├── CNN-LSTM-Attention.ipynb
├── README.md
└── data.csv


/CNN-LSTM-Attention.ipynb:
--------------------------------------------------------------------------------
  1 | {
  2 |  "cells": [
  3 |   {
  4 |    "cell_type": "code",
  5 |    "execution_count": 15,
  6 |    "metadata": {},
  7 |    "outputs": [],
  8 |    "source": [
  9 |     "#导入必要的库\n",
 10 |     "import numpy as np\n",
 11 |     "import matplotlib.pyplot as plt\n",
 12 |     "import pandas as pd\n",
 13 |     "from sklearn import preprocessing\n",
 14 |     "from sklearn.metrics import mean_squared_error\n",
 15 |     "from sklearn.metrics import mean_absolute_error\n",
 16 |     "from math import sqrt\n",
 17 |     "from keras.layers import *\n",
 18 |     "from keras.models import *\n",
 19 |     "from keras.optimizers import Adam"
 20 |    ]
 21 |   },
 22 |   {
 23 |    "cell_type": "code",
 24 |    "execution_count": 16,
 25 |    "metadata": {},
 26 |    "outputs": [
 27 |     {
 28 |      "data": {
 29 |       "text/html": [
 30 |        "<div>\n",
 31 |        "<style scoped>\n",
 32 |        "    .dataframe tbody tr th:only-of-type {\n",
 33 |        "        vertical-align: middle;\n",
 34 |        "    }\n",
 35 |        "\n",
 36 |        "    .dataframe tbody tr th {\n",
 37 |        "        vertical-align: top;\n",
 38 |        "    }\n",
 39 |        "\n",
 40 |        "    .dataframe thead th {\n",
 41 |        "        text-align: right;\n",
 42 |        "    }\n",
 43 |        "</style>\n",
 44 |        "<table border=\"1\" class=\"dataframe\">\n",
 45 |        "  <thead>\n",
 46 |        "    <tr style=\"text-align: right;\">\n",
 47 |        "      <th></th>\n",
 48 |        "      <th>open</th>\n",
 49 |        "      <th>low</th>\n",
 50 |        "      <th>close</th>\n",
 51 |        "      <th>high</th>\n",
 52 |        "    </tr>\n",
 53 |        "  </thead>\n",
 54 |        "  <tbody>\n",
 55 |        "    <tr>\n",
 56 |        "      <th>26272</th>\n",
 57 |        "      <td>7.1042</td>\n",
 58 |        "      <td>7.1042</td>\n",
 59 |        "      <td>7.1048</td>\n",
 60 |        "      <td>7.1055</td>\n",
 61 |        "    </tr>\n",
 62 |        "    <tr>\n",
 63 |        "      <th>26273</th>\n",
 64 |        "      <td>7.1047</td>\n",
 65 |        "      <td>7.1038</td>\n",
 66 |        "      <td>7.1038</td>\n",
 67 |        "      <td>7.1047</td>\n",
 68 |        "    </tr>\n",
 69 |        "    <tr>\n",
 70 |        "      <th>26274</th>\n",
 71 |        "      <td>7.1039</td>\n",
 72 |        "      <td>7.1039</td>\n",
 73 |        "      <td>7.1046</td>\n",
 74 |        "      <td>7.1047</td>\n",
 75 |        "    </tr>\n",
 76 |        "    <tr>\n",
 77 |        "      <th>26275</th>\n",
 78 |        "      <td>7.1048</td>\n",
 79 |        "      <td>7.1043</td>\n",
 80 |        "      <td>7.1045</td>\n",
 81 |        "      <td>7.1049</td>\n",
 82 |        "    </tr>\n",
 83 |        "    <tr>\n",
 84 |        "      <th>26276</th>\n",
 85 |        "      <td>7.1049</td>\n",
 86 |        "      <td>7.1049</td>\n",
 87 |        "      <td>7.1057</td>\n",
 88 |        "      <td>7.1057</td>\n",
 89 |        "    </tr>\n",
 90 |        "  </tbody>\n",
 91 |        "</table>\n",
 92 |        "</div>"
 93 |       ],
 94 |       "text/plain": [
 95 |        "         open     low   close    high\n",
 96 |        "26272  7.1042  7.1042  7.1048  7.1055\n",
 97 |        "26273  7.1047  7.1038  7.1038  7.1047\n",
 98 |        "26274  7.1039  7.1039  7.1046  7.1047\n",
 99 |        "26275  7.1048  7.1043  7.1045  7.1049\n",
100 |        "26276  7.1049  7.1049  7.1057  7.1057"
101 |       ]
102 |      },
103 |      "execution_count": 16,
104 |      "metadata": {},
105 |      "output_type": "execute_result"
106 |     }
107 |    ],
108 |    "source": [
