analyzis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "0d758cc2-d687-4b9e-b328-0da4df7c9574",
   "metadata": {},
   "source": [
    "# Analize results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c1f3af14-1fec-45da-86af-0753763be02f",
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'AB-TRAP_CIC'"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from supportFiles.myFunc import loadModel, getDSName, setTarget, loadDataset, getFeatureList\n",
    "from joblib import dump, load\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "611fefce-604e-4d2d-8217-4c05f928ef54",
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Package                            Version\n",
      "---------------------------------- -------------------\n",
      "alabaster                          0.7.12\n",
      "anaconda-client                    1.7.2\n",
      "anaconda-navigator                 2.0.3\n",
      "anaconda-project                   0.9.1\n",
      "anyio                              2.2.0\n",
      "appdirs                            1.4.4\n",
      "argh                               0.26.2\n",
      "argon2-cffi                        20.1.0\n",
      "asn1crypto                         1.4.0\n",
      "astroid                            2.5\n",
      "astropy                            4.2.1\n",
      "async-generator                    1.10\n",
      "atomicwrites                       1.4.0\n",
      "attrs                              20.3.0\n",
      "autopep8                           1.5.6\n",
      "Babel                              2.9.0\n",
      "backcall                           0.2.0\n",
      "backports.functools-lru-cache      1.6.4\n",
      "backports.shutil-get-terminal-size 1.0.0\n",
      "backports.tempfile                 1.0\n",
      "backports.weakref                  1.0.post1\n",
      "bcrypt                             3.2.0\n",
      "beautifulsoup4                     4.9.3\n",
      "bitarray                           1.9.2\n",
      "bkcharts                           0.2\n",
      "black                              19.10b0\n",
      "bleach                             3.3.0\n",
      "bokeh                              2.3.2\n",
      "boto                               2.49.0\n",
      "Bottleneck                         1.3.2\n",
      "brotlipy                           0.7.0\n",
      "certifi                            2020.12.5\n",
      "cffi                               1.14.5\n",
      "chardet                            4.0.0\n",
      "click                              7.1.2\n",
      "cloudpickle                        1.6.0\n",
      "clyent                             1.2.2\n",
      "colorama                           0.4.4\n",
      "comtypes                           1.1.9\n",
      "conda                              4.10.1\n",
      "conda-build                        3.21.4\n",
      "conda-content-trust                0+unknown\n",
      "conda-package-handling             1.7.3\n",
      "conda-repo-cli                     1.0.4\n",
      "conda-token                        0.3.0\n",
      "conda-verify                       3.4.2\n",
      "contextlib2                        0.6.0.post1\n",
      "cryptography                       3.4.7\n",
      "cycler                             0.10.0\n",
      "Cython                             0.29.23\n",
      "cytoolz                            0.11.0\n",
      "dask                               2021.4.0\n",
      "decorator                          5.0.6\n",
      "defusedxml                         0.7.1\n",
      "Note: you may need to restart the kernel to use updated packages.\n",
      "diff-match-patch                   20200713\n",
      "distributed                        2021.4.0\n",
      "docutils                           0.17\n",
      "emlearn                            0.13.1\n",
      "entrypoints                        0.3\n",
      "et-xmlfile                         1.0.1\n",
      "fastcache                          1.1.0\n",
      "filelock                           3.0.12\n",
      "flake8                             3.9.0\n",
      "Flask                              1.1.2\n",
      "fsspec                             0.9.0\n",
      "future                             0.18.2\n",
      "gevent                             21.1.2\n",
      "glob2                              0.7\n",
      "graphviz                           0.18.2\n",
      "greenlet                           1.0.0\n",
      "h5py                               2.10.0\n",
      "HeapDict                           1.0.1\n",
      "html5lib                           1.1\n",
      "idna                               2.10\n",
      "imagecodecs                        2021.3.31\n",
      "imageio                            2.9.0\n",
      "imagesize                          1.2.0\n",
      "importlib-metadata                 3.10.0\n",
      "iniconfig                          1.1.1\n",
      "intervaltree                       3.1.0\n",
      "ipykernel                          5.3.4\n",
      "ipython                            7.22.0\n",
      "ipython-genutils                   0.2.0\n",
      "ipywidgets                         7.6.3\n",
      "isort                              5.8.0\n",
      "itsdangerous                       1.1.0\n",
      "jdcal                              1.4.1\n",
      "jedi                               0.17.2\n",
      "Jinja2                             2.11.3\n",
      "joblib                             1.0.1\n",
      "json5                              0.9.5\n",
      "jsonschema                         3.2.0\n",
      "jupyter                            1.0.0\n",
      "jupyter-client                     6.1.12\n",
      "jupyter-console                    6.4.0\n",
      "jupyter-core                       4.7.1\n",
      "jupyter-packaging                  0.7.12\n",
      "jupyter-server                     1.4.1\n",
      "jupyterlab                         3.0.14\n",
      "jupyterlab-pygments                0.1.2\n",
      "jupyterlab-server                  2.4.0\n",
      "jupyterlab-widgets                 1.0.0\n",
      "kamene                             0.32\n",
      "keyring                            22.3.0\n",
      "kiwisolver                         1.3.1\n",
      "lazy-object-proxy                  1.6.0\n",
      "libarchive-c                       2.9\n",
      "llvmlite                           0.36.0\n",
      "locket                             0.2.1\n",
      "lxml                               4.6.3\n",
