2Detect2_Reproducibility.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c491d25a-eeb4-4814-8418-1210ff14df03",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pickle\n",
    "from utils.partition import weight_update_statistics\n",
    "from utils.flower_detection import mal_agents_update_statistics\n",
    "from sklearn.metrics import confusion_matrix, accuracy_score\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "bd93386f-00a8-4caf-8128-591e9ba17605",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.   0.05 0.1  0.15 0.2  0.25 0.3  0.35 0.4  0.45 0.5  0.55 0.6  0.65\n",
      " 0.7  0.75 0.8  0.85 0.9  0.95 1.   1.2  1.4  1.6  1.8  2.   2.2 ]\n",
      "27\n"
     ]
    }
   ],
   "source": [
    "kappas = np.round(np.concatenate((np.arange(0, 1, 0.05), np.arange(1,2.2,0.2))),3)\n",
    "print(kappas)\n",
    "print(len(kappas))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "b14628e4-656e-44dc-804e-c4c01cf268cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "meta_metrics = {}\n",
    "\n",
    "for i in range(1,21):\n",
    "    with open(f'data/reproducibility_detection2/{i}_metrics.pkl', 'rb') as handle:\n",
    "        metrics = pickle.load(handle)\n",
    "    meta_metrics[i] = metrics\n",
    "    \n",
    "\n",
    "results = {}\n",
    "\n",
    "for kappa in kappas:\n",
    "    result_at_kappa = []\n",
    "    #print(kappa)\n",
    "    for i in meta_metrics.keys():\n",
    "        detected = mal_agents_update_statistics(meta_metrics[i], kappa=kappa)\n",
    "        result_at_kappa.append(detected)\n",
    "    results[str(np.round(kappa,2))] = result_at_kappa\n",
    "    \n",
    "with open('data/reproducibility_detection2_kappas.pkl', 'wb') as handle:\n",
    "    pickle.dump(results,handle,protocol=pickle.HIGHEST_PROTOCOL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "465b7a58-8dad-4407-9e38-b74a2bc3de9a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['2', '6', '7', '0', '3', '5', '8', '9', '4', '1'])"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "meta_metrics[1].keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "4bc1928b-3357-44d8-bd0f-fc66cb0307c0",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('data/reproducibility_detection2_kappas.pkl', 'rb') as handle:\n",
    "    results = pickle.load(handle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "dd28f106-51fb-4a11-a5a1-61aff39b665e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['0.0', '0.05', '0.1', '0.15', '0.2', '0.25', '0.3', '0.35', '0.4', '0.45', '0.5', '0.55', '0.6', '0.65', '0.7', '0.75', '0.8', '0.85', '0.9', '0.95', '1.0', '1.2', '1.4', '1.6', '1.8', '2.0', '2.2'])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "0f46630c-471a-4ec2-9d3b-82f9708b59f9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([0., 0., 0., 0., 0., 0., 0., 0., 1., 0.]),\n",
       " array([0., 0., 0., 0., 0., 0., 1., 0., 0., 0.]),\n",
       " array([0., 0., 0., 0., 0., 0., 0., 0., 1., 0.]),\n",
       " array([0., 0., 0., 0., 0., 0., 0., 0., 0., 1.]),\n",
       " array([0., 0., 0., 0., 0., 0., 1., 0., 0., 0.]),\n",
       " array([0., 0., 0., 0., 0., 0., 1., 0., 0., 0.]),\n",
       " array([0., 0., 1., 0., 0., 0., 0., 0., 0., 0.]),\n",
       " array([0., 0., 0., 1., 0., 0., 0., 0., 0., 0.]),\n",
       " array([0., 0., 0., 0., 0., 1., 0., 0., 0., 0.]),\n",
       " array([0., 0., 0., 0., 0., 0., 0., 1., 0., 0.]),\n",
       " array([0., 0., 0., 0., 0., 0., 1., 0., 0., 0.]),\n",
       " array([0., 0., 1., 0., 0., 0., 0., 0., 0., 0.]),\n",
       " array([0., 0., 0., 0., 0., 0., 0., 0., 0., 1.]),\n",
       " array([0., 0., 0., 0., 0., 0., 0., 0., 0., 1.]),\n",
       " array([0., 0., 1., 0., 0., 0., 0., 0., 0., 0.]),\n",
       " array([0., 0., 0., 0., 0., 0., 0., 0., 1., 0.]),\n",
       " array([0., 0., 1., 0., 0., 0., 0., 0., 0., 0.]),\n",
       " array([0., 1., 0., 0., 0., 0., 0., 0., 0., 0.]),\n",
       " array([0., 0., 0., 1., 0., 0., 0., 0., 0., 0.]),\n",
       " array([0., 0., 0., 1., 0., 0., 0., 0., 0., 0.])]"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#results\n",
    "results['2.0']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "3b834d6b-2b44-47b8-b644-e8913193e245",
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_true = [0, 0, 0, 1, 0, 0, 0, 0 ,0 ,0] * 20\n",