109 |     "#设置LSTM的时间窗等参数\n",
110 |     "window=5\n",
111 |     "lstm_units = 16\n",
112 |     "dropout = 0.01\n",
113 |     "epoch=60\n",
114 |     "#读取数据\n",
115 |     "df1=pd.read_csv('data.csv') \n",
116 |     "df1=df1.iloc[:,2:]\n",
117 |     "df1.tail()"
118 |    ]
119 |   },
120 |   {
121 |    "cell_type": "code",
122 |    "execution_count": 17,
123 |    "metadata": {},
124 |    "outputs": [],
125 |    "source": [
126 |     "#进行数据归一化\n",
127 |     "from sklearn import preprocessing\n",
128 |     "min_max_scaler = preprocessing.MinMaxScaler()\n",
129 |     "df0=min_max_scaler.fit_transform(df1)\n",
130 |     "df = pd.DataFrame(df0, columns=df1.columns)\n",
131 |     "input_size=len(df.iloc[1,:])"
132 |    ]
133 |   },
134 |   {
135 |    "cell_type": "code",
136 |    "execution_count": 18,
137 |    "metadata": {},
138 |    "outputs": [
139 |     {
140 |      "name": "stdout",
141 |      "output_type": "stream",
142 |      "text": [
143 |       "(23644, 5, 4) (23644,) (2627, 5, 4) (2627,)\n"
144 |      ]
145 |     },
146 |     {
147 |      "name": "stderr",
148 |      "output_type": "stream",
149 |      "text": [
150 |       "E:\\anoconda\\lib\\site-packages\\ipykernel_launcher.py:5: FutureWarning: Method .as_matrix will be removed in a future version. Use .values instead.\n",
151 |       "  \"\"\"\n"
152 |      ]
153 |     }
154 |    ],
155 |    "source": [
156 |     "#构建lstm输入\n",
157 |     "stock=df\n",
158 |     "seq_len=window\n",
159 |     "amount_of_features = len(stock.columns)#有几列\n",
160 |     "data = stock.as_matrix() #pd.DataFrame(stock) 表格转化为矩阵\n",
161 |     "sequence_length = seq_len + 1#序列长度\n",
162 |     "result = []\n",
163 |     "for index in range(len(data) - sequence_length):#循环数据长度-sequence_length次\n",
164 |     "    result.append(data[index: index + sequence_length])#第i行到i+sequence_length\n",
165 |     "result = np.array(result)#得到样本，样本形式为6天*3特征\n",
166 |     "row = round(0.9 * result.shape[0])#划分训练集测试集\n",
167 |     "train = result[:int(row), :]\n",
168 |     "x_train = train[:, :-1]\n",
169 |     "y_train = train[:, -1][:,-1]\n",
170 |     "x_test = result[int(row):, :-1]\n",
171 |     "y_test = result[int(row):, -1][:,-1]\n",
172 |     "#reshape成 6天*3特征\n",
173 |     "X_train = np.reshape(x_train, (x_train.shape[0], x_train.shape[1], amount_of_features))\n",
174 |     "X_test = np.reshape(x_test, (x_test.shape[0], x_test.shape[1], amount_of_features))  \n",
175 |     "print(X_train.shape, y_train.shape, X_test.shape, y_test.shape)"
176 |    ]
177 |   },
178 |   {
179 |    "cell_type": "code",
180 |    "execution_count": 19,
181 |    "metadata": {},
182 |    "outputs": [
183 |     {
184 |      "name": "stdout",
185 |      "output_type": "stream",
186 |      "text": [
187 |       "Model: \"model_2\"\n",
188 |       "__________________________________________________________________________________________________\n",
189 |       "Layer (type)                    Output Shape         Param #     Connected to                     \n",
190 |       "==================================================================================================\n",
191 |       "input_2 (InputLayer)            (None, 5, 4)         0                                            \n",
192 |       "__________________________________________________________________________________________________\n",
193 |       "conv1d_2 (Conv1D)               (None, 5, 16)        80          input_2[0][0]                    \n",
194 |       "__________________________________________________________________________________________________\n",
195 |       "max_pooling1d_2 (MaxPooling1D)  (None, 1, 16)        0           conv1d_2[0][0]                   \n",
196 |       "__________________________________________________________________________________________________\n",