      "MarkupSafe                         1.1.1\n",
      "matplotlib                         3.3.4\n",
      "mccabe                             0.6.1\n",
      "menuinst                           1.4.16\n",
      "mistune                            0.8.4\n",
      "mkl-fft                            1.3.0\n",
      "mkl-random                         1.2.1\n",
      "mkl-service                        2.3.0\n",
      "mock                               4.0.3\n",
      "more-itertools                     8.7.0\n",
      "mpmath                             1.2.1\n",
      "msgpack                            1.0.2\n",
      "multipledispatch                   0.6.0\n",
      "mypy-extensions                    0.4.3\n",
      "navigator-updater                  0.2.1\n",
      "nbclassic                          0.2.6\n",
      "nbclient                           0.5.3\n",
      "nbconvert                          6.0.7\n",
      "nbformat                           5.1.3\n",
      "nest-asyncio                       1.5.1\n",
      "networkx                           2.5\n",
      "nltk                               3.6.1\n",
      "nose                               1.3.7\n",
      "notebook                           6.3.0\n",
      "numba                              0.53.1\n",
      "numexpr                            2.7.3\n",
      "numpy                              1.20.1\n",
      "numpydoc                           1.1.0\n",
      "olefile                            0.46\n",
      "openpyxl                           3.0.7\n",
      "packaging                          21.3\n",
      "pandas                             1.2.4\n",
      "pandocfilters                      1.4.3\n",
      "paramiko                           2.7.2\n",
      "parso                              0.7.0\n",
      "partd                              1.2.0\n",
      "path                               15.1.2\n",
      "pathlib2                           2.3.5\n",
      "pathspec                           0.7.0\n",
      "patsy                              0.5.1\n",
      "pep8                               1.7.1\n",
      "pexpect                            4.8.0\n",
      "pickleshare                        0.7.5\n",
      "Pillow                             8.2.0\n",
      "pip                                21.0.1\n",
      "pkginfo                            1.7.0\n",
      "pluggy                             0.13.1\n",
      "ply                                3.11\n",
      "prometheus-client                  0.10.1\n",
      "prompt-toolkit                     3.0.17\n",
      "psutil                             5.8.0\n",
      "ptyprocess                         0.7.0\n",
      "py                                 1.10.0\n",
      "pybind11                           2.8.1\n",
      "pycodestyle                        2.6.0\n",
      "pycosat                            0.6.3\n",
      "pycparser                          2.20\n",
      "pycurl                             7.43.0.6\n",
      "pydocstyle                         6.0.0\n",
      "pyerfa                             1.7.3\n",
      "pyflakes                           2.2.0\n",
      "Pygments                           2.8.1\n",
      "pylint                             2.7.4\n",
      "pyls-black                         0.4.6\n",
      "pyls-spyder                        0.3.2\n",
      "PyNaCl                             1.4.0\n",
      "pyodbc                             4.0.0-unsupported\n",
      "pyOpenSSL                          20.0.1\n",
      "pyparsing                          2.4.7\n",
      "pyreadline                         2.1\n",
      "pyrsistent                         0.17.3\n",
      "PySocks                            1.7.1\n",
      "pytest                             6.2.3\n",
      "python-dateutil                    2.8.1\n",
      "python-jsonrpc-server              0.4.0\n",
      "python-language-server             0.36.2\n",
      "pytz                               2021.1\n",
      "PyWavelets                         1.1.1\n",
      "pywin32                            227\n",
      "pywin32-ctypes                     0.2.0\n",
      "pywinpty                           0.5.7\n",
      "PyYAML                             5.4.1\n",
      "pyzmq                              20.0.0\n",
      "QDarkStyle                         2.8.1\n",
      "QtAwesome                          1.0.2\n",
      "qtconsole                          5.0.3\n",
      "QtPy                               1.9.0\n",
      "regex                              2021.4.4\n",
      "requests                           2.25.1\n",
      "rope                               0.18.0\n",
      "Rtree                              0.9.7\n",
      "ruamel-yaml-conda                  0.15.100\n",
      "scapy                              2.4.5\n",
      "scapy-python3                      0.26\n",
      "scikit-image                       0.18.1\n",
      "scikit-learn                       0.24.1\n",
      "scipy                              1.6.2\n",
      "seaborn                            0.11.1\n",
      "Send2Trash                         1.5.0\n",
      "setuptools                         52.0.0.post20210125\n",
      "shap                               0.40.0\n",
      "simplegeneric                      0.8.1\n",
      "singledispatch                     0.0.0\n",
      "sip                                4.19.13\n",
      "six                                1.15.0\n",
      "slicer                             0.0.7\n",
      "sniffio                            1.2.0\n",
      "snowballstemmer                    2.1.0\n",
      "sortedcollections                  2.1.0\n",
      "sortedcontainers                   2.3.0\n",
      "soupsieve                          2.2.1\n",
      "Sphinx                             4.0.1\n",
      "sphinxcontrib-applehelp            1.0.2\n",
      "sphinxcontrib-devhelp              1.0.2\n",
      "sphinxcontrib-htmlhelp             1.0.3\n",
      "sphinxcontrib-jsmath               1.0.1\n",
      "sphinxcontrib-qthelp               1.0.3\n",
      "sphinxcontrib-serializinghtml      1.1.4\n",
      "sphinxcontrib-websupport           1.2.4\n",
      "spyder                             4.2.5\n",
      "spyder-kernels                     1.10.2\n",
      "SQLAlchemy                         1.4.7\n",
      "statsmodels                        0.12.2\n",
      "sympy                              1.8\n",
      "tables                             3.6.1\n",
      "tblib                              1.7.0\n",
      "terminado                          0.9.4\n",
