week5_ml_concepts.ipynb · kevin/teaching_AI

加入 Gitee

与超过 1200万开发者一起发现、参与优秀开源项目，私有仓库也完全免费：）

文件

该仓库未声明开源许可证文件（LICENSE），使用请关注具体项目描述及其代码上游依赖。

克隆/下载

week5_ml_concepts.ipynb 237.02 KB

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "%pylab inline\n",
    "from sklearn import datasets, metrics, model_selection, neighbors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = datasets.load_iris()\n",
    "X, y = data.data, data.target"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 使用 train_test_split 把所有样本按 7:3 的比例分为测试样本、训练样本"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#-samples in train, test sets are:  105 45\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = model_selection.train_test_split(X, y, test_size=0.3)\n",
    "\n",
    "print('#-samples in train, test sets are: ', len(X_train), len(X_test), )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### train_test_split 会把数据的顺序也随机打乱"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      " 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
      " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2\n",
      " 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2\n",
      " 2 2] \n",
      " [0 1 0 0 0 2 2 2 0 0 0 1 2 1 2 1 2 1 0 1 1 2 2 1 2 1 0 1 0 0 1 2 0 2 1 1 0\n",
      " 2 2 1 1 2 1 2 0 1 0 2 1 1 1 0 2 0 0 2 1 0 2 2 1 2 2 0 0 2 0 2 1 1 2 0 2 0\n",
      " 0 0 0 1 1 1 1 0 1 0 0 0 2 0 2 2 1 2 2 0 0 0 1 2 0 2 2 1 1 0 1]\n"
     ]
    }
   ],
   "source": [
    "print(y, '\\n', y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 使用 test sample 给模型打分"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy on the train sample:  0.9809523809523809\n",
      "Accuracy on the test  sample:  0.9777777777777777\n"
     ]
    }
   ],
   "source": [
    "model = neighbors.KNeighborsClassifier(n_neighbors=5)\n",
    "\n",
    "model.fit(X_train, y_train)\n",
    "\n",
    "print('Accuracy on the train sample: ', metrics.accuracy_score( model.predict(X_train), y_train))\n",
    "print('Accuracy on the test  sample: ', metrics.accuracy_score( model.predict(X_test), y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### k-折交叉检验（k-fold cross-validation），常取  cv = 5， 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/utils/deprecation.py:125: FutureWarning: You are accessing a training score ('train_score'), which will not be available by default any more in 0.21. If you need training scores, please set return_train_score=True\n",
      "  warnings.warn(*warn_args, **warn_kwargs)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'fit_time': array([0.0007    , 0.00043464, 0.00026917, 0.00051665, 0.00060582]),\n",
       " 'score_time': array([0.00139117, 0.00106096, 0.00105977, 0.00097752, 0.00078702]),\n",
       " 'test_score': array([0.96666667, 1.        , 0.93333333, 0.96666667, 1.        ]),\n",
       " 'train_score': array([0.96666667, 0.96666667, 0.975     , 0.975     , 0.96666667])}"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_selection.cross_validate(model, X,y, cv=5, scoring='accuracy' )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.96666667, 1.        , 0.93333333, 0.96666667, 1.        ])"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_selection.cross_val_score(model, X, y, cv=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9733333333333334"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_selection.cross_val_score(model, X, y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 验证曲线 （validation curve）选择超参数 n_neighbors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = neighbors.KNeighborsClassifier()\n",
    "\n",
    "train_scores, test_scores = model_selection.validation_curve(model, X, y,  \n",
    "                            param_name='n_neighbors', param_range=range(1,31), cv=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'cv score')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(15,5))\n",
    "ax.plot(range(1,31), [mean(xx) for xx in test_scores], \n",
    "        lw=2, marker='o', c='k', label = 'test_scores (fitting error + generalization error)')\n",
    "ax.plot(range(1,31), [mean(xx) for xx in train_scores], \n",
    "        lw=2,  marker='*', c='k', ls = '--', label = 'train scores (fitting error)')\n",
    "ax.legend(frameon=False, fontsize=16); ax.grid()\n",
    "ax.set_xlabel('n_neighbors', fontsize=14)\n",
    "ax.set_ylabel('cv score', fontsize=14)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 练习题：\n",
    "\n",
    "####      1. 使用 knn 拟合 digits, 绘制 n_neighbors = 1-31 的 train_score 以及 test_score 的 cross_validation curve. (你将看到 knn 的缓慢速度)\n",