197 |       "dropout_2 (Dropout)             (None, 1, 16)        0           max_pooling1d_2[0][0]            \n",
198 |       "__________________________________________________________________________________________________\n",
199 |       "bilstm (Bidirectional)          (None, 32)           4224        dropout_2[0][0]                  \n",
200 |       "__________________________________________________________________________________________________\n",
201 |       "attention_vec (Dense)           (None, 32)           1056        bilstm[0][0]                     \n",
202 |       "__________________________________________________________________________________________________\n",
203 |       "multiply_2 (Multiply)           (None, 32)           0           bilstm[0][0]                     \n",
204 |       "                                                                 attention_vec[0][0]              \n",
205 |       "__________________________________________________________________________________________________\n",
206 |       "dense_2 (Dense)                 (None, 1)            33          multiply_2[0][0]                 \n",
207 |       "==================================================================================================\n",
208 |       "Total params: 5,393\n",
209 |       "Trainable params: 5,393\n",
210 |       "Non-trainable params: 0\n",
211 |       "__________________________________________________________________________________________________\n"
212 |      ]
213 |     }
214 |    ],
215 |    "source": [
216 |     "#建立LSTM模型 训练\n",
217 |     "inputs=Input(shape=(window, input_size))\n",
218 |     "model=Conv1D(filters = lstm_units, kernel_size = 1, activation = 'sigmoid')(inputs)#卷积层\n",
219 |     "model=MaxPooling1D(pool_size = window)(model)#池化层\n",
220 |     "model=Dropout(dropout)(model)#droupout层\n",
221 |     "model=Bidirectional(LSTM(lstm_units, activation='tanh'), name='bilstm')(model)#双向LSTM层\n",
222 |     "attention=Dense(lstm_units*2, activation='sigmoid', name='attention_vec')(model)#求解Attention权重\n",
223 |     "model=Multiply()([model, attention])#attention与LSTM对应数值相乘\n",
224 |     "outputs = Dense(1, activation='tanh')(model)\n",
225 |     "model = Model(inputs=inputs, outputs=outputs)\n",
226 |     "model.compile(loss='mse',optimizer='adam',metrics=['accuracy'])\n",
227 |     "model.summary()#展示模型结构"
228 |    ]
229 |   },
230 |   {
231 |    "cell_type": "code",
232 |    "execution_count": 20,
233 |    "metadata": {},
234 |    "outputs": [
235 |     {
236 |      "name": "stderr",
237 |      "output_type": "stream",
238 |      "text": [
239 |       "E:\\anoconda\\lib\\site-packages\\ipykernel_launcher.py:1: UserWarning: The `nb_epoch` argument in `fit` has been renamed `epochs`.\n",
240 |       "  \"\"\"Entry point for launching an IPython kernel.\n"
241 |      ]
242 |     },
243 |     {
244 |      "name": "stdout",
245 |      "output_type": "stream",
246 |      "text": [
247 |       "Train on 23644 samples, validate on 2627 samples\n",
248 |       "Epoch 1/60\n",
249 |       "23644/23644 [==============================] - 1s 58us/step - loss: 0.0993 - accuracy: 4.2294e-05 - val_loss: 0.0560 - val_accuracy: 0.0000e+00\n",
250 |       "Epoch 2/60\n",
251 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0585 - accuracy: 4.2294e-05 - val_loss: 0.0550 - val_accuracy: 0.0000e+00\n",
252 |       "Epoch 3/60\n",
253 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0561 - accuracy: 4.2294e-05 - val_loss: 0.0523 - val_accuracy: 0.0000e+00\n",
254 |       "Epoch 4/60\n",
255 |       "23644/23644 [==============================] - 0s 16us/step - loss: 0.0538 - accuracy: 4.2294e-05 - val_loss: 0.0501 - val_accuracy: 0.0000e+00\n",
256 |       "Epoch 5/60\n",