      "testpath                           0.4.4\n",
      "textdistance                       4.2.1\n",
      "threadpoolctl                      2.1.0\n",
      "three-merge                        0.1.1\n",
      "tifffile                           2021.4.8\n",
      "toml                               0.10.2\n",
      "toolz                              0.11.1\n",
      "tornado                            6.1\n",
      "tqdm                               4.59.0\n",
      "traitlets                          5.0.5\n",
      "typed-ast                          1.4.2\n",
      "typing-extensions                  3.7.4.3\n",
      "ujson                              4.0.2\n",
      "unicodecsv                         0.14.1\n",
      "urllib3                            1.26.4\n",
      "watchdog                           1.0.2\n",
      "wcwidth                            0.2.5\n",
      "webencodings                       0.5.1\n",
      "Werkzeug                           1.0.1\n",
      "wheel                              0.36.2\n",
      "widgetsnbextension                 3.5.1\n",
      "win-inet-pton                      1.1.0\n",
      "win-unicode-console                0.5\n",
      "wincertstore                       0.2\n",
      "wrapt                              1.12.1\n",
      "xgboost                            1.5.0\n",
      "xlrd                               2.0.1\n",
      "XlsxWriter                         1.3.8\n",
      "xlwings                            0.23.0\n",
      "xlwt                               1.3.0\n",
      "xmltodict                          0.12.0\n",
      "yapf                               0.31.0\n",
      "zict                               2.0.0\n",
      "zipp                               3.4.1\n",
      "zope.event                         4.5.0\n",
      "zope.interface                     5.3.0\n"
     ]
    }
   ],
   "source": [
    "%pip list"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "59dfb756-bd95-42f5-afc9-0a48b3b7ea10",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e7b0e4ce-0523-44ac-ba7c-50d57e8554b4",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loading models from SCAN_AB-TRAP_CIC\n",
      "Models fetched: ['SCAN_AB-TRAP_CIC_DT.joblib', 'SCAN_AB-TRAP_CIC_LR.joblib', 'SCAN_AB-TRAP_CIC_MLP.joblib', 'SCAN_AB-TRAP_CIC_NB.joblib', 'SCAN_AB-TRAP_CIC_SVM.joblib', 'SCAN_AB-TRAP_CIC_XGB.joblib']\n",
      "DT's index of best performance: 5\n",
      "LR's index of best performance: 0\n",
      "MLP's index of best performance: 0\n",
      "NB's index of best performance: 0\n",
      "SVM's index of best performance: 0\n",
      "XGB's index of best performance: 0\n",
      "DT\n"
     ]
    }
   ],
   "source": [
    "#Datasets available are :\n",
    "# {0: 'AB-TRAP', 1: 'NB15', 2: 'CIC-IDS', 3: 'ToN-IoT', 4: 'BoT-IoT'}\n",
    "trainerDS = 0\n",
    "testerDS = 1\n",
    "modelName = \"DT\"\n",
    "# Set names\n",
    "trainerDSName = getDSName(trainerDS,1,True,True)\n",
    "testerDSName = getDSName(testerDS,1,True,True)\n",
    "\n",
    "models, prep, table, algo = loadModel(trainerDSName)\n",
    "cols = getFeatureList(0)[0:67]\n",
    "#X, y = setTarget(loadDataset(testerDS, 96, 1), testerDS, True, True, trainerDS)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec2d356f-47af-48b8-a9d5-c1c910e2ec6b",
   "metadata": {},
   "source": [
    "### DT"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7927c3ba-02c9-4bd4-8e3c-dc544a039d30",
   "metadata": {},
   "source": [
    "Get feature importance directly from model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "dea4dd4a-d36d-4896-9d8a-0f0523cbd2d7",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "reading file: \n"
     ]
    }
   ],
   "source": [
    "clf = models['DT'].best_estimator_\n",
    "relevantes = clf.feature_importances_"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "24ca2567-15de-48d2-961d-1925d50e2cfb",
   "metadata": {},
   "source": [
    "Process feature importance for plotting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "c16dceac-f1c3-4278-b852-66a7d84f88b0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "import os\n",
    "\n",
    "fi_df = pd.DataFrame({'feature_importance': relevantes, 'feature_name': cols})\n",
    "\n",
    "fi_df.sort_values(by=['feature_importance'], ascending=False, inplace=True)\n",
    "\n",
    "relevantes = fi_df[fi_df.feature_importance > 0.004]\n",
    "\n",
    "#plt.figure(figsize=(10,8))\n",
    "g=sns.barplot(x=relevantes['feature_importance'], y=relevantes['feature_name'])\n",
    "\n",
    "plt.xlabel('Importance', fontsize=20)\n",
    "plt.ylabel('Feature', fontsize=20);\n",
    "i=0\n",
    "for index, row in relevantes.iterrows():\n",
    "    g.text(row.feature_importance+0.03, i, round(row.feature_importance, 4), color='black', ha=\"center\", va=\"center\", fontsize=9)\n",
    "    i+=1\n",
    "filename = 'dissertation/feat_imp_{0}_{1}.eps'.format(trainerDSName, modelName)\n",
    "plt.savefig(filename, dpi=300, bbox_inches = \"tight\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f2a91d18-0595-4d44-91b5-917deef2a8c2",
   "metadata": {},
   "source": [
    "### LR"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c2374126-3c0d-44e8-8add-4e43ed510869",
   "metadata": {},
   "source": [
    "Get feature importance directly from model coefficient's absolute values and normalize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6f268056-0fe5-4678-8a07-cb8be175f389",
   "metadata": {},
   "outputs": [],
   "source": [
    "relevantes = list(map(abs, models[\"LR\"].best_estimator_.coef_))[0]\n",
    "relevantes = relevantes/max(relevantes)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "03f325e6-9962-4067-80e2-143eaf078f40",
   "metadata": {},
   "source": [
    "Process feature importance for plotting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "4c2ac7c9-180d-4ad7-a9d0-501ffff1dc6e",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fi_df = pd.DataFrame({'feature_importance': relevantes, 'feature_name': cols})\n",
    "\n",
    "fi_df.sort_values(by=['feature_importance'], ascending=False, inplace=True)\n",
    "\n",
    "relevantes = fi_df[fi_df.feature_importance > 0.2]\n",
    "\n",
    "#plt.figure(figsize=(10,8))\n",
    "g=sns.barplot(x=relevantes['feature_importance'], y=relevantes['feature_name'])\n",
    "\n",
    "plt.xlabel('Importance', fontsize=20)\n",
    "plt.ylabel('Feature', fontsize=20);\n",
    "i=0\n",
    "for index, row in relevantes.iterrows():\n",
    "    if row.feature_importance==1:\n",
    "        temp=0.025\n",
