performance_evaluation_GAT-bias2false-faster.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2022.3.31"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "DATASET = \"../dataset/MMS.csv\"\n",
    "TOPOLOGY = \"../dataset/MMS_topology.pk\"\n",
    "RESULT=\"../topomad-1-results/MMS_GAT\"\n",
    "NAME=\"performance_evaluation_GAT\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from graph_lstm_vae_ad_faster import GraphLSTM_VAE_AD\n",
    "from graph_lstm_vae_ad_faster import GraphLSTM_VAE\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import roc_auc_score\n",
    "import numpy as np\n",
    "import os\n",
    "import pickle\n",
    "import time\n",
    "import torch\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"0\"\n",
    "gpu_choice =0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 读数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv(DATASET, header=[0,1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 参数"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "seq_len = 10\n",
    "num_epo = 150\n",
    "hid_dim = 3\n",
    "lea_rat = 1e-4\n",
    "nod_num = 50\n",
    "sam_num = 20\n",
    "del_tol = 5\n",
    "tra_siz = len(data)//3*2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 预处理标准化"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/user/anaconda3/lib/python3.7/site-packages/pandas/core/generic.py:3936: PerformanceWarning: dropping on a non-lexsorted multi-index without a level parameter may impact performance.\n",
      "  obj = obj._drop_axis(labels, axis, level=level, errors=errors)\n"
     ]
    }
   ],
   "source": [
    "metric = data.drop(['TimeStamp', 'label'], axis = 1)\n",
    "metric.columns.names = ['pod','metric']\n",
    "tempm = metric.swaplevel('metric','pod',axis=1).stack()\n",
    "tempm = (tempm-tempm.mean())/(tempm.std())\n",
    "metric = tempm.unstack().swaplevel('metric','pod',axis=1).stack().unstack()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 读拓扑"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(TOPOLOGY, 'rb') as f:\n",
    "    edge_index = pickle.load(f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 训练VAE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [150/150], Iter [60/63], total_loss: -3139.9795, recon_loss: -3251.1792, kl_loss: 111.1997\n",
      "valid_loss: [1797.2425220630787, 1792.7706344039352, 1794.1688006365741, 1796.021701388889, 1791.7787407769097, 1791.3928900824653, 1791.4117250795719, 1787.3687789351852, 1774.737765842014, 1757.5021656177662, 1735.7508770978009, 1700.36456072772, 1657.8076669198495, 1598.8004421657986, 1534.607132523148, 1464.7627179181134, 1393.8099591290509, 1318.2068549262153, 1241.9591200086807, 1167.203547724971, 1088.9635484483506, 1021.0960806387442, 951.195246943721, 881.5971498842592, 811.7305207429109, 745.710139521846, 683.5525625723379, 617.8870759186922, 555.2191648130064, 495.01986355251734, 432.7162939001013, 376.3482315628617, 318.09761103877315, 259.3723104971427, 204.5432691220884, 147.08499484592014, 95.11597866482205, 39.608976929276075, -11.23226766233091, -60.956920835706924, -115.09814862851744, -165.38826553909868, -212.9390174018012, -259.55972177010995, -308.5170581958912, -353.0006369131583, -398.8398844401042, -443.47210015190973, -488.46857480649595, -528.3235993561922, -572.0284469039352, -613.4020125777633, -653.8664663809317, -691.8036679868345, -731.2739574291088, -767.3073255750868, -805.4630104347511, -844.292955186632, -878.0325339988426, -915.3562056929977, -948.6951746057581, -985.0469767252604, -1016.1740180121528, -1048.8782574688946, -1082.7193852177372, -1114.6886234989872, -1149.1742666739005, -1177.9172227647568, -1206.9558512369792, -1237.2553620515046, -1266.9187373408565, -1299.334802698206, -1327.2678991247105, -1355.6784984447338, -1384.1067798755787, -1414.112951208044, -1443.0340395326969, -1466.565872757523, -1495.3166368272568, -1522.6987440321182, -1545.5413999204281, -1573.7779088903355, -1600.0789840133102, -1626.7043954354745, -1648.263355396412, -1671.5140832971645, -1701.2433132595486, -1726.1482340494792, -1748.9267261646412, -1772.308394820602, -1799.4995478877315, -1820.6731092664932, -1842.080982349537, -1864.3832013165509, -1888.8787389684605, -1910.92596661603, -1934.5868372034145, -1956.472109194155, -1981.030318648727, -2002.5057237413193, -2028.7577311197917, -2047.9793882