257 |       "23644/23644 [==============================] - 0s 18us/step - loss: 0.0516 - accuracy: 4.2294e-05 - val_loss: 0.0484 - val_accuracy: 0.0000e+00\n",
258 |       "Epoch 6/60\n",
259 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0493 - accuracy: 4.2294e-05 - val_loss: 0.0465 - val_accuracy: 0.0000e+00\n",
260 |       "Epoch 7/60\n",
261 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0470 - accuracy: 8.4588e-05 - val_loss: 0.0445 - val_accuracy: 0.0000e+00\n",
262 |       "Epoch 8/60\n",
263 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0446 - accuracy: 8.4588e-05 - val_loss: 0.0426 - val_accuracy: 0.0000e+00\n",
264 |       "Epoch 9/60\n",
265 |       "23644/23644 [==============================] - 0s 16us/step - loss: 0.0416 - accuracy: 8.4588e-05 - val_loss: 0.0400 - val_accuracy: 0.0000e+00\n",
266 |       "Epoch 10/60\n",
267 |       "23644/23644 [==============================] - 0s 18us/step - loss: 0.0382 - accuracy: 8.4588e-05 - val_loss: 0.0368 - val_accuracy: 0.0000e+00\n",
268 |       "Epoch 11/60\n",
269 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0344 - accuracy: 8.4588e-05 - val_loss: 0.0325 - val_accuracy: 0.0000e+00\n",
270 |       "Epoch 12/60\n",
271 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0297 - accuracy: 8.4588e-05 - val_loss: 0.0267 - val_accuracy: 0.0000e+00\n",
272 |       "Epoch 13/60\n",
273 |       "23644/23644 [==============================] - 0s 18us/step - loss: 0.0245 - accuracy: 8.4588e-05 - val_loss: 0.0196 - val_accuracy: 0.0000e+00\n",
274 |       "Epoch 14/60\n",
275 |       "23644/23644 [==============================] - 0s 16us/step - loss: 0.0194 - accuracy: 8.4588e-05 - val_loss: 0.0123 - val_accuracy: 0.0000e+00\n",
276 |       "Epoch 15/60\n",
277 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0141 - accuracy: 8.4588e-05 - val_loss: 0.0061 - val_accuracy: 0.0000e+00\n",
278 |       "Epoch 16/60\n",
279 |       "23644/23644 [==============================] - 0s 18us/step - loss: 0.0102 - accuracy: 8.4588e-05 - val_loss: 0.0028 - val_accuracy: 0.0000e+00\n",
280 |       "Epoch 17/60\n",
281 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0075 - accuracy: 8.4588e-05 - val_loss: 0.0016 - val_accuracy: 0.0000e+00\n",
282 |       "Epoch 18/60\n",
283 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0058 - accuracy: 8.4588e-05 - val_loss: 0.0012 - val_accuracy: 0.0000e+00\n",
284 |       "Epoch 19/60\n",
285 |       "23644/23644 [==============================] - 0s 16us/step - loss: 0.0051 - accuracy: 8.4588e-05 - val_loss: 0.0011 - val_accuracy: 0.0000e+00\n",
286 |       "Epoch 20/60\n",
287 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0046 - accuracy: 8.4588e-05 - val_loss: 0.0010 - val_accuracy: 0.0000e+00\n",
288 |       "Epoch 21/60\n",
289 |       "23644/23644 [==============================] - 0s 18us/step - loss: 0.0044 - accuracy: 8.4588e-05 - val_loss: 0.0010 - val_accuracy: 0.0000e+00\n",
290 |       "Epoch 22/60\n",
291 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0043 - accuracy: 8.4588e-05 - val_loss: 9.9080e-04 - val_accuracy: 0.0000e+00\n",
292 |       "Epoch 23/60\n",
293 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0041 - accuracy: 8.4588e-05 - val_loss: 8.8500e-04 - val_accuracy: 0.0000e+00\n",
294 |       "Epoch 24/60\n",
295 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0042 - accuracy: 8.4588e-05 - val_loss: 8.5094e-04 - val_accuracy: 0.0000e+00\n",
296 |       "Epoch 25/60\n",
297 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0040 - accuracy: 8.4588e-05 - val_loss: 9.0211e-04 - val_accuracy: 0.0000e+00\n",
298 |       "Epoch 26/60\n",
299 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0038 - accuracy: 8.4588e-05 - val_loss: 8.6092e-04 - val_accuracy: 0.0000e+00\n",
300 |       "Epoch 27/60\n",
301 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0039 - accuracy: 8.4588e-05 - val_loss: 7.9445e-04 - val_accuracy: 0.0000e+00\n",