    "    else:\n",
    "        temp=0.05\n",
    "    g.text(row.feature_importance+temp, i, round(row.feature_importance, 4), color='black', ha=\"center\", va=\"center\", fontsize=9)\n",
    "    i+=1\n",
    "filename = 'dissertation/feat_imp_{0}_{1}.eps'.format(trainerDSName, \"LR\")\n",
    "plt.savefig(filename, dpi=300, bbox_inches = \"tight\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aeffce10-94c6-4e95-bd36-f9b3e02105dd",
   "metadata": {},
   "source": [
    "### MLP"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2e5d8a6c-c14a-4d7a-bf81-e2f3c4d59647",
   "metadata": {},
   "source": [
    "Get feature importance directly from model coefficient's by adding absolute values for every path weight\n",
    "\n",
    "Weight is calculated by multiplying path's coefficients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "30c261f5-c846-4522-902e-6ba46b23c1e9",
   "metadata": {},
   "outputs": [],
   "source": [
    "models[\"MLP\"].best_estimator_.coefs_[1]\n",
    "hocuspocus = lambda a,b: [[r*q for r,q in zip(p,b)] for p in a]\n",
    "c = hocuspocus(models[\"MLP\"].best_estimator_.coefs_[0],models[\"MLP\"].best_estimator_.coefs_[1])\n",
    "relevantes = []\n",
    "for entry in c:\n",
    "    relevantes.append(sum(list(map(abs, entry))[0]) )\n",
    "    \n",
    "relevantes = relevantes/max(relevantes)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3e5490e0-26e2-4d86-b5c7-9085532db001",
   "metadata": {},
   "source": [
    "Process feature importance for plotting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "ae514c22-f6b8-4d6f-8aa0-a785fbe5daef",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fi_df = pd.DataFrame({'feature_importance': relevantes, 'feature_name': cols})\n",
    "\n",
    "fi_df.sort_values(by=['feature_importance'], ascending=False, inplace=True)\n",
    "\n",
    "relevantes = fi_df[fi_df.feature_importance > 0.3]\n",
    "\n",
    "#plt.figure(figsize=(10,8))\n",
    "g=sns.barplot(x=relevantes['feature_importance'], y=relevantes['feature_name'])\n",
    "\n",
    "plt.xlabel('Importance', fontsize=20)\n",
    "plt.ylabel('Feature', fontsize=20);\n",
    "i=0\n",
    "for index, row in relevantes.iterrows():\n",
    "    if row.feature_importance==1:\n",
    "        temp=0.025\n",
    "    else:\n",
    "        temp=0.05\n",
    "    g.text(row.feature_importance+temp, i, round(row.feature_importance, 4), color='black', ha=\"center\", va=\"center\", fontsize=9)\n",
    "    i+=1\n",
    "filename = 'dissertation/feat_imp_{0}_{1}.eps'.format(trainerDSName, \"MLP\")\n",
    "plt.savefig(filename, dpi=300, bbox_inches = \"tight\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "26d78c7f-5b08-478b-8e5c-f50dda06e022",
   "metadata": {},
   "source": [
    "### SVM"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "83845574-7d18-4827-9ca1-a43e17c0a2c7",
   "metadata": {},
   "source": [
    "Get feature importance directly from model coefficient's absolute values and normalize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0e2b0e15-f353-41a9-b9bb-01adb9fdcd3c",
   "metadata": {},
   "outputs": [],
   "source": [
    "relevantes = list(map(abs, models[\"SVM\"].best_estimator_.coef_[0]))\n",
    "relevantes = relevantes/max(relevantes)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a18f6e97-46c7-4d53-be5d-d3c187a7ca93",
   "metadata": {},
   "source": [
    "Process feature importance for plotting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "1c1dc087-24bc-4155-abb4-0960e989df12",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fi_df = pd.DataFrame({'feature_importance': relevantes, 'feature_name': cols})\n",
    "\n",
    "fi_df.sort_values(by=['feature_importance'], ascending=False, inplace=True)\n",
    "\n",
    "relevantes = fi_df[fi_df.feature_importance > 0.1]\n",
    "\n",
    "#plt.figure(figsize=(10,8))\n",
    "g=sns.barplot(x=relevantes['feature_importance'], y=relevantes['feature_name'])\n",
    "\n",
    "plt.xlabel('Importance', fontsize=20)\n",
    "plt.ylabel('Feature', fontsize=20);\n",
    "i=0\n",
    "for index, row in relevantes.iterrows():\n",
    "    if row.feature_importance==1:\n",
    "        temp=0.025\n",
    "    else:\n",
    "        temp=0.05\n",
    "    g.text(row.feature_importance+temp, i, round(row.feature_importance, 4), color='black', ha=\"center\", va=\"center\", fontsize=9)\n",
    "    i+=1\n",
    "filename = 'dissertation/feat_imp_{0}_{1}.eps'.format(trainerDSName, \"SVM\")\n",
    "plt.savefig(filename, dpi=300, bbox_inches = \"tight\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ecdd72c2-9e12-4227-b708-e73a110e57da",
   "metadata": {},
   "source": [
    "### SVM by prediction"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "297a2f44-8f48-4aa6-9edf-2ba7c8fe66e6",
   "metadata": {},
   "source": [
    "Get feature importance directly from model's predictions normalized absolute values\n",
    "\n",
    "Predictions are made with each feature set to one while others are set to zero"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "efebcfdd-c470-4170-b791-0f560a4d4b24",
   "metadata": {},
   "outputs": [],
   "source": [
    "relevantes = list(map(abs, models[\"SVM\"].best_estimator_.decision_function(np.identity(67))))\n",
    "relevantes = relevantes/max(relevantes)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5ac4bd00-9086-4c8e-8063-89261d9d7eb0",
   "metadata": {},
   "source": [
    "Process feature importance for plotting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "fb059cc0-d9dd-4d35-9343-808a6260e6cc",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fi_df = pd.DataFrame({'feature_importance': relevantes, 'feature_name': cols})\n",
    "\n",
    "fi_df.sort_values(by=['feature_importance'], ascending=False, inplace=True)\n",
    "\n",
    "relevantes = fi_df[fi_df.feature_importance > 0.2]\n",
    "\n",
    "#plt.figure(figsize=(10,8))\n",
    "g=sns.barplot(x=relevantes['feature_importance'], y=relevantes['feature_name'])\n",
    "\n",
    "plt.xlabel('Importance', fontsize=20)\n",
    "plt.ylabel('Feature', fontsize=20);\n",
    "i=0\n",
    "for index, row in relevantes.iterrows():\n",
    "    if row.feature_importance==1:\n",
    "        temp=0.025\n",
    "    else:\n",
    "        temp=0.05\n",
    "    g.text(row.feature_importance+temp, i, round(row.feature_importance, 4), color='black', ha=\"center\", va=\"center\", fontsize=9)\n",
    "    i+=1\n",
    "filename = 'dissertation/feat_imp_{0}_{1}.eps'.format(trainerDSName, \"SVM\")\n",
