FireAI_016_OutputClassConfidence.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "np.random.seed(37) # 使得每次运行得到的随机数都一样"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 1200 entries, 0 to 1199\n",
      "Data columns (total 3 columns):\n",
      "0    1200 non-null float64\n",
      "1    1200 non-null float64\n",
      "2    1200 non-null int64\n",
      "dtypes: float64(2), int64(1)\n",
      "memory usage: 28.2 KB\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "# 准备数据集\n",
    "data_path='E:\\PyProjects\\DataSet\\FireAI/data_multivar_imbalance.txt'\n",
    "df=pd.read_csv(data_path,header=None)\n",
    "# print(df.head()) # 没有问题\n",
    "print(df.info()) # 查看数据信息，确保没有错误\n",
    "dataset_X,dataset_y=df.iloc[:,:-1],df.iloc[:,-1]\n",
    "# print(dataset_X.head())\n",
    "# print(dataset_X.info())\n",
    "# print(dataset_y.head()) # 检查没问题\n",
    "dataset_X=dataset_X.values\n",
    "dataset_y=dataset_y.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 将整个数据集划分为train set和test set\n",
    "from sklearn.model_selection import train_test_split\n",
    "train_X, test_X, train_y, test_y=train_test_split(\n",
    "    dataset_X,dataset_y,test_size=0.25,random_state=42)\n",
    "\n",
    "# 看来直接使用简单的线性SVM分类器难以将class_0区分出来，故而我们要调整数据集中样本的数量权重\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.svm import SVC\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sample: [2.  1.5], probs: [0.08066588 0.91933412]\n",
      "sample: [8. 9.], probs: [0.08311977 0.91688023]\n",
      "sample: [4.8 5.2], probs: [0.14367183 0.85632817]\n",
      "sample: [4. 4.], probs: [0.06178594 0.93821406]\n",
      "sample: [2.5 7. ], probs: [0.21050117 0.78949883]\n",
      "sample: [7.6 2. ], probs: [0.07548128 0.92451872]\n",
      "sample: [5.4 5.9], probs: [0.45817727 0.54182273]\n"
     ]
    }
   ],
   "source": [
    "# 计算某个新样本的置信度\n",
    "new_samples=np.array([[2,1.5],\n",
    "                      [8,9],\n",
    "                      [4.8,5.2],\n",
    "                      [4,4],\n",
    "                      [2.5,7],\n",
    "                      [7.6,2],\n",
    "                      [5.4,5.9]])\n",
    "classifier3=SVC(kernel='rbf',probability=True) # 比上面的分类器增加了 probability=True参数\n",
    "classifier3.fit(train_X,train_y)\n",
    "\n",
    "# 使用训练好的SVM分类器classifier3对新样本进行预测，并给出置信度\n",
    "for sample in new_samples:\n",
    "    print('sample: {}, probs: {}'.format(sample,classifier3.predict_proba([sample])[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 将分类器绘制到图中\n",
    "def plot_classifier(classifier, X, y):\n",
    "    x_min, x_max = min(X[:, 0]) - 1.0, max(X[:, 0]) + 1.0 # 计算图中坐标的范围\n",
    "    y_min, y_max = min(X[:, 1]) - 1.0, max(X[:, 1]) + 1.0\n",
    "    step_size = 0.01 # 设置step size\n",
    "    x_values, y_values = np.meshgrid(np.arange(x_min, x_max, step_size),\n",
    "                                     np.arange(y_min, y_max, step_size))\n",
    "    # 构建网格数据\n",
    "    mesh_output = classifier.predict(np.c_[x_values.ravel(), y_values.ravel()])\n",
    "    mesh_output = mesh_output.reshape(x_values.shape) \n",
    "    plt.figure()\n",
    "    plt.pcolormesh(x_values, y_values, mesh_output, cmap=plt.cm.gray)\n",
    "    plt.scatter(X[:, 0], X[:, 1], c=y, s=80, edgecolors='black', \n",
    "                linewidth=1, cmap=plt.cm.Paired)\n",
    "    # specify the boundaries of the figure\n",
    "    plt.xlim(x_values.min(), x_values.max())\n",
    "    plt.ylim(y_values.min(), y_values.max())\n",
    "\n",
    "    # specify the ticks on the X and Y axes\n",
    "    plt.xticks((np.arange(int(min(X[:, 0])), int(max(X[:, 0])), 1.0)))\n",
    "    plt.yticks((np.arange(int(min(X[:, 1])), int(max(X[:, 1])), 1.0)))\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_classifier(classifier3,new_samples,classifier3.predict(new_samples))"
   ]
  },
  {
   "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}