    "results_uni = {}\n",
    "for key in results.keys():\n",
    "    results_uni[key] = np.concatenate(results[key])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "0040a098-c632-4fd6-aa28-9928ee101516",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "200"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(Y_true)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "012e7830-67ca-4c45-8be2-17058ddadfea",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 1.,\n",
       "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0.,\n",
       "       0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.,\n",
       "       0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0.,\n",
       "       0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
       "       1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.,\n",
       "       0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0.,\n",
       "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0.,\n",
       "       0., 0., 0., 1., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
       "       0., 0., 0., 0., 0., 1., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0.,\n",
       "       0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.,\n",
       "       0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0.])"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_uni['2.0']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "d2be512e-115c-47f8-92f1-10c9ac4c0ec2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.2\n",
      "[[ 47 133]\n",
      " [  2  18]]\n",
      "0.6\n",
      "[[151  29]\n",
      " [ 13   7]]\n"
     ]
    }
   ],
   "source": [
    "l_accu = []\n",
    "l_TPR = []\n",
    "l_FPR = []\n",
    "\n",
    "for key in results_uni.keys():\n",
    "    accu = accuracy_score(Y_true, results_uni[key])\n",
    "    cmtx = confusion_matrix(Y_true, results_uni[key])\n",
    "    TN = cmtx[0,0]\n",
    "    FN = cmtx[1,0]\n",
    "    TP = cmtx[1,1]\n",
    "    FP = cmtx[0,1]\n",
    "    TPR = TP / (TP+FN)\n",
    "    FPR = FP / (FP+TN)\n",
    "    \n",
    "    if key in ['0.2', '0.6']:\n",
    "        print(key)\n",
    "        print(cmtx)\n",
    "    l_accu.append(accu)\n",
    "    l_TPR.append(TPR)\n",
    "    l_FPR.append(FPR) \n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "99a32529-b320-4aaf-a227-06bb16f19cbe",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7efe43555850>]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig1 = plt.figure(figsize=(8,6))\n",
    "ax1 = fig1.add_subplot(1,1,1)\n",
    "\n",
    "ax1.plot(kappas, l_accu)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "45f086f5-fb5a-4796-806f-f6e0f8bd7deb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'True Positive Rate')"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig1 = plt.figure(figsize=(5,5))\n",
    "ax1 = fig1.add_subplot(1,1,1)\n",
    "ax1.plot(l_FPR, l_TPR)\n",
    "ax1.plot([0,1], [0,1], color = 'r', ls='--')\n",
    "for i, (FPR, TPR) in enumerate(zip(l_FPR, l_TPR)):\n",
    "    if (i % 2 == 0 or i == 5) and (i != 10):\n",
    "        ax1.text(FPR, TPR+0.01, kappas[i], horizontalalignment='center', verticalalignment='bottom')\n",
    "    if i > 15:\n",
    "        break\n",
    "ax1.set_xlabel(\"False Positive Rate\", fontsize = 14)\n",
    "ax1.set_ylabel(\"True Positive Rate\", fontsize = 14)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "46623abf-3d4b-49b7-a0ee-8707de8cb0a1",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Sanity check"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "f29a6f6a-6390-46c4-aa18-44f3acc8b190",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0., 1., 0., 0., 0.])"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_metrics = {}\n",
    "test_metrics[0] = np.random.rand(2000,2)\n",
    "test_metrics[1] = np.random.rand(2000,2)*(4)\n",
    "test_metrics[2] = np.random.rand(2000,2)\n",
    "test_metrics[3] = np.random.rand(2000,2)\n",
    "test_metrics[4] = np.random.rand(2000,2)\n",
    "\n",
    "mal_agents_update_statistics(test_metrics, kappa=1, fix=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fc8f5c84-cc7b-49f4-8431-0b68a382559c",
   "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": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}