    "plt.savefig(filename, dpi=300, bbox_inches = \"tight\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cee73526-5122-4f1c-a2c2-dd1b8af90b0c",
   "metadata": {},
   "source": [
    "### XGB"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bac0d466-896f-4cf4-b566-a8484582dd29",
   "metadata": {},
   "source": [
    "Get feature importance directly from model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2868f763-c3e1-46bc-9d94-ecb18914da7c",
   "metadata": {},
   "outputs": [],
   "source": [
    "relevantes = list(map(abs, models[\"XGB\"].best_estimator_.feature_importances_))\n",
    "relevantes = relevantes/max(relevantes)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "880afc13-ee00-4f6a-9c90-b6f89406006c",
   "metadata": {},
   "source": [
    "Process feature importance for plotting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "286e23fb-d14e-4503-a707-5f044a66465d",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdIAAAEPCAYAAAD2wEXHAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAABEEklEQVR4nO3de3hVxb3/8fdHJNzNRa0VPFwUBEEEayAVCInQqii0eGq1GEWMohYrtadF7WmrYK03zrGCbS1oNSKi/uqlWo0giiEENYgYCOIdKYieqoREEIMRvr8/1kpMNrnv3PN9Pc9+stfMrFkzG558M7NnrZGZ4Zxzzrn6Oai5G+Ccc861Zh5InXPOuSh4IHXOOeei4IHUOeeci4IHUueccy4KBzd3A1zTO+yww6xv377N3QznnGtVXnvttc/M7PDIdA+k7VDfvn1Zu3ZtczfDOedaFUn/qizdp3adc865KPiItB1688MdnDRrUXM3wznnmtRrc6c2Sr0+InXOOeei4IHUOeeci4IHUueccy4KHkidc865KHggdc45166cdtppHH744dx4440H5BUXF5OWlkZycjJpaWkUFxfXWF+7DaSS+kra2Bx1VHWepOGSzoimTc4556r3t7/9jblz51aal5GRwaBBg1i1ahUDBw4kIyOjxvrabSBtoYYDHkidc64RHXXUUVXmZWVlMXHiRAAmTZpEdnZ2jfW190B6sKT7JW2Q9KiksZIeB5D0Q0lfSoqR1FnS5jD9JEnrJb0MXFFd5ZKmSXpS0lJJb0u6vpIyR0t6XVIScANwrqQ8SedKSgnf54VlelRxnSMlZYflNkpKrqTMpZLWSlr79Z5d9fionHOu7SsoKCA+Ph6AuLg4duzYUeM57f2BDAOBi81staR7gZHAiWFeMrARGEHwOeWG6fcBV5rZSkmVzw1UNBI4HtgDvCrpGeAzAEkDgYeBi8wsT9J1QKKZ/SzM/ydwRdi+7kBVk/XnAcvM7A+SOgBdIwuY2UJgIUC3b/ezWrTbOefanYSEBAoLCwEoKioiISGhxnPa+4h0m5mtDt8vBsYA70k6jiAA3g6MJQiqqyTFAnFmtjI854FaXGO5me0wsy+Bx8NrABwOPAmcb2Z5VZy7Grhd0szwul9XUe5V4CJJs4GhZuZDTuecq4eUlBQyMzMByMzMJCUlpcZz2nsgjRyZGbAKmACUAM8TBL4xQDagSs6pzzUAioBtwOgqTzS7BbgE6AK8ImlQFeWyCQL+duABSY3zHCznnGsDpk+fzty5c8nIyGDy5Mnk5eWVLT6aNm0a+fn5JCcnk5+fz7Rp02qsr71P7faWdLKZvQxMAXKA14BFwCIz+1TSocC3gTfMzCQVSRpjZjlAWi2u8X1JCcCXwGQgPUz/KjxeJmm3mS0BdgFl34NKOsbM8oF8SScDg4C3Ii8gqQ+w3czultQN+E7YB+eccxHuvvvuA9KGDx8OQJcuXXjooYfqVF97D6RvAhdKWgC8C9xFMGI8gmAECrAB+MTMSkeSFwH3StoDLKvFNXIIpoD7A0vMbK2kvgBm9oWkicBySV8ALwLXSsoDbgbGSDoF2AdsAp6t4hqpwCxJJcBuwEekzjnXRPRNfHANTdI0yi0eaim6fbufDbpgTnM3wznnmlS0u79Ies3MEiPT2/t3pM4551xU2vvUboOQdBpwa0TyB2Z2FpDRgNcZyoErhfeaWVJDXcM551zd+NRuO5SYmGhr165t7mY451yr4lO7zjnnXCPwQOqcc85FwQOpc845FwVfbNQOffXxG2y9YWhzN8M1gd7X5Td3E5xr83xE6pxzzkXBA6lzzjkXBQ+kzjnnXBQ8kDrnnHNR8EDqXBuXkZHBqFGjGD16NOvWrauQ9/DDDzNmzBjGjh3LxIkT+fzzzwF49tlnGTFiBMnJyaSlpfH118FWuP369SM1NZXU1FT+8Ic/ALBlyxbi4+PL0p955pmm7aBzzcyfbNQOndCriz19Wf/mboZrAj2uzGb8+PG88sorbN++nQsuuICcnJyy/K+++oqYmBgArrvuOo444giuuOIKEhMTeeyxx+jTpw/Tpk3j3HPPZcKECfTv35/33nuvwjW2bNnCJZdcwvPPP9+kfXOuqfmTjSJI2iLpsIi0OEkzajivr6SNjds65xpGbm4uycnJxMTE0K9fP3bv3s3evXvL8kuDKMCePXsYMmQIAEOGDKGwsBAzo6ioiMMPPxwAM+OUU07h9NNPJy8vr+zc9evXk5yczNSpU9mxY0fTdM65FqLdBtIqxAHVBlLnWpOCggLi4+PLjmNjYykoKKhQ5m9/+xtDhw4lOzu7LJBOnTqV008/nUGDBtGxY0cSE4M/wnNzc3nxxRe59dZbSUsL9rU/8sgj2bx5M6tWrWLMmDFcffXVTdQ751qGNh9IwxHkW5Lul7RB0qOSupbL7yJpqaTpwC3AMZLyJM2tRd0dJM2V9GpY92VheqqkrPBab0l6UJKqqWeLpJskvSxpraTvSFom6X1Jl4dlukt6QdI6SfmSfhimjwiv3VlSN0lvSDq+kmtcGta9tuCLfXX/IF2rlJCQQGFhYdlxUVERCQkJFcpcfPHF5Ofnc/bZZzN3bvDf/rLLLmPNmjW8/fbbJCQk8Pe//x2Aww4LJnGGDRtGt27d2LlzJ506daJHjx4AnH/++fiGCK69afOBNDQQWGhmJwCf882oszvwT2CJmd0NXAu8b2bDzWxWLeq9GCgysxHACGC6pH5h3onAVcBg4GhgdA11bTOzk4FVBFuvnQ18F7ghzC8GzjKz7wCnAP8rSWb2KvAUcCNwG7DYzA6YejazhWaWaGaJCd061KJrri1ISkoiJyeHkpIStm7dSvfu3enUqVNZfnFxcdn7uLg4unYN/sbs0KFD2Uj28MMPp6CggL1795aV3759O4WFhcTFxVFUVFRWx4oVKxg4cGBTdM25FqO9PCJwm5mtDt8vBmaG758EbjOzB+tZ76nACZLODo9jgQHAV8AaM/sQQFIe0BfIqaSOUk+FP/OB7ma2C9glqVhSHPAFcJOkscB+oBdwBPB/BMH2VYJgOzOyYtd+xcfHM2PGDFJSUpDEvHnzyMvLY/ny5cyaNYu5c+fywgsvAMHo9d577wXgxhtvZNy4cXTu3Jm4uDiuueYaPvnkE374wx/SrVs39u3bx4IFC5DEiy++yA033ECPHj3o3Lkzd999d3N22bkm1+ZX7UrqC6w0sz7h8TjgSoIRYyZwCHCBmVlY9mkzO2BqNKK+p83seEmPEYx0l0WUSQV+ZWYTw+M/AWvNLKOKOrcAiWb2maRp4fuflc8DJgITgPPNrCRMTzWzLZK+DawG9gIjzOyL6j4TX7Xbfvizdp1rOO191W5vSSeH76fwzcjwOmAH8JfweBfQow71LgN+KqkjgKRjJXVrgPZWJhb4JAyipwB9yuUtBH4HPAjc2kjXd845V4n2EkjfBC6UtAFIAO4ql3cV0FnSbWa2A1gtaWNtFhsB9wCbgHXhLTELaLzp8geBRElrgTTgLQBJU4GvzWwJwWKpEeGo2znnXBNoL1O71U7Xtjc+tdt++NSucw2nvU/tOuecc42iza/aNbMtQJ1Ho5KGAg9EJO81s6T6tkXSE0C/iORrIhcrOeecaz3a/NSuO1BiYqL5TfPOOVc3PrXrnHPONQIPpM4551wUPJA655xzUWjzi43cgd765C1G31nTo3/bh9VXrq65kHPOVcNHpM4551wUPJA655xzUfBA6pxzzkXBA6lzzjkXBQ+kzgEZGRmMGjWK0aNHs27dugp5t912G0lJSYwePZorr7yS0oeYbNmyhXHjxjF69GhuuummsvILFiwgKSmJ5OTksr0+zYzp06czduxYTjvtNLZt29Z0nXPONaoWGUgl9Q13U2m0OiRNC/cJbVAN0fZyde1uiHpc9Xbu3Mn8+fPJyspi8eLFzJxZcW/0s846i9zcXFavXs2///1vVqxYAcC1117LnDlzWL16NStWrOCtt97ik08+YcGCBeTk5JCZmck111zDvn37ePLJJ+nQoQPZ2dnccMMNXHvttc3RVedcI2iRgbQ9kdShudvQ3uXm5pKcnExMTAz9+vVj9+7d7N27tyx/wIABZe9jYmI4+ODgrrG8vDySk5MBOPPMM8nOzmbLli0MHjyYjh070qNHD7p168b777/PO++8Q2Ji8GSxESNGsHLlyibsoXOuMbXkQHqwpPslbZD0qKSxkh4HkPRDSV9KipHUWdLmMP0kSeslvQxcUYtr9JS0VNK7km4rTZR0qqSXJa2T9HdJ3cP06yS9Gu5XulCSqruupA6S5obnbJB0WZieKulFSUuAWu1zJWlWuXrmhGl9Jb0p6W5Jb0h6TlKX2tTnvlFQUEB8fHzZcWxsLAUFBQeUy8rK4uOPP2bs2LEA7N+/vywvLi6OHTt20L9/f/Ly8vj888/Zvn0769evp6CggKFDh7Js2TLMjGeffZbPPvus8TvmnGsSLTmQDgQWmtkJwOfASODEMC8Z2AiMAJKA3DD9PmCmmZ1cy2sMB84FhgLnSvoPSYcBvwW+Z2bfAdYC/xWW/5OZjQj3Nu0CTKzhuhcDRWY2ImzrdEmlu7+MBH5jZoNraqSkU4EB4TnDgZMkjQ2zBwB/NrMhQCHwoyrquFTSWklrS3aX1HTJdiUhIYHCwsKy46KiIhISEiqU2bBhA7/+9a955JFHCP9+4qCDDjrgnISEBObMmcOkSZP4xS9+wbBhw+jZsycTJkxg8ODBpKam8vzzzzNkyJAm6ZtzrvG15EC6zcxKHzuzGBgDvCfpOIKAcjswliCorpIUC8SZWemcWeQWaJV5wcyKzKwY2AT0Ab4LDAZWS8oDLgzTAU6RlCspHxgHDKnhuqcCU8N6coFDCQIfwBoz+6CWn8Wp4et1YB0wqFw9H5hZXvj+NaBvZRWY2UIzSzSzxI7dO9bysu1DUlISOTk5lJSUsHXrVrp3706nTp3K8t977z3S09N5+OGHOeyww8rShw0bxksvvQTAs88+WzZS/dGPfsTKlSuZN28eXbt2pXfv3gDMmTOHlStXMmHCBMaPH9+EPXTONaaW/IjAyP3dDFgFTABKgOeBDKAD8CtAlZxTk73l3u8j+DwELDezKeULSuoM/AVINLNtkmYDnWu4roArI/cblZQKfFGHdgq42cwWRNTTt5I++NRuHcXHxzNjxgxSUlKQxLx588jLy2P58uXMmjWLq666isLCQi688EIAZs2axZlnnsnNN9/MxRdfzFdffcWECRM47rjjAJg6dSrbtm2ja9eu3HnnnUCwoOmss86iQ4cO9OnTpyzdOdf6teRA2lvSyWb2MjAFyCEYcS0CFpnZp5IOBb4NvGFmJqlI0hgzywHS6nndV4A/S+pvZu9J6gocBXwS5n8Wfmd6NvComRVWc91lwE8lrTCzEknHAtvr0aZlwO8lPWhmuyX1IvhjwjWQ9PR00tPTK6QNHz4cgKeffrrSc44++mhefPHFA9IXLVp0QFp8fDxZWVlRt9M51/K05ED6JnChpAXAu8BdBCO/I4DssMwG4BP7Znfyi4B7Je0hCD51FgboacBDkkrn935rZu9IuptgcdAW4NVyp1V13XsIplrXhQuTPgUm16NNz4VT2i+H38/tBs4nGIE655xrRvomBrn2onvv7jZs1rDmbkaL4Lu/OOdqS9JrZpYYmd6SFxs555xzLV5LntptEJJOA26NSP7AzM5qjvZECr/nfaGSrPFmtqOp2+Occ65ufGq3HUpMTLS1a9c2dzOcc65V8ald55xzrhF4IHXOOeei4IHUOeeci4IHUueccy4