302 |       "Epoch 28/60\n",
303 |       "23644/23644 [==============================] - 0s 18us/step - loss: 0.0037 - accuracy: 8.4588e-05 - val_loss: 7.8038e-04 - val_accuracy: 0.0000e+00\n",
304 |       "Epoch 29/60\n",
305 |       "23644/23644 [==============================] - 0s 16us/step - loss: 0.0037 - accuracy: 8.4588e-05 - val_loss: 7.7221e-04 - val_accuracy: 0.0000e+00\n",
306 |       "Epoch 30/60\n",
307 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0036 - accuracy: 8.4588e-05 - val_loss: 7.2390e-04 - val_accuracy: 0.0000e+00\n",
308 |       "Epoch 31/60\n",
309 |       "23644/23644 [==============================] - 0s 18us/step - loss: 0.0037 - accuracy: 8.4588e-05 - val_loss: 6.6263e-04 - val_accuracy: 0.0000e+00\n",
310 |       "Epoch 32/60\n",
311 |       "23644/23644 [==============================] - 0s 18us/step - loss: 0.0035 - accuracy: 8.4588e-05 - val_loss: 7.0413e-04 - val_accuracy: 0.0000e+00\n",
312 |       "Epoch 33/60\n",
313 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0034 - accuracy: 8.4588e-05 - val_loss: 6.8634e-04 - val_accuracy: 0.0000e+00\n",
314 |       "Epoch 34/60\n",
315 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0034 - accuracy: 8.4588e-05 - val_loss: 6.2003e-04 - val_accuracy: 0.0000e+00\n",
316 |       "Epoch 35/60\n",
317 |       "23644/23644 [==============================] - 0s 19us/step - loss: 0.0033 - accuracy: 8.4588e-05 - val_loss: 5.9458e-04 - val_accuracy: 0.0000e+00\n",
318 |       "Epoch 36/60\n",
319 |       "23644/23644 [==============================] - 0s 18us/step - loss: 0.0032 - accuracy: 8.4588e-05 - val_loss: 5.4686e-04 - val_accuracy: 0.0000e+00\n",
320 |       "Epoch 37/60\n",
321 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0033 - accuracy: 8.4588e-05 - val_loss: 5.2805e-04 - val_accuracy: 0.0000e+00\n",
322 |       "Epoch 38/60\n",
323 |       "23644/23644 [==============================] - 0s 16us/step - loss: 0.0031 - accuracy: 8.4588e-05 - val_loss: 5.1706e-04 - val_accuracy: 0.0000e+00\n",
324 |       "Epoch 39/60\n",
325 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0031 - accuracy: 8.4588e-05 - val_loss: 4.6490e-04 - val_accuracy: 0.0000e+00\n",
326 |       "Epoch 40/60\n",
327 |       "23644/23644 [==============================] - 0s 18us/step - loss: 0.0030 - accuracy: 8.4588e-05 - val_loss: 4.7443e-04 - val_accuracy: 0.0000e+00\n",
328 |       "Epoch 41/60\n",
329 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0030 - accuracy: 8.4588e-05 - val_loss: 4.1126e-04 - val_accuracy: 0.0000e+00\n",
330 |       "Epoch 42/60\n",
331 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0029 - accuracy: 8.4588e-05 - val_loss: 4.3156e-04 - val_accuracy: 0.0000e+00\n",
332 |       "Epoch 43/60\n",
333 |       "23644/23644 [==============================] - 0s 16us/step - loss: 0.0028 - accuracy: 8.4588e-05 - val_loss: 3.9280e-04 - val_accuracy: 0.0000e+00\n",
334 |       "Epoch 44/60\n",
335 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0029 - accuracy: 8.4588e-05 - val_loss: 3.3466e-04 - val_accuracy: 0.0000e+00\n",
336 |       "Epoch 45/60\n",
337 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0028 - accuracy: 8.4588e-05 - val_loss: 3.2231e-04 - val_accuracy: 0.0000e+00\n",
338 |       "Epoch 46/60\n",
339 |       "23644/23644 [==============================] - 0s 18us/step - loss: 0.0026 - accuracy: 8.4588e-05 - val_loss: 3.1442e-04 - val_accuracy: 0.0000e+00\n",
340 |       "Epoch 47/60\n",
341 |       "23644/23644 [==============================] - 0s 18us/step - loss: 0.0026 - accuracy: 8.4588e-05 - val_loss: 2.8463e-04 - val_accuracy: 0.0000e+00\n",
342 |       "Epoch 48/60\n",
343 |       "23644/23644 [==============================] - 0s 16us/step - loss: 0.0026 - accuracy: 8.4588e-05 - val_loss: 2.5236e-04 - val_accuracy: 0.0000e+00\n",
