10.2 主成分分析.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# PCA方法\n",
    "- 函数原型: sklearn.decomposition.PCA(n_components=None, copy=True, whiten=False)\n",
    "- 官方文档: https://scikit-learn.org/stable/modules/generated/sklearn.decomposition.PCA.html?highlight=pca#sklearn.decomposition.PCA\n",
    "- 相关说明: https://www.jianshu.com/p/8642d5ea5389"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 0.41724458  0.90879423]\n",
      " [-0.90879423  0.41724458]]\n",
      "\n",
      "[[1.16076484 0.        ]]\n",
      "[[-0.96095731 -0.32068398]]\n",
      "\n",
      "[0.97468327 0.02531673]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Dimensionality Reduced')"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "from sklearn.decomposition import PCA\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "np.random.seed(0)\n",
    "\n",
    "X = np.linspace(-5, 5, 20)\n",
    "X = X.reshape((X.shape[0], 1))\n",
    "X = np.concatenate((X, 2*X), axis=1)\n",
    "X = X + np.random.normal(loc=0, scale=1, size=X.shape)\n",
    "\n",
    "\n",
    "# 计算关联系数矩阵\n",
    "# np.cov中的bias=False表示除以N-1而不是N\n",
    "R = np.cov(X.T, bias=False)\n",
    "# print(R)\n",
    "eigenvalues, eigenvectors = np.linalg.eig(R)\n",
    "order = eigenvalues.argsort()[::-1]\n",
    "eigenvalues_order = eigenvalues[order]\n",
    "eigenvectors = eigenvectors[order]\n",
    "eigenvalues[1] = -eigenvalues[1]\n",
    "\n",
    "# PCA训练器\n",
    "pca = PCA(n_components=2, copy=True, whiten=False)\n",
    "# 训练并降维\n",
    "pca.fit(X)\n",
    "x0 = [pca.components_[:, 0]]\n",
    "newX = pca.transform(x0)\n",
    "\n",
    "print(pca.components_)\n",
    "print()\n",
    "print(np.matmul(x0, pca.components_*eigenvalues))\n",
    "print(newX)\n",
    "# 将降维后数据逆变换回原数据\n",
    "pca.inverse_transform(newX)\n",
    "# 各个特征的方差百分比\n",
    "print()\n",
    "print(pca.explained_variance_ratio_)\n",
    "#print(pca.explained_variance_ratio_) # 即特征值百分比\n",
    "#print(pca.explained_variance_) # 即特征值\n",
    "\n",
    "newX = pca.transform(X)\n",
    "plt.figure(figsize=(10, 5))\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.scatter(X[:,0], X[:, 1], color='b')\n",
    "plt.xlim(xmin=-6, xmax=6)\n",
    "plt.ylim(ymin=-8, ymax=8)\n",
    "plt.title('Dimensionality Unreduced')\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.scatter(newX[:,0], newX[:, 1], color='r')\n",
    "plt.xlim(xmin=-10, xmax=10)\n",
    "plt.ylim(ymin=-7, ymax=7)\n",
    "plt.title('Dimensionality Reduced')"
   ]
  }
 ],
 "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}