KbX7VrjvQrrffZuXYlOZuRoNLyfatyZxzTc9HpM4551wUPJA655xzUfBA6pxzzkXBA6lzzjkXBQ+krk3JyMhg1KhRjB49mnXr1lXIe+mllxg6dCidO3fmww8/LEvfs2cPl1xyCePHjyc1NZWdO3cCcPPNNzNixAhGjhzJ7bffDsCKFStITU0lNTWVE088kZNOOqnpOueca5Fa5KpdSXHAeWb2l2rK9AVGmdmS8DgV+JWZTWzAduw2s+7NVU9l50X2231j586dzJ8/n1deeYXt27dzwQUXkJOTU5Y/ZMgQXn75ZSZOrPhfZM6cOZxzzjmceuqpZWm7du3i3nvv5a233sLMGDx4MJdddhnjxo1j3LhxANx2223s37+/aTrnnGuxWuqINA6YUUOZvsB5jd6Slqcv7bPfNcrNzSU5OZmYmBj69evH7t272bt3b1l+bGws3bsf+PfMCy+8wNKlS0lNTeX6668HoEuXLvTs2ZMvv/ySL7/8ki5dutCxY8cK5y1ZsoTzzvN/Cufau3oHUkmTJD0sab2k98qlHyfpakm9omjXLcAxkvIkzQ1fGyXlSzq3XJnksMwvItrWTdK9kl6V9LqkH4bp0yQ9LmmppHcl3VaLfv6vpHWSXpB0uKRvSXotzBsmyST1Do/fl9RVUj9JL4fX/30N9adKypb0hKRNkv4q6aCIMoeF9Z0Z2W9JQyStCY83SBpQxXUulbRW0tqikpKaut0qFRQUEB8fX3YcGxtLQUFBjefl5+czbtw4XnzxRTZt2sTSpUs5+OCDOeOMMxg4cCDHHnssl156KTExMRXOiY2NpXfv3o3SF+dc61HnQKrA/cA/gB8DxwD9yhXZCdwEnB9Fu64F3jez4cArwHBgGPA9YK6kI8Myq8xsuJn9MeL83wArzGwEcEp4TrcwbzhwLjAUOFfSf1TTjm7AOjP7DrASuN7MPgE6SzoESAbWEgS2PsAnZrYHmAfcFV7//2rR35HAL8M2HQP8Z2mGpCOAZ4DrzOyZSvp9OTAv/KwSgQ+phJktNLNEM0uMjRhZtRUJCQkUFhaWHRcVFZGQkFCr804//XQkcdppp7FhwwbeeecdHnvsMTZv3szmzZu5//772b59e9k5DzzwAOefH81/cedcW1GfEekM4ALgPiAB+J/ymWb2f8Bq4MyoWxcYAzxkZvvM7N8EAW1EDeecClwrKQ/IAjoDpUOHF8ysyMyKgU1An2rq2Q88Er5fHLYF4CVgNDCW4I+GsQRBdVWYPxp4KHz/QA1tBVhjZpvNbF94Xul1OhLsVXq1mS2v4tyXgf+WdA3Qx8y+rMX12qSkpCRycnIoKSlh69atdO/enU6dOtV4XmpqKqXbyq1du5b+/ftjZvTo0YNOnTrRpUsXOnXqxO7duwHYv38/jz/+OD/+8Y8btT/OudahPoH0YmA9MN3MioDKNjR9l4qj1Gionuf8KBy1DTez3mb2Zpi3t1y5fdRtwVVpX1cRBM4+wJMEo+UxQHYlZetSb+Tx18BrwGlVnhgsOvoB8CWwTNK4Oly3TYmPj2fGjBmkpKQwZcoU7rjjDvLy8pg7dy4A77zzDt/73vdYv349U6ZM4a677gLglltu4brrrmPs2LGUlJQwefJkBg4cyMiRI/nud79LUlISo0ePZuDAgQBkZWUxbNgw4uLimqurzrkWpD6rdgcCC6z6HcE/AQ6vX5MA2AX0CN9nA5eF08kJBKO/WUCvcmUiLQOulHSlmZmkE83s9Xq04yDgbOBhggU+pUtAs4EbgWwz2y+pADgD+HWYvxr4CcEoNq0W1xkpqR/wL4Jp54VhugHpwN8lXWtmt1Dxs0HS0cBmM5sfvj8BWFGPvrYJ6enppKenV0gbPnw4AMceeyzPP//8Aef06dOH55577oD0m2++udJrlF+565xz9QmkXxNMlVanF7C7HnUDYGY7JK2WtBF4FthAMAo2gmnO/5O0A/ha0nogAygfKH8P3AFskCRgC1Cf22K+AIaEi4uKCIIcZrYlqLZsBJoDHGVmO8PjnwNLJP0ceKwW13mZYBHR0LDOJ0ozzGyfpJ8A/5T0OXA3FfvdGThfUgnB97E31KOfzjnn6knVDywrOUF6GYgFhoSjvesJFsJ0CPM7E0ztvmVm32/oBrc1jXH/a00G9uhhC0/8TlNdrsn47i/OucYk6TUzS4xMr893pA8Ag4A/VnKbRgfgdqAnwWjJOeeca9PqM7W7gGBxy0yC2192AUh6FPguQRB90swebKhGNjZJuUDk8s4LzCy/Aa8xlANX8O41sySClcXOOedaoTpP7QJIOhj4LXAFcGi5rELgTuD3ZvZ1QzTQNbzExEQrvd3DOedc7VQ1tVuvZ+2GQXK2pDnAsQTBtIjge9F9UbXUOeeca0XqHEgl7QMeMbPzwltg3m74ZjnnnHOtQ30WG+0iuN/ROeeca/fqE0hfBwY3dEOcc8651qg+35HeSvBwgO9X8/xX14J98mERf/rlP5v8uj/730lNfk3nnGts9Qmk3wKWAs9K+gfwKsETdQ5Y/mtmi6JqnXPOOdfC1SeQZhAETRFs91W65Vf5QKrw2AOpc865Nq0+gfSiBm+Fc84510rVOZCa2f2N0RDnnHOuNarPql3XACT9dyPUOVvSrxq63saUkZHBqFGjGD16NOvWrauQV1xcTFpaGsnJyaSlpVFcXAzAyy+/zKhRoxg7dmzZXqMAmzdvZtKkSYwbN46pU6c2aT+cc+2XB9JGFD7EvyoNHkhbm507dzJ//nyysrJYvHgxM2fOrJCfkZHBoEGDWLVqFQMHDiQjIwOAmTNn8vDDD5OdnU1WVhbvvPMOAD/72c9YuHAhK1asYNEi/3reOdc06hxIJW2u5ev9xmhwSyGpr6S3JN0vaYOkRyV1lbRF0nWScoAfS5oiKV/SRkm3hufeAnSRlCfpwTDtv8IyGyVdVe46U8P610t6IEzrI+mFMP0FSb2b4SOIWm5uLsnJycTExNCvXz92797N3r17y/KzsrKYODHYXW7SpElkZwfbvxYVFdG7d9DlxMREsrKy+Ne//sWePXv4+c9/TmpqKo89VpttYJ1zLnr1WWx0EJXc6kKwR2lc+P4joKSebWpNBgIXm9lqSfcCM8L0YjMbI6kn8ApwErATeE7SZDO7VtLPzGw4gKSTCBZxJRGseM6VtBL4CvgNMNrMPpOUENb/J2CRmd0vKR2YD0yurqGSLgUuBYjvcXgDdT86BQUFxMfHlx3HxsZSUFDAkUceeUB+XFwcO3bsAOCwww5j/fr1HHfccTz//PNMmjSJjz76iNdff51NmzbRo0cPRo0axbhx4yrU75xzjaE+i436VpUnqT/BL/VuwGn1b1arsc3MVofvFxNsLQfwSPhzBJBlZp8ChKPPscA/IuoZAzxhZl+E5R4Hkgn+YHnUzD4DMLOCsPzJfHPb0QPAbTU11MwWAgsBen97QN23/GkECQkJFBYWlh0XFRWRkJBQaX75vLvvvptf/vKXmBlHH300PXv2JCEhgaFDh9KrVy8Ahg8fzrvvvsvIkSObrD/OufapQb8jNbP3CH7B9wKub8i6W6jIgFR6/EX4U7Wsp6pypffj1rUdrUJSUhI5OTmUlJSwdetWunfvTqdO32wLm5KSQmZmJgCZmZmkpKQAMGTIEJYuXco///lPCgoKmDBhAv3792fPnj3s2rWLr7/+mk2bNtGnT59m6Zdzrn1p8MVGZlYMLAemNHTdLVBvSSeH76cAORH5uUCKpMPChUdTgJVhXomkjuH7bGBy+B1rN+AsYBXwAnCOpEMByk3tvgT8JHyfVsl1W4X4+HhmzJhBSkoKU6ZM4Y477iAvL69sJe60adPIz88nOTmZ/Px8pk2bBsDtt9/OKaecwmmnncYVV1zB4YcfTocOHbj11luZMGECo0ePZvr06RxxxBHN2DvnXHtRr429a6xUuhOYbmadG7zyFkJSXyCTIAiOAt4FLgA2AYml07GSzgN+TTC6zDSzq8P0W4EfAOvMLE3SfwHpYfX3mNkdYbkLgVnAPuB1M5sWXvte4DDgU+AiM9sqaTaw28z+p7q29/72ALs67faG+BjqxJ+165xrzara2LvBA6mkw4A84EszG9CglbcgYTB72syOb+621JUHUuecq7uqAml9Nva+roqsg4H/AH5IsIL313Wt2znnnGtt6nP7y+wa8j8HbjSzGleStmZmtgVodaNR55xzDas+gfSUKtL3E9wr+ZaZfV3/JjnnnHOtR6MsNnItW2Jioq1du7a5m+Gcc61KVd+R1ucRgddJGltDmeRqvkt1zjnn2oz63Ec6G0itocxY2scDGZxzzrVzjbX7y8EE35k655xzbVp9FhvVxknAZ41Ut4vSxx+8zx/OP7vB6vvN4kcbrC7nnGttahVIJa2ISJomKbWSoh0I7iXtAzwUVcucc865VqC2I9LUcu8N6Bu+Iu0HdhDsfvKLKNrlnHPOtQq1CqRmVvZdqqT9wGwzu6HRWuWcc861EvX5jvQi4PWGbohzzjnXGtV51a6Z3W9mGxqjMa51y8jIYNSoUYwePZp169ZVyCsuLiYtLY3k5GTS0tIoLi4GYM+ePVxyySWMHz+e1NRUdu7cWXZOQUEB8fHxLF68GIAVK1aQmppKamoqJ554IieddFLTdc4556oQ1apdSUcRbOLdqbJ8M8uOpn7XeuzcuZP58+fzyiuvsH37di644AJycr7ZJjUjI4NBgwbx4IMPcsMNN5CRkcHll1/OnDlzOOecczj11FMPqPPmm29m9OjRZcfjxo1j3LhxANx2223s3+93WDnnml+97iOVdKqkN4B/EWwy/WIVr1ZH0pZwK7jyaXGSZtRwXl9JGxu3dSBpsqTBzdmGyuTm5pKcnExMTAz9+vVj9+7d7N27tyw/KyuLiRMnAjBp0iSys4O/sV544QWWLl1Kamoq11//zTM8tm7dyscff0xi4gFP4wJgyZIlnHfeeY3YI+ecq536PCIwCXgaiAP+RLBhdTZwN/BWePxPoC0tRooDqg2kTWgyUGkgbU6l07ClYmNjKSgoqDQ/Li6OHTt2AJCfn8+4ceN48cUX2bRpE0uXLgVgzpw5/OY3v6n0Wvn5+cTGxtK7d+/G6o5zztVafUak/w0UAyPM7Odh2otmdjnBtmK/B74HtOi79MPR21uS7pe0QdKjkrqWy+8iaamk6cAtwDGS8iTNrUXdHSTNlfRqWPdlYXqqpKzwWm9JelCSqqnnFkmbwjr+R9Io4AfA3LAtx0g6SdJ6SS8DV1RT16WS1kpa+0Xx3qqK1VtCQgKFhYVlx0VFRSQkJFSaXz4vISGB008/HUmcdtppbNiwgfz8fCRx3HHHVXqtBx54gPPPP7/B++Ccc/VRn0B6MvCUmX0UWY8FrgfeBOY0QPsa20BgoZmdQLCPaumoszvBqHqJmd0NXAu8b2bDzWxWLeq9GCgysxHACGC6pH5h3onAVQSjyqOB0ZVVICkBOAsYErbvRjN7CXgKmBW25X3gPmCmmZ1cXYPMbKGZJZpZYrfOlX6lHZWkpCRycnIoKSlh69atdO/enU6dvrlOSkoKmZmZAGRmZpKSkgJAamoqpTvRrF27lv79+/Paa6/x9ttvc/rpp7N48WLmzp3LmjVrANi/fz+PP/44P/7xjxu8D845Vx/1WWwUC2wtd/wV0C2izGqgNXyBtc3MVofvFwMzw/dPAreZ2YP1rPdU4ARJpc/hiwUGEHxWa8zsQwBJeQQPtsippI7PCUb+90h6hmA6vQJJsUCcma0Mkx4AJtSzzVGJj49nxowZpKSkIIl58+aRl5fH8uXLmTVrFtOmTSM9PZ3k5GSOOuoo7rvvPgBuueUWpk+fTnFxMQMGDGDy5MkcdNBBTJs2DYDZs2fTv39/Ro4cCQTftQ4bNoy4uLjm6KZzzh2gPoH0EyA+4viYiDIdgS71bVQTityMtfR4NTBB0hKr34atAq40s2UVEoPHKpafV91HFf8GZva1pJHAeOAnwM+AcZVcp8VsKJuenk56enqFtOHDhwPQpUsXHnrowKdG9unTh+eee67KOmfPnl3huPzKXeecawnqM7X7DhUD5yvA9yUdCyDp28CPgHejb16j6y2pdEp0Ct+MDK8jeNThX8LjXUCPOtS7DPippI4Ako6VFDlqr5ak7kCsmWUSTAUPj2yLmRUCRZLGhHlpdbmGc8656NUnkC4FUsLv8ADmEYw+X5f0KsHK3cOBOxqkhY3rTeBCSRuABOCucnlXAZ0l3WZmO4DVkjbWZrERcA+wCVgX3o6ygLqP/nsAT4dtW8k3zy5+GJgl6XVJxxA8aerP4WKjL+t4Deecc1FSXWcuJR0CHAdsMrNdYdpZBKt1jwG2AH80s4UN29SGJakv8LSZHd/cbWlqvQ6NtxkTxjdYfb6NmnOuPZD0mpkdcHN7nb8jNbPPgdyItCeAJ+rfPOecc651aqyNvVs8M9tCcN9rnUgaSrA6try9ZpZU37ZIegLoF5F8TeRiJeeccy1Pnad2y06UDidYVHQc0M3MLimX3g/INzP/zq4FSkxMtNJ7N51zztVOg03thpVdDMwHOvPNLRiXhNlHAC8DlwJ/q1drnXPOuVaiPs/a/T6wkOA2mLOouNIVM9sIvEHwTFjnnHOuTavPiPQa4GMgxcw+l3RiJWU2EDxK0DnnnGvT6hNIE4GHw9W7VfkQ+Hb9muQaW/HHu3jzDyvqff5xv/EnCznnXKn6PJAhBviihjJxBI+/c84559q0+gTSLcBJNZRJAt6uR93OOedcq1KfQPokkCyp0n2sJF0EnAA8Fk3DnHPOudagPt+R3kawG8lD4TZhsQCSfgYkA/9J8MD6Oxuqkc4551xLVecRqZntBFIIdkr5McHemyK4r/THwEvAeDOr6XtU14plZGQwatQoRo8ezbp16yrkFRcXk5aWRnJyMmlpaRQXF1fIT0lJ4ZJLLik7XrBgAUlJSSQnJ/PCCy8AsGXLFuLj40lNTSU1NZVnnnmm8TvlnHP1UJ+pXcxsq5mlAicCPwV+C1wJjDCzFDPbXpt6JL1UizL3SBocvv/vWpSfKelNSXXelFvSFkmH1fW8cudnSoqrx3lZkg54WkY15SeXfibNYefOncyfP5+srCwWL17MzJkzK+RnZGQwaNAgVq1axcCBA8nIyCjLe/rppznkkEPKjj/55BMWLFhATk4OmZmZXHPNNezbF6xTO+mkk8jKyiIrK4szzzyzSfrmnHN1VatAKmmqpBMi081svZktMLObzOzPZvZaXS5uZqNqUeYSM9sUHtYYSIEZwBlm1uR7c5rZGeEeoY1tMtBsgTQ3N5fk5GRiYmLo168fu3fvZu/eb/Yrz8rKYuLEiQBMmjSJ7OxsAPbv38+f//xnrrjiirKyW7ZsYfDgwXTs2JEePXrQrVs33n//fQDWr19PcnIyU6dOZceOHU3YQ+ecq73ajkgziHhSkaQLJdX/ZsSgjt3hz9RwVPaopLckPShJYV6WpERJtwBdJOVVNdqU9FfgaOApSb+UtFmBOEn7JY0Ny62S1F/SoZKeC/f2XEAwRV1VW6+WNDN8/8fSvksaL2lx+H6LpMMk9Q1HxXdLeiO8RpcaPo7zJb0U7nk6UtJBkt4Nn11MePxe2IcfAHPDz+KYcBS+SdIGSQ/X+h+gngoKCoiPjy87jo2NpaCgoNL8uLi4siB4//3385//+Z907ty5rGz//v3Jy8vj888/Z/v27axfv56CggKOPPJINm/ezKpVqxgzZgxXX311Y3fLOefqpV5Tu6G+BN+VNpQTCTbTHkwQDEeXzzSza4EvzWx4VaNNM7sc+Ag4xcz+l+AxhoOBMcBrBKuNOwFHmdl7wPVAjpmdCDwF9K6mfdkEi6kgeChFd0kdw7pXVVJ+APBnMxsCFBI84L863cIR+gzgXjPbDywGSvv6PWC9mWWHbZ0VfhbvA9cCJ5rZCcDllVUu6VJJayWtLfiisIamVC8hIYHCwm/qKCoqIiEhodL80rzi4mIefPBBLrroogPqmjNnDpMmTeIXv/gFw4YNo2fPnnTq1IkePXoAcP755+MP2XfOtVTRBNKGtsbMPgwDSB5BoI7WKmBs+LqZIOiNAF4N88cSBCvM7BlgZzV1vQacJKkHsJfgwfyJBMG1skD6gZnllTu3bw1tfShsRzZwSPhd673A1DA/HbivinM3AA9KOh/4urICZrbQzBLNLDGhW1wNTaleUlISOTk5lJSUsHXrVrp3706nTp3K8lNSUsjMzAQgMzOTlJQUPvjgAwoLC5k4cSJXX301y5Yt45577gHgRz/6EStXrmTevHl07dqV3r17U1RUVFbfihUrGDhwYFRtds65xtKS9iPdW+79PhqmbasIRmg9geuAWUAqweiyVK32kTOzEklbgIsIViZvAE4BjgHerOSUyP7UNLUb2Q4zs22S/i1pHMFDLqr63vdMgj8KfgD8TtIQM6s0oDaE+Ph4ZsyYQUpKCpKYN28eeXl5LF++nFmzZjFt2jTS09NJTk7mqKOO4r777qNz585lo8rSRUqlK3enTp3Ktm3b6Nq1K3feGdw19eKLL3LDDTfQo0cPOnfuzN13391Y3XHOuai0pEBaGyWSOppZSS3L5wKLgM1mViwpD7gMmBjmZxMEpxslTQDiK63lG9nArwhGh/nA7cBrVt9NXSs6F3hR0higyMxKh2T3EIyaHzCz0scu7gJ6QPDdKfAfZvaipBzgPKA7wXRyo0lPTyc9Pb1C2vDhwwHo0qULDz30UJXnlt7SUmrRokUHlJk8eTKTJ09uiKY651yjqsvUbkMEi2gtBDbU9tYWM9sLbANeCZNWEQSg/PB4DjBW0jqC+2G31lDlKuBI4GUz+zdQTOXTuvWxM7wd6K/AxeXSnyIIjOWndR8GZkl6neC72MWS8oHXgT820cph55xzgGozmJK0n7oHUjOz1jbibXHC+0v/aGbJNRaupeN7DbS/z7ir5oJV8N1fnHPtkaTXzOyAe/7rEuiqvDWkgcq7CJKuJXjgRZPfE+ucc652ahVIzawlre5F0qHAC5VkjTezqO7cb+S6/0zEbT3APDOrdDWumd0C3BLNNZ1zzjWuWk3turYlMTHR/L5M55yrm6qmdlvUSNM555xrbTyQOuecc1HwQOqcc85FwQOpc845FwW/z7Md+uijj5g9e3a1ZWrKd845F/ARqXPOORcFD6TOOedcFDyQOuecc1HwQOqcc85FwQOpq5eMjAxGjRrF6NGjWbduXYW84uJi0tLSSE5OJi0tjeLiYgBmzZpFSkoKI0eOZNasWc3RbOeca3DtMpBKminpTUk7wwfD16eOuZLeCH/OlvSrhm5nHdoyXNIZTXW9nTt3Mn/+/LINumfOnFkhPyMjg0GDBrFq1SoGDhxIRkYGAH/4wx9YuXIla9asYc2aNbzxxhtN1WTnnGs07TKQAjOAM8wsPnwwfH1cBnzHzFrC0Go40GSBNDc3l+TkZGJiYujXrx+7d+9m7969ZflZWVlMnBjsnT5p0iSys7MBiImJAaCkpIRu3brRs2fPpmqyc841mnYXSCX9FTgaeErSLyT9KUzPkDRf0kuSNks6u5o6ngK6AbmSzo3Imy7pVUnrJT0mqWuYfoykV8K8GyTtrqGdV0vKD+u5JUzLknSrpDWS3pGULCkGuAE4V1JeZHvK1XeppLWS1u7Zs6f2H1glCgoKiI+PLzuOjY2loKCg0vy4uDh27Phm05wrr7ySo48+miOPPJLY2Nio2uGccy1BuwukZnY58BFwCrAzIvtIYAwwkWq2LzOzHwBfmtlwM3skIvtxMxthZsOAN4GLw/R5BFumjQivXyVJE4DJQFJYz23lsg82s5HAVcD1ZvYVcB3wSBXtKW3zQjNLNLPErl27Vnf5GiUkJFBYWFh2XFRUREJCQqX5kXl33nknH3zwAZ999hlLly6Nqh3OOdcStLtAWoN/mNl+M9sEHFHPOo6XtEpSPsGG3EPC9JOBv4fvl9RQx/eA+8xsD4CZFZTLezz8+RrQt55tjEpSUhI5OTmUlJSwdetWunfvTqdOncryU1JSyMzMBCAzM5OUlBSAskVHBx98MN26dSPagO6ccy2BPyKwor3l3quedWQAk81svaRpQGo96hBQ1UaxpW3cRzP9+8XHxzNjxgxSUlKQxLx588jLy2P58uXMmjWLadOmkZ6eTnJyMkcddRT33RfsW56WlsaOHTsoKSlhzJgxpKamNkfznXOuQXkgbXg9gI8ldSQYkW4P018BfgQ8AvykhjqeA66TtMTM9khKiBiVRtoVXrfJpKenk56eXiFt+PDhAHTp0oWHHnrogHMee+yxpmiac841KZ/abXi/A3KB5cBb5dKvAv5L0hqC72KLqqrAzJYCTwFrJeUBNd1a8yIwuLrFRs455xqHzKqaQXQNKVy9+6WZmaSfAFPM7IfN0ZaePXvapZdeWm0Z3/3FOecqkvSamSVGpvvUbtM5CfiTJAGFQHr1xZ1zzrUGHkirIWko8EBE8l4zS6prXWa2ChjWWPU755xrHj612w4lJiba2rVrm7sZzjnXqlQ1teuLjZxzzrkoeCB1zjnnouCB1DnnnIuCLzZqh3bufJP/9/eRVeaf8+M1Tdga55xr3XxE6pxzzkXBA6lzzjkXBQ+kzjnnXBQ8kDrnnHNR8EDqnHPORcEDqatSRkYGo0aNYvTo0axbt65CXnFxMWlpaSQnJ5OWlla2afc555zDqFGjSEpKIiMjo6z85s2bmTRpEuPGjWPq1Kll6cuWLWP8+PGccsop3H777U3SL+eca0gt7hGBkmYCPwXWmVlaHc7bAiSa2Wd1vN4B50mKA84zs79Uc15f4GkzO74u14uWpJ7AfDM7u751HHNMN7v5liFV5p/z4zXs3LmT8ePH88orr7B9+3YuuOACcnJyysr89a9/5dNPP+V3v/sdN9xwA9/61re4/PLLeffddxkwYADFxcUcf/zxbNy4kc6dO3PGGWfwt7/9jSOPPLKsjs8++4xp06bx+OOPExMTU9/uOOdck2hNjwicAZxRlyDaCOLCdrQ4ZvZRNEG0tnJzc0lOTiYmJoZ+/fqxe/du9u7dW5aflZXFxIkTAZg0aRLZ2dkADBgwAICOHTty0EEHIYl//etf7Nmzh5///OekpqaWbfD9zDPPkJCQwA9+8ANOP/103njjjcbulnPONbgW9UAGSX8FjgaeknQ8QaCPBQqAVDPLlrQKuAjYCTwEHA6sAVRNvX2BpQQbbp8IvANMNbM95cp0AZ4AHgPGA8eEm2ovN7NZNbS7A3ALkAp0Av5sZgskpQKzgc+A44HXgPOtimmAcHS8BDgF6AhcCtwM9Afmmtlfy4+EJU0DfgB0BY4BnjCzq6uo+9KwPg47rObRX0FBAfHx8WXHsbGxFBQUlI0oy+fHxcWxY8eOCuffdNNNTJkyhU6dOvHRRx/x+uuvs2nTJnr06MGoUaMYN24cH330Ee+99x5ZWVm8+eabXHrppaxevbrGtjnnXEvSokakZnY58BFBIFkGDAbGEASgZEmdgKPM7D3geiDHzE4EngJ611D9QGChmZ0AfE7FEWd34J/AEjO7G7gWeN/MhtcUREMXA0VmNgIYAUyX1C/MOxG4KuzL0cDoGuraZmYnA6uADOBs4LvADVWUHw6cCwwFzpX0H5UVMrOFZpZoZomHHFLz308JCQkUFhaWHRcVFZGQkFBpfmTeokWL2LhxI9dff31Z2aFDh9KrVy8OOeQQhg8fzrvvvktCQgLjxo0jJiaGYcOG8emnn9bYLueca2laVCCNsAoYG75uJgioI4BXw/yxwGIAM3uGYIRanW1mVjrcWRzWV+pJ4D4zW1TPtp4KTA1HsLnAocCAMG+NmX1oZvuBPKBvDXU9Ff7MB3LNbJeZfQoUh9/dRnrBzIrMrBjYBPSpZx8qSEpKIicnh5KSErZu3Ur37t3p1KlTWX5KSgqZmZkAZGZmkpKSAsCTTz7JkiVLeOCBBzjooOC/V//+/dmzZw+7du3i66+/ZtOmTfTp04fU1NSyRUzbtm3jkEMOaYimO+dck2pRU7sRVgGXAz2B64BZBFOn2eXK1GWlVGTZ8sergQmSllQ17VoDAVea2bIKicHU7t5ySfuo+TMvLb8/4tz9VZxb1/prJT4+nhkzZpCSkoIk5s2bR15eHsuXL2fWrFlMmzaN9PR0kpOTOeqoo7jvvvsASEtLY9CgQZx66qkAPPjgg/Tq1Ytbb72VCRMmUFJSwvTp0zniiCM44ogjSE1NZezYsZSUlDBv3ryGaLpzzjWplhxIc4FFwGYzKw5He5cBE8P8bCANuFHSBCC+0lq+0VvSyWb2MjAFyCmXdx3wO+AvBCuGdwE96tDWZcBPJa0wsxJJxwLb63B+i5Senk56enqFtOHDhwPQpUsXHnrooQPO2b17d6V1ff/73+f73//+AelXX301V19d6de6zjnXKrTYqV0z2wtsA14Jk1YRBLf88HgOMFbSOoKp1a01VPkmcKGkDUACcFdE/lVAZ0m3mdkOYLWkjZLm1qK59xBMq66TtBFYQMv+I8U551wDaXH3kTaG5rrns6WqzX2kzjnnKmpN95E655xzrUabmn6UdCjwQiVZ4+s7GpU0FHggInmvmSXVp76wzieAfhHJ10QuVnLOOdfytYupXVdRYmKirV27trmb4ZxzrYpP7TrnnHONwAOpc845FwUPpM4551wU2tRiI1c7m3Z+zrBHK1/XtP7s05q4Nc4517r5iNQ555yLggdS55xzLgoeSJ1zzrkoeCB1zjnnouCB1B0gIyODUaNGMXr06LL9QksVFxeTlpZGcnIyaWlpFBcXA3DOOecwatQokpKSyMjIKCufnp7OkUceySWXXFKWtmvXLk4++WTi4uJYvHhxk/TJOecaiwdSV8HOnTuZP38+WVlZLF68mJkzZ1bIz8jIYNCgQaxatYqBAweWBc0//OEPvPTSS6xcuZIbb7yxLMD+/ve/P2C7tS5duvDEE09w1VVXNUWXnHOuUXkgBSTNlPSmpAfreN4WSYfV43oHnCcpTtKMutbV0HJzc0lOTiYmJoZ+/fqxe/du9u79Zu/wrKwsJk4MtoSdNGkS2dnBPusDBgwAoGPHjhx00EFIAqBXr14HXOPggw/m29/+dmN3xTnnmoQH0sAM4AwzS2vGNsSF7WhWBQUFxMd/s0d6bGwsBQUFlebHxcWxY8eOCuffdNNNTJkyhU6dOjVNg51zrpm1+0Aq6a/A0cBTkkyBOEn7