344 |       "Epoch 49/60\n",
345 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0026 - accuracy: 8.4588e-05 - val_loss: 2.5213e-04 - val_accuracy: 0.0000e+00\n",
346 |       "Epoch 50/60\n",
347 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0025 - accuracy: 8.4588e-05 - val_loss: 2.5187e-04 - val_accuracy: 0.0000e+00\n",
348 |       "Epoch 51/60\n",
349 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0025 - accuracy: 8.4588e-05 - val_loss: 2.2000e-04 - val_accuracy: 0.0000e+00\n",
350 |       "Epoch 52/60\n"
351 |      ]
352 |     },
353 |     {
354 |      "name": "stdout",
355 |      "output_type": "stream",
356 |      "text": [
357 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0024 - accuracy: 8.4588e-05 - val_loss: 2.1267e-04 - val_accuracy: 0.0000e+00\n",
358 |       "Epoch 53/60\n",
359 |       "23644/23644 [==============================] - 0s 16us/step - loss: 0.0024 - accuracy: 8.4588e-05 - val_loss: 1.9909e-04 - val_accuracy: 0.0000e+00\n",
360 |       "Epoch 54/60\n",
361 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0023 - accuracy: 8.4588e-05 - val_loss: 1.8918e-04 - val_accuracy: 0.0000e+00\n",
362 |       "Epoch 55/60\n",
363 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0022 - accuracy: 8.4588e-05 - val_loss: 1.9034e-04 - val_accuracy: 0.0000e+00\n",
364 |       "Epoch 56/60\n",
365 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0023 - accuracy: 8.4588e-05 - val_loss: 1.8040e-04 - val_accuracy: 0.0000e+00\n",
366 |       "Epoch 57/60\n",
367 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0022 - accuracy: 8.4588e-05 - val_loss: 1.7450e-04 - val_accuracy: 0.0000e+00\n",
368 |       "Epoch 58/60\n",
369 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0022 - accuracy: 8.4588e-05 - val_loss: 1.7322e-04 - val_accuracy: 0.0000e+00\n",
370 |       "Epoch 59/60\n",
371 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0022 - accuracy: 8.4588e-05 - val_loss: 1.6940e-04 - val_accuracy: 0.0000e+00\n",
372 |       "Epoch 60/60\n",
373 |       "23644/23644 [==============================] - 0s 17us/step - loss: 0.0021 - accuracy: 8.4588e-05 - val_loss: 1.6800e-04 - val_accuracy: 0.0000e+00\n"
374 |      ]
375 |     }
376 |    ],
377 |    "source": [
378 |     "history=model.fit(X_train, y_train, nb_epoch = epoch, batch_size = 256,shuffle=False,validation_data=(X_test, y_test)) #训练模型epoch次"
379 |    ]
380 |   },
381 |   {
382 |    "cell_type": "code",
383 |    "execution_count": 21,
384 |    "metadata": {},
385 |    "outputs": [
386 |     {
387 |      "data": {
388 |       "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAEICAYAAABRSj9aAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nO3deZgU1dX48e/pZRaYBWQRZSAgoLIJwQFxV0DFlWjwVRFXDGhC4pYo5pfEqDGvmkTiGiWKUfEFEaMhLiEqLokLMOyb4IAiI9uwMzBbd5/fH1U90zPMQAPN9HTN+TxPP9NVdav63GE499at6luiqhhjjPEuX7IDMMYYc3hZojfGGI+zRG+MMR5nid4YYzzOEr0xxnicJXpjjPE4S/SmURMRv4iUiEjHRhDLf0Xk+lQ7tjGW6E1CuUk5+oqISGnM8tUHejxVDatqlqp+ezjiTQQRuUZEVtWxPk1ENovI0EM8/u9E5G+HcgzTtFmiNwnlJuUsVc0CvgUujln3Su3yIhJo+CgT7nWgjYicVmv9BUAF8F7Dh2RMNUv0pkG5vdNXRWSyiOwCRorIySLyhYhsF5H1IvK4iATd8gERURHp5C5Pcre/KyK7RORzEelcz2f5RGSaiGxwj/2RiHSP2b7PY4nIUBFZISI7ROQxQOr6HFXdA0wDrq216VpgkqqGRaSViLwjIsUisk1E/iki7Q/+N1kVY08R+dit32IRuTBm20UistytW5GI3O6ub+vGsl1EtorIJ4cah2ncLNGbZLgU+D8gF3gVCAG3Aq2BU4GhwJh97D8C+DVwBM5ZwwP7KPsW0A1oBywBXo7nWCLSFid5j3PjKgJO2sfnvAj8j4hkuPu3BC4EXnK3+4C/Ah2B7wGVwGP7ON5+iUiaW7+3gTbA7cCrItLVLfICMEpVs4ETgI/d9b8AVrv7tMOpv/EwS/QmGf6rqv9U1YiqlqrqHFWdpaohVV0NTADO3Mf+01S1QFUrgVeAvnUVco//N1XdpaplwG+BE0WkeRzHughYoKpvuNv+BBTvI6ZPgK3AJe7ylcASVV3ixlLsHqtUVXcCv99PHeNxKpAG/EFVK1X1feBd97PBaUx6iEi2qm5V1Xkx648GOqpqhap+vNeRjadYojfJsDZ2QUSOF5G33SGWncD9OL3o+myIeb8HyKqrkHvHziMisto9bqG7KfbY9R3r6Ng4VTWC06uvkzqzA75M9fDNNTi9/GgszUXkORH51o1lJvuuYzyOBr7VmjMTrgGiQ0KX4jQ837rDVtEzkofcch+IyCoR+cUhxmEaOUv0JhlqT5n6LM6wSldVzQF+Qz3j4QfoWpwLooNwhomiQxrxHHs90CG6ICI+IG8/+7wEnCsipwD5wOSYbXcBnYEBbh0HxVOB/VgHdBCR2Pp0BL4DcM+SLgHa4gzxTHHX71TV21W1E/AD4G4ROdSzC9OIWaI3jUE2sAPY7V4s3df4/IEetxzYAjQDHjyAfd8C+orIMPfOoNtxxrTrpaqrgFk41x/eVdXYoZ5snDOGbSLSCqcxOxB+EcmIeaUDn+Fc37hTRIIiMginYZsqIpkiMkJEctyhp11AGEBELhaRLm4DscNdHz7AeEwKsURvGoM7getwktGzOBdoE+EFnF7vOmApTmKMi6puBK4A/oDTUHTESeL78yLOxdaXaq1/FOesYosbx7vxxuIaCZTGvFaoajlwMTAM2Aw8DoxQ1ZXuPtcBa9yholE4w0kAx+EMHZUAnwKPqep/DzAek0LEHjxijDHeZj16Y4zxOEv0xhjjcZbojTHG4yzRG2OMxzW6CaVat26tnTp1SnYYxhiTUubOnbtZVeu8BbjRJfpOnTpRUFCQ7DCMMSaliMia+rbZ0I0xxnicJXpjjPE4S/TGGONxjW6M3hjjHZWVlRQVFVFWVpbsUDwjIyODvLw8gsFg3PvElejdZ14+BviB51T1oVrbzwD+jPNwgytVdVrMtuuAX7mLv1PVFzHGNAlFRUVkZ2fTqVMnak6yaQ6GqrJlyxaKioro3LnOB6vVab9DNyLiB54Czgd6AFeJSI9axb4FrseZtS923yOAe3GezDMAuNd98o4xpgkoKyujVatWluQTRERo1arVAZ8hxTNGPwAoVNXVqlqBM6f1sNgCqvqNqi4CIrX2PQ94z326zTachyQPPaAIjTEpzZJ8Yh3M7zOeRN+emk8EKqL6CTYJ2VdERotIgYgUFBfv62lt9Vu3vZRH/72CrzfvPqj9jTHGq+JJ9HU1H/HObRzXvqo6QVXzVTW/TZt9PtuhXlt3V/D4zEJWbtx1UPsbY7xly5Yt9O3bl759+9KuXTvat29ftVxRURHXMW644QZWrFgR92c+99xz3HbbbQcb8mETz8XYImIeqYbzOLV1cR6/CDir1r4fxbnvAcnJcK5A7yytPByHN8akmFatWrFgwQIAfvvb35KVlcXPf/7zGmVUFVXF56u7z/vCCy8c9jgbQjw9+jlANxHpLCJpOE+Ynx7n8WfgPEOzpXsR9lx3XcLlZDpt1s6y0OE4vDHGIwoLC+nVqxc333wz/fr1Y/369YwePZr8/Hx69uzJ/fffX1X2tNNOY8GCBYRCIVq0aMG4cePo06cPJ598Mps2bYr7MydNmkTv3r3p1asXv/zlLwEIhUJcc801Vesff/xxAMaPH0+PHj3o06cPI0eOTEid99ujV9WQiIzFSdB+YKKqLhWR+4ECVZ0uIv2BN4CWwMUicp+q9lTVrSLyAE5jAXC/qm5NSOS1ZFuP3phG7b5/LmXZup0JPWaPo3O49+KeB7zfsmXLeOGFF3jmmWcAeOihhzjiiCMIhUKcffbZDB8+nB49at5cuGPHDs4880weeugh7rjjDiZOnMi4ceP2+1lFRUX86le/oqCggNzcXIYMGcJbb71FmzZt2Lx5M4sXLwZg+/btADzyyCOsWbOGtLS0qnWHKq5vxqrqO6p6rKp2UdUH3XW/UdXp7vs5qpqnqs1VtZWq9ozZd6KqdnVfh+08yO8TstMD7LBEb4zZjy5dutC/f/+q5cmTJ9OvXz/69evH8uXLWbZs2V77ZGZmcv755wNw4okn8s0338T1WbNmzWLQoEG0bt2aYDDIiBEj+OSTT+jatSsrVqzg1ltvZcaMGeTm5gLQs2dPRo4cySuvvHJAX4raF099MzYnM8jOMkv0xjRGB9PzPlyaN29e9f6rr77iscceY/bs2bRo0YKRI0fWeZ96Wlpa1Xu/308oFN8wcX3P5W7VqhWLFi3i3Xff5fHHH+f1119nwoQJzJgxg48//ph//OMf/O53v2PJkiX4/f4DrGFNnprrJjsjwM5SG6M3xsRv586dZGdnk5OTw/r165kxI7GXEQcOHMiHH37Ili1bCIVCTJkyhTPPPJPi4mJUlcsvv5z77ruPefPmEQ6HKSoqYtCgQfzhD3+guLiYPXv2HHIMnurR51qP3hhzgPr160ePHj3o1asXxxxzDKeeeuohHe/5559n2rSqWWAoKCjg/vvv56yzzkJVufjii7nwwguZN28eo0aNQlURER5++GFCoRAjRoxg165dRCIR7r77brKzsw+1ikh9pxXJkp+frwf74JEfvVTA2q17+NdtZyQ4KmPMwVi+fDndu3dPdhieU9fvVUTmqmp+XeU9NXSTkxG0u26MMaYWbyX6zIDdR2+MMbV4KtHnZgYpKQ8RCteeW80YY5ouTyX66DQIJeXWqzfGmChvJfpMJ9Hbl6aMMaaatxJ9hjvfjd1Lb4wxVTyV6HPdHr3dS2+MScQ0xQATJ05kw4YNdW4bOXIkb775ZqJCPmw89YWp6NCN3WJpjIlnmuJ4TJw4kX79+tGuXbtEh9hgPNWjtzF6Y0w8XnzxRQYMGEDfvn358Y9/TCQSqXPa4FdffZUFCxZwxRVXxH0mEIlEuOOOO+jVqxe9e/eu+pbsd999x2mnnUbfvn3p1asXn332Wb1TFSeat3r00TF6G7oxpvF5dxxsWJzYY7brDec/dEC7LFmyhDfeeIPPPvuMQCDA6NGjmTJlCl26dNlr2uAWLVrwxBNP8OSTT9K3b9+4jv/aa6+xbNkyFi5cSHFxMf379+eMM85g0qRJXHzxxdx9992Ew2FKS0uZO3dunVMVJ5qnEn1WegCf2MVYY0z93n//febMmUN+vjNbQGlpKR06dOC8886rmjb4ggsu4Nxzzz2o4//3v/9lxIgR+P1+2rVrx2mnnUZBQQH9+/dnzJgxlJWV8YMf/IA+ffrUmKr4UD5zfzyV6EXEpio2prE6wJ734aKq3HjjjTzwwAN7batr2uCDOX5dBg0axEcffcTbb7/N1VdfzT333MPVV1+dkM/cH0+N0YPzpSkbozfG1GfIkCFMnTqVzZs3A87dOd9++22d0wYDZGdns2vXrriPf8YZZzBlyhTC4TAbN27k008/JT8/nzVr1tCuXTtGjx7N9ddfz/z58+v9zETzVI8e3PluLNEbY+rRu3dv7r33XoYMGUIkEiEYDPLMM8/g9/v3mjYY4IYbbuCmm24iMzOT2bNn13gACcBNN93E2LFjAejcuTMff/wxX3zxBX369EFEePTRR2nbti0TJ07k0UcfJRgMkpWVxaRJk1i7dm2dn5lonpqmGODq576grDLC67ecksCojDEHw6YpPjya9DTFYFMVG2NMbZ5M9DZGb4wx1byX6DMDdteNMY1IYxseTnUH8/v0XKLPzQxSVhmhPBROdijGNHkZGRls2bLFkn2CqCpbtmwhIyPjgPbz4F03zjQIu8pCpGf5kxyNMU1bXl4eRUVFFBcXJzsUz8jIyCAvL++A9vFeos+ontisdVZ6kqMxpmkLBoN07tw52WE0eZ4busnJdNouuyBrjDEOzyX66jnpbb4bY4wBDyb62KEbY4wxXkz09pQpY4ypwXuJPsMePmKMMbE8l+gzgj7S/D6bk94YY1xxJXoRGSoiK0SkUETG1bE9XURedbfPEpFO7vqgiLwoIotFZLmI3JPY8OuM1b4da4wxMfab6EXEDzwFnA/0AK4SkR61io0CtqlqV2A8EJ1r83IgXVV7AycCY6KNwOFkE5sZY0y