2014-v2.root\n", "\n", "the source of which is given on the `Higgschallenge.pdf`.\n", "\n", "It will take some time, but after some hours of invested CPU time, you will be rewarded with your personal score for the Higgs Challenge 2014!" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.9.6" } }, "nbformat": 4, "nbformat_minor": 4 }