JY0Ny6yS1F/SoZKek/S6pAWAqqm3r6S3JN0vaYOkRyV1jSjTRdJSSdOBW4BjJOVJmivpSEnZ4fFGSclVXKeDpIywTL6kX1RR7lJJayWt/frzoio/j4SEBAoLC8uOi4qKSEhIqDQ/Mm/RokVs3LiR66+/vsr6nXOurWn3gdTMLgc+Ak4BlgGDgTHAa0CypE7AUWb2HnA9kGNmJwJPAb1rqH4gsNDMTgA+p+KIszvwT2CJmd0NXAu8b2bDzWwWcB6wzMyGA8OAvCquMRzoZWbHm9lQ4L4q+rnQzBLNLPHgQ2KrbHBSUhI5OTmUlJSwdetWunfvXmF0mZKSQmZmJgCZmZmkpKQA8OSTT7JkyRIeeOABDjqo3f+3cs61I/4br6JVwNjwdTNBQB0BvBrmjwUWA5jZM8DOGurbZmarw/eLw/pKPQncZ2aLqjj3VeAiSbOBoWa2q4pym4GjJd0p6XSCgF1v8fHxzJgxg5SUFKZMmcIdd9xBXl4ec+fOBWDatGnk5+eTnJxMfn4+06ZNAyAtLY3PPvuMU089ldTUVLZv3w7Ab3/7W375y1/y7LPP8r3vfY8vvvgCCL5fXbRoEbfccguXX355NE12zrlm5fuREiz+ARIJRqOXAz2B04EXgWeAQjP7k6Q84Cwz+yA8rwA41sw+q6TOvsBKM+sTHo8DrjSzs8LrZQKHABeYmYXlny6/AbmknsCZwExgblVBV1J34DRgGvCpmaVX19+uxxxrA269s9I8f9auc85VzvcjrZ1cYBSw38yKCaZTLyMYqQJkA2kAkiYA8ZXUUV5vSSeH76cAOeXyrgN2AH8Jj3cBPUozJfUBPgmnff8GfKeyC4Srfw8ys8eA31VVzjnnXOPwQFqOme0FtgGvhEmrCIJbfng8BxgraR1wKrC1hirfBC6UtAFIAO6KyL8K6CzpNjPbAawOFw3NBVKBPEmvAz8C5lVxjV5AVjhazgB+XXNPnXPONRSf2m0klU3VthQ+teucc3XnU7vOOedcI/CNvaMk6VDghUqyxjf0aFRSLhD5pIMLzCy/svLOOecanwfSKIXfbQ5vomslNUQ9g+MPYa1P4TrnXIPwqV3nnHMuCr7YqB2StAt4u7nb0YwOAw6497edaM99B++/9z+6/vcxs8MjE31qt316u7KVZ+2FpLXttf/tue/g/ff+N07/fWrXOeeci4IHUueccy4KHkjbp4XN3YBm1p773577Dt5/738j8MVGzjnnXBR8ROqcc85FwQOpc845FwUPpG2UpNMlvS3pPUnXVpIvSfPD/A2S2tT2a7Xof1rY7w2SXpI0rDna2Vhq6n+5ciMk7ZN0dlO2r7HVpv+SUiXlSXpD0sqmbmNjqsX//1hJ/5S0Puz/Rc3RzsYg6V5Jn0jaWEV+w//uMzN/tbEX0AF4HzgaiAHWA4MjypwBPAsI+C6Q29ztbuL+jwLiw/cT2lv/y5VbQbDJ/NnN3e4m/vePAzYBvcPjbzV3u5u4//8N3Bq+PxwoAGKau+0N1P+xBPsyb6wiv8F/9/mItG0aCbxnZpvN7CvgYeCHEWV+CCyywCtAnKQjm7qhjaTG/pvZS2a2Mzx8BTiqidvYmGrz7w9wJfAY8ElTNq4J1Kb/5wGPm9lWADNrS59BbfpvQA9JAroTBNKvm7aZjcPMsgn6U5UG/93ngbRt6kWwQXmpD8O0upZprerat4sJ/kJtK2rsv6RewFnAX5uwXU2lNv/+xwLxkrIkvSZpapO1rvHVpv9/Ao4DPgLygZ+b2f6maV6za/Dfff6IwLZJlaRF3udUmzKtVa37JukUgkA6plFb1LRq0/87gGvMbF8wKGlTatP/g4GTgPFAF+BlSa+Y2TuN3bgmUJv+nwbkAeOAY4DlklaZ2eeN3LaWoMF/93kgbZs+BP6j3PFRBH951rVMa1Wrvkk6AbgHmGDBdnhtRW36nwg8HAbRw4AzJH1tZv9okhY2rtr+///MzL4AvpCUDQwD2kIgrU3/LwJuseBLw/ckfQAMAtY0TRObVYP/7vOp3bbpVWCApH6SYoCfAE9FlHkKmBquYPsuUGRmHzd1QxtJjf2X1Bt4nGBj9Lbwy7O8GvtvZv3MrK+Z9QUeBWa0kSAKtfv//ySQLOlgSV2BJODNJm5nY6lN/7cSjMaRdAQwENjcpK1sPg3+u89HpG2QmX0t6WfAMoIVfPea2RuSLg/z/0qwUvMM4D1gD8FfqG1CLft/HXAo8JdwVPa1tZFdMWrZ/zarNv03szclLQU2APuBe8ys0tslWpta/vv/HsiQlE8w1XmNmbWJ7dUkPQSkAodJ+hC4HugIjfe7zx8R6JxzzkXBp3adc865KHggdc4556LggdQ555yLggdS55xzLgoeSJ1zzrkoeCB1zjnnouCB1LlWRpJJavP3rUnqG/Y1o7nb4lx1PJA655xzUfBA6pxzzkXBA6lzbUD5aVBJx0h6VNIOSbskPSfp+LDc4ZIWSvpYUrGkV8MdcCLrmx3WlyrpQkmvS/pS0ieS7pX07SraMUDSIknbJX0l6aPweEAN1zhPUq6k3ZK2SJoNfBAWvbB0Ojt8TQvPj5H0M0mZkv4laa+kAknPS5pQRfu2hK+ukuZK2hqe956ka1TFVjiSRkp6JOzX3vDze07SOZWUTQo///8LP4NtkhZI6ln5v55r7fwRgc61MqXfj5qZyqX1JQg8K4HjCR7AvgboS7DvaAFwMrAU+Dwsl0DwQPP9wLGlm1yH9c0meEbpU8CpwCPAxwTbzY0Jr5VkZp+WO2cE8DzQIzxvE8GOIj8EdgPjzWxtJdd4Gvg+8E/gfSA2vN5k4OfAeuAf5T6Cf5hZXhjMtwMvAW8DnwJHApPCvk03s3siPrstBM9d/QDoCSwn2NB6cng828zmRJwzHbgL2Bf2613gWwQ76BSaWWq5shcBdwN7w7LbgAHAD4B/A98t/zm7NsLM/OUvf7WiF8HeiRaR1rc0HfhNRN7vwvQCgo28DyqXd0GY98eIc2aH6V8BJ0bk/THM+1u5NBEEbwPSIsqfG6a/FXHt0mt8EXmNiD5lVPE5dAKOqiQ9FtgY9rdLRN6WsM7M8nkEgbEwfHUslz4YKAnrGlLJtY4q9/7Y8PN6D+gVUW4cQSB+orn///ir4V8+tetc27IFuCUi7f7wZydglpntL5e3hGBENryK+h4ws9cj0mYDRcB5kjqFaaMIRp8vm9mD5Qub2SNADsFWXZVtoL6wkmvUyMz2mtmHlaQXAfcC8cCIKk6faWZfljvnE4Kt1WLDdpb6KcEuWb83szcqudaHEWU7Aj83s+0R5VYQjFAnSepRi+65VsS3UXOubckzs30RaaWbFr9jZrvKZ5jZPkn/JtjcuDIrIxPMrEhSHpACHAfkAd8Js1dUUc8KgiB6IpAdkVfvzaQlDQFmAWMJpnU7RxTpVclpRWb2XiXp28Kf8eXSvhv+fLYWzTk5/JkSTnNH+hbBtmbHAq/Voj7XSnggda5tKYpMsGB/ykrzQl8T7tdYiX9Xkf5/4c/YiJ9VbZBcmh5XTV11Em7KvILg99gLBCO+zwm+8x1O8N1sp0pOLayiyq/Dnx3KpcWFP7dTs0PDn7NqKNe9FnW5VsQDqXOuOkdUkV66arco4melq3kJRovly5VX3xWPvwW6AKeYWVb5DEm/Jgik0SoMf/Yi+I63OqV9izWzzxvg2q6V8O9InXPVSYlMkBRLMOIrJlhgBFD6HWdqFfWUpq+rw7VLp6g7VJHfHyiIDKKhA9pdT6+EPyu9naaKsskNdG3XSnggdc5V5wJJJ0akzSaYyn3IzPaGaasJbkEZI+ns8oXD47HAOwSLjmprJ8FotXcV+VuABEknRFzvYuC0OlynOncRTPn+TtLgyExJ5b9b/hPBCt8/Sjq2krIxkjzItkE+teucq86zwGpJ/4+K95FuAa4tLWRmJulCgvsyH5H0JMFU6ECCezR3AVMjVgxXy8x2S8oFkiU9SBCI9wFPmdkG4A6CgJkTtq+I4N7OMcCjwNmVVlwHZrZJ0gyC24ZeD/v1LsH3oYlhv04Jy74lKZ1gxfAbkpaGbe5I8MdAMsG9roOibZdrWTyQOueq80fgCeAqgvtBdwMZwH+Ht4yUMbPccLXqb4HvETwY4TPgIYLbR96ux/UvCNtwOjCF4H7VD4ENZrZU0qTweucSBNk1BIHtaBogkAKY2d2SNgK/IpiinkzQrw3APRFlF0taD/wybMepBPfJfkQQ3B9piDa5lsWfbOScO0C5pw4dsJDHOVeRf0fqnHPORcEDqXPOORcFD6TOOedcFPw7Uueccy4KPiJ1zjnnouCB1DnnnIuCB1LnnHMuCh5InXPOuSh4IHXOOeei8P8BEEjlb7lHrfUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fi_df = pd.DataFrame({'feature_importance': relevantes, 'feature_name': cols})\n",
    "\n",
    "fi_df.sort_values(by=['feature_importance'], ascending=False, inplace=True)\n",
    "\n",
    "relevantes = fi_df[fi_df.feature_importance > 0.02]\n",
    "\n",
    "#plt.figure(figsize=(10,8))\n",
    "g=sns.barplot(x=relevantes['feature_importance'], y=relevantes['feature_name'])\n",
    "\n",
    "plt.xlabel('Importance', fontsize=20)\n",
    "plt.ylabel('Feature', fontsize=20);\n",
    "i=0\n",
    "for index, row in relevantes.iterrows():\n",
    "    if row.feature_importance==1:\n",
    "        temp=0.025\n",
    "    else:\n",
    "        temp=0.05\n",
    "    g.text(row.feature_importance+temp, i, round(row.feature_importance, 4), color='black', ha=\"center\", va=\"center\", fontsize=9)\n",
    "    i+=1\n",
    "filename = 'dissertation/feat_imp_{0}_{1}.eps'.format(trainerDSName, \"SVM\")\n",
    "plt.savefig(filename, dpi=300, bbox_inches = \"tight\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d373c79e-6d45-49ad-b66c-67cafa4d01cf",
   "metadata": {},
   "source": [
    "## Build LaTeX tables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "5d248a0e-cf3c-4d6e-8286-5c9a07d7b568",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['fscore_AB-TRAP_CIC.csv', 'fscore_BoT-IoT_CIC.csv', 'fscore_NB15_CIC.csv', 'fscore_SCAN_AB-TRAP_CIC.csv', 'fscore_SCAN_BoT-IoT_CIC.csv', 'fscore_SCAN_CIC-IDS_CIC.csv', 'fscore_SCAN_NB15_CIC.csv', 'fscore_SCAN_ToN-IoT_CIC.csv', 'fscore_ToN-IoT_CIC.csv']\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "from supportFiles.myFunc import saveTable\n",
    "\n",
    "files = [s for s in os.listdir(\"./dissertation/\") if ((\".csv\" in s) and (\"fscore_\" in s))]\n",
    "print(files)\n",
    "for file in files:\n",
    "    tableName = file.replace(\".csv\",\"\")\n",
    "    table = pd.read_csv(\"./dissertation/\"+file, sep=',')\n",
    "    caption = \"Feature importance {0}\".format(tableName)\n",
    "    label = tableName\n",
    "    saveTable(table, tableName, caption, label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "759912b5-6d6a-4549-ae14-e2b862001b7c",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}