1eHr0A4BCVV2tqhXAFGBYrTLDgBfd99OAwSIigALNRSQAZAIVwM6ERL4P2Zk2sZkxxkTFk+jbA2tjlovcdXWWUdUQsANohZP0dwPrgW+BP6rq1tofICKjRaRARAoS8VXp3Myg3UdvjDGueBK91LGu9gxF9ZUZAISBo4HOwJ0icsxeBVUnqGq+qua3adMmjpD2LScjwC7r0RtjDBBfoi8COsQs5wHr6ivjDtPkAluBEcC/VLVSVTcBnwJ1PgElkewB4cYYUy2eRD8H6CYinUUkDbgSmF6rzHTgOvf9cGCmOvOSfgsMEkdzYCDwZWJCr1/04SM2NaoxxsSR6N0x97HADGA5MFVVl4rI/SJyiVvseaCViBQCdwDRWzCfArKAJTgNxguquijBddhLbmaQyrBSVhk53B9ljDGNXlzTFKvqO8A7tdb9JuZ9Gc6tlLX3K6lr/eEWncFyZ1klmWk2J70xpmnz3DdjwSY2M8aYWN5M9DaxmTHGVPFkoo/OSW9fmjLGGI8m+pwMd4zeJjYzxhiPJnobujHGmCreTPR2MdYYY6p4MtGnBXxkBv02Rm+MMXg00YNzL72N0RtjjJcTfYbNd2OMMeDlRG8TmxljDODhRJ9rDx8xxhjAw4k+J8PG6I0xBryc6G3oxhhjAC8nevcB4TYnvTGmqfNsos/NDBJRKCm34RtjTNPm2URfPSe9JXpjTNPm3URv0yAYYwzg5USfaYneGGPAw4k+t2oGSxu6McY0bZ5N9NGhG/vSlDGmqfNuoo9ejLVEb4xp4jyb6LMz7OEjxhgDHk70fp+QnW7TIBhjjGcTPTh33tgYvTGmqfN0os/OCNjQjTGmyfN0os/JDNrFWGNMk+fpRJ+bGbT76I0xTZ6nE310BktjjGnKvJ3oMwOW6I0xTZ63E31GkF3lIcIRm5PeGNN0xZXoRWSoiKwQkUIRGVfH9nQRedXdPktEOsVsO0FEPheRpSKyWEQyEhf+vkXnuymxcXpjTBO230QvIn7gKeB8oAdwlYj0qFVsFLBNVbsC44GH3X0DwCTgZlXtCZwFNNhYSnQGS7uX3hjTlMXTox8AFKrqalWtAKYAw2qVGQa86L6fBgwWEQHOBRap6kIAVd2iquHEhL5/ORnRh49YojfGNF3xJPr2wNqY5SJ3XZ1lVDUE7ABaAccCKiIzRGSeiNx16CHHz+akN8YYCMRRRupYV/vqZn1lAsBpQH9gD/CBiMxV1Q9q7CwyGhgN0LFjxzhCik/1nPSW6I0xTVc8PfoioEPMch6wrr4y7rh8LrDVXf+xqm5W1T3AO0C/2h+gqhNUNV9V89u0aXPgtahHdY/eLsYaY5queBL9HKCbiHQWkTTgSmB6rTLTgevc98OBmaqqwAzgBBFp5jYAZwLLEhP6/rVsFiToF96Y/x27yy3ZG2Oapv0menfMfSxO0l4OTFXVpSJyv4hc4hZ7HmglIoXAHcA4d99twKM4jcUCYJ6qvp34atStWVqA31/am1lfb2Hk87PYvqeioT7aGGMaDXE63o1Hfn6+FhQUJPSYM5Zu4KeT59OpVTNeuvEk2uU22K38xhjTINzrn/l1bfP0N2OjzuvZjr/d0J/vtpUy/JnP+Hrz7mSHZIwxDcY7iT4SgWXTIVRe5+ZTurRm8uiB7KkIc/kzn7F03Y4GDtAYY5LDO4n+m//A1Gvg0e7w71/DllV7FTkhrwVTx5xMmt/Hlc9+wWerNichUGOMaVjeSfSdToeRf4eOJ8PnT8ET/eClYbD0TYhUfxm3a9sspt1yCu1yM7h+4hzeWlT7TlFjjPEW7yR6nw+6DoYrX4Hbl8LZ/w82F8Jr18G0GyFcfXvl0S0ymXbzKfTpkMtPJ8/nb59+ncTAjTHm8PJOoo+VcxSceRfctgiG/BaWvQnTxzrj+K7cZkFeHnUSQ7ofyW//uYxH/vUlje0OJGOMSQRvJvoonx9Oux3O/hUsnAzv3AkxyTwj6OcvV/fjqgEdefqjVdz52kIqQpF9HNAYY1JPPHPdpL4zfg6Vu+G/4yGQCec9COJMzxPw+/j9pb1ol5PB+PdXsn57Gc+MPJHcZsEkB22MMYnh7R59lAgMvhdOuhm+eAo+fLDWZuHWId0Yf0Uf5q7ZxqV/+ZQ1W+xee2OMNzSNRA9Osh/6EPS7Fj75A3zwQI1hHIBLv5/HpJtOYuvuCi59+jPmrtmapGCNMSZxmk6iByfZX/Rn+P418J8/wrQboLK0RpEBnY/gjR+fSk5GgKv+OovpC+32S2NMamtaiR6cC7SXPAHn3O/cY//CBbBzfY0inVs3540fn0rfvBb8bPJ8nvl4ld2RY4xJWU0v0YPTsz/1Vrjy/6B4Bfx1EKxbUKNIy+ZpvHzTAC7uczQPvfslv/7HEkJhuyPHGJN6mmaijzr+Ahj1b6eXP3EoLHm9xub0gJ/HrujLzWd2YdIX33LzpLnsqbB57Y0xqaVpJ3qAdr3gRzPhqBOcb9D+fQyUbq/a7PMJ484/ngeG9WTml5u4asIXFO+qe+I0Y4xpjCzRA2S1hevfhjPHweLX4C+nwKoPaxS55uROPHtNPis27uKyv3zK6uKSJAVrjDEHxhJ9lD8IZ98DN70HwWbw8g/gnbugYk9VkXN6HMmU0SezpzzMD/9it18aY1KDJfra2p8IYz5xvlw1+1mYcCZsWFK1uW+HFvz9x6eQmxlkxF9n8a8lG5IYrDHG7J8l+rqkNYPzH4Zr3oSync5dOXOeq/qC1fdaNef1W06h+1E53PLKXF787JvkxmuMMftgiX5fupwNN/8XOp0Gb98JU6+tulDbKiudyT8ayODjj+Te6Uv533eWE4nYvfbGmMbHEv3+ZLWBq6c5X7Ba8Q48ezoUOQ8vz0zz8+w1JzJyYEee/WQ1d7++iLAle2NMI2OJPh4+n/MFqxtnOMsTz4PPnwZV/D7hgWG9uHVwN16bW8StU+ZTaV+sMsY0IpboD0RePoz5Dxw7FGbc4wzllO1ARLj9nGP55QXH89ai9dwyaS5lleH9H88YYxqAJfoDldkCrpgE5/4OvnwbJpwF6xcBMPqMLjzwg168v3wTo16cY9+iNcY0CpboD4YInPJT50tWlaXw/Dkw7yUArhn4Pf54eR8+X7WFa5+fzc6yyiQHa4xp6izRH4rvnewM5XQ4Cab/1Jk+obyE4Sfm8eSIfixYu50xL821xxMaY5LKEv2hymoD17wBZ90Di6c6X7Bav4gLeh/FI8NP4PPVW7hr2kKb5tgYkzSW6BPB54ezxsG106G8BJ4bArP/ymXfb88vzjuONxes4w8zViQ7SmNME2WJPpE6nw63fAqdz4B3fg5Tr+XHJ7fhqgEdefqjVUz6Yk2yIzTGNEGW6BOteWsYMbXqC1by8mU8MLQDg45vy2/+sYT3l21MdoTGmCbGEv3hEP2C1eUvwvoFBP7vcp78YRd6tc9l7OR5LFy7ff/HMMaYBIkr0YvIUBFZISKFIjKuju3pIvKqu32WiHSqtb2jiJSIyM8TE3aK6H4RXP43WDePZlOvZOJV3Wmdlc4tk+aybXdFsqMzxjQR+030IuIHngLOB3oAV4lIj1rFRgHbVLUrMB54uNb28cC7hx5uCup+MQyfCEUFtH5zBM9cfiybSyq4Y+oCmwTNGNMg4unRDwAKVXW1qlYAU4BhtcoMA150308DBouIAIjID4DVwNLEhJyCegyD4c9D0Rx6fXQTvx36PT5cUcyzn6xOdmTGmCYgnkTfHlgbs1zkrquzjKqGgB1AKxFpDtwN3LevDxCR0SJSICIFxcXF8caeWnpeCj98DtbO5qo1v+GiXm35479XMPtre0qVMebwiifRSx3