016781, -2066.4502043547454, -2092.1611644603586, -2115.5165201822915, -2131.2792742693864, -2152.483556676794, -2173.695222077546, -2192.7771086516204, -2215.5243688512733, -2236.8589319299767, -2262.4892849392363, -2272.68810583044, -2295.048886899595, -2315.049442997685, -2335.0339717158563, -2355.7670808015046, -2373.5344871238426, -2394.659722222222, -2415.7025915075233, -2432.0976110387733, -2450.3046784577546, -2472.5390172887733, -2494.4460539641204, -2506.555329499421, -2530.8907335069443, -2545.906964337384, -2563.7138038917824, -2580.855523003472, -2603.3242277922454, -2621.626202618634, -2638.379322193287, -2656.296214916088, -2672.9157262731483, -2686.0326515480324, -2706.07921911169, -2725.6223777488426, -2744.1565031828704, -2759.6338975694443, -2782.8327003761574, -2794.8412724247687, -2808.931007667824, -2832.070674189815, -2844.3180881076387, -2859.7759150752313, -2879.047417534722, -2891.883617259838, -2912.3979040075233, -2928.165753399884]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 63/63 [00:20<00:00,  3.05it/s]\n",
      "100%|██████████| 27/27 [00:04<00:00,  6.69it/s]\n",
      "/home/user/hzl/bishe-mms/faster/graph_lstm_vae_ad_faster/graphlstm_vae_ad.py:238: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  X.interpolate(inplace=True)\n",
      "/home/user/anaconda3/lib/python3.7/site-packages/pandas/core/frame.py:4153: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  downcast=downcast,\n",
      "100%|██████████| 46/46 [02:05<00:00,  2.73s/it]\n",
      "100%|██████████| 91/91 [04:11<00:00,  2.76s/it]\n",
      "/home/user/hzl/bishe-mms/faster/graph_lstm_vae_ad_faster/graphlstm_vae_ad.py:312: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  X.interpolate(inplace=True)\n",
      "100%|██████████| 46/46 [00:08<00:00,  5.28it/s]\n"
     ]
    }
   ],
   "source": [
    "#checklist：ver，name，seedi，filename\n",
    "from sklearn.metrics import average_precision_score\n",
    "import pickle\n",
    "aps = []\n",
    "maps = []\n",
    "traintime150s = []\n",
    "testtimes = []\n",
    "for seedi in range(5):\n",
    "    model = GraphLSTM_VAE_AD(name=NAME+\"VAE\", sequence_length=seq_len, num_epochs=num_epo, hidden_dim=hid_dim, lr=lea_rat, gpu = gpu_choice, variational = True, kind = 'GAT',seed=seedi, head=(8,8),dropout=(0.4,0.4), bias=(True, False))\n",
    "    traintime150_start=time.time()\n",
    "    model.fit(metric[:tra_siz],nod_num, edge_index)\n",
    "    traintime150_end=time.time()\n",
    "    traintime150s.append(traintime150_end-traintime150_start)\n",
    "    \n",
    "    testtime_start=time.time()\n",
    "    pred = model.predict(metric[tra_siz:],nod_num, edge_index, sam_num, del_tol)[1]\n",
    "    testtime_end=time.time()\n",
    "    testtimes.append(testtime_end-testtime_start)\n",
    "    \n",
    "    \n",
    "    pred_t = model.predict(metric[:tra_siz], nod_num, edge_index, sam_num, del_tol)[1]\n",
    "    itp=model.interpret(metric[tra_siz:],nod_num, edge_index, sam_num, del_tol)\n",
    "    \n",
    "    ap = average_precision_score(data['label'][tra_siz+seq_len-del_tol:],pred)\n",
    "    aps.append(ap)\n",
    "    \n",
    "    mapp = np.mean((average_precision_score(data['label'][tra_siz+seq_len-del_tol:],pred), average_precision_score(1-data['label'][tra_siz+seq_len-del_tol:],-pred)))\n",
    "    maps.append(mapp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.44546379299195243, 0.446228654079362, 0.4429879274242494, 0.4463454411725116, 0.445277909737149]\n",
      "[0.7010292781148034, 0.7013511990329346, 0.7001372768209961, 0.7011541733236337, 0.7004709792910512]\n"
     ]
    }
   ],
   "source": [
    "print(aps)\n",
    "print(maps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7008285813166838"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(maps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.4452607450810449"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(aps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[3732.759813785553, 3788.3246517181396, 3635.5087356567383, 3523.102861881256, 3559.9291837215424]\n"
     ]
    }
   ],
   "source": [
    "print(traintime150s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "60.7987508225441"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(traintime150s)/60"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}