    "\n",
    "####       2. 学习保存/读入你的模型。\n",
    "    \n",
    "    选出你认为最好的 knn 模型，保存到文件。再使用 joblib 读入，对 digits data 进行预测。\n",
    "    \n",
    "    提示：相关语句示例为，\n",
    "        import sklearn.externals.joblib as jb\n",
    "        jb.dump(model, 'knn_nb3')\n",
    "        model = jb.load('knn_nb3')\n",
    "\n",
    "#### 3. 自学神经网络的使用方法。写一个单层神经网络的模型学习 iris 数据。测试神经元个数对 performance 的影响。\n",
    "\n",
    "    已知多层感知机（Multiple Layers Perception, MLP）分类器的建模语句为\n",
    "        from sklearn import neural_network\n",
    "        model = neural_network.MLPClassifier( )\n",
    "    其中 model 的第一个选项\n",
    "        hidden_layer_sizes\n",
    "    代表需要多少层神经网络，每层几个神经元。例如\n",
    "        hidden_layer_sizes = (10, 15)\n",
    "    表示有两层神经元，个数分别为 10, 15\n",
    "    \n",
    "    问题：\n",
    "        请用单层 layer 的 nerual network 拟合 iris 数据库；使用 validation_curve 画出 accuracy score 随 hidden_layer_sizes 的变化。看多少个 neuron 可以满足问题需求（达到高准确度的预测结果）"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "########## 练习题答案在下面，除非你做完了，不要往下翻看\n",
    "########## 练习题答案在下面，除非你做完了，不要往下翻看\n",
    "########## 练习题答案在下面，除非你做完了，不要往下翻看\n",
    "########## 练习题答案在下面，除非你做完了，不要往下翻看\n",
    "########## 练习题答案在下面，除非你做完了，不要往下翻看\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = neighbors.KNeighborsClassifier()\n",
    "\n",
    "digits = datasets.load_digits()\n",
    "X, y = digits.data, digits.target\n",
    "\n",
    "train_scores, test_scores = model_selection.validation_curve(model, X, y,  \n",
    "                            param_name='n_neighbors', param_range=range(1,31), cv=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'cv score')"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(15,5))\n",
    "ax.plot(range(1,31), [mean(xx) for xx in test_scores], \n",
    "        lw=2, marker='o', c='k', label = 'test_scores (fitting error + generalization error)')\n",
    "ax.plot(range(1,31), [mean(xx) for xx in train_scores], \n",
    "        lw=2,  marker='*', c='k', ls = '--', label = 'train scores (fitting error)')\n",
    "ax.legend(frameon=False, fontsize=16); ax.grid()\n",
    "ax.set_xlabel('n_neighbors', fontsize=14)\n",
    "ax.set_ylabel('cv score', fontsize=14)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9661143537042125"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import sklearn.externals.joblib as jb\n",
    "model = neighbors.KNeighborsClassifier(n_neighbors=3)\n",
    "jb.dump(model, 'knn_nb3')\n",
    "model = jb.load('knn_nb3')\n",
    "model_selection.cross_val_score(model, X, y, cv=5).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n",
      "/home/xiaodongli/software/anaconda3/lib/python3.7/site-packages/sklearn/neural_network/multilayer_perceptron.py:562: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached and the optimization hasn't converged yet.\n",
      "  % self.max_iter, ConvergenceWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'accuracy score')"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn import neural_network\n",
    "nn = neural_network.MLPClassifier( )\n",
    "\n",
    "num_neurons = par_range = range(5,200,5)\n",
    "hidden_layer_sizes = [ num_neuron for num_neuron in num_neurons  ]\n",
    "\n",
    "iris = datasets.load_iris()\n",
    "train_scores, test_scores = model_selection.validation_curve(nn, iris.data, iris.target, \n",
    "                                                             'hidden_layer_sizes', num_neurons, cv=5)\n",
    "\n",
    "fig = plt.figure(figsize=(14,8))\n",
    "ax = fig.add_subplot(111)\n",
    "ax.plot(par_range, [mean(xx) for xx in test_scores], \n",
    "        label = 'cross-validation curve, test_scores (fitting error + generalization error), cv=5', lw=3)\n",
    "ax.plot(par_range, [mean(xx) for xx in train_scores], \n",
    "        label = 'cross-validation curve, train_scores (fitting error), cv=5', lw=3, ls = '--')\n",
    "\n",
    "ax.grid()\n",
    "ax.legend(fontsize=16, frameon=False)\n",
    "ax.set_xlabel('num_neuron', fontsize=20)\n",
    "ax.set_ylabel('accuracy score', fontsize=20)"
   ]
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}

一键复制编辑原始数据按行查看历史

提交于 2019-09-23 12:26 . minor changes

马建仓 AI 助手

尝试更多

代码解读

代码找茬

代码优化

https://gitee.com/huangxin1999/teaching_AI.git

git@gitee.com:huangxin1999/teaching_AI.git

huangxin1999

teaching_AI

master

kevin/teaching_AI

简介

发行版

贡献者

近期动态

kevin/teaching_AI .gitee-modal { width: 500px !important; }

简介

发行版

贡献者

近期动态

搜索帮助

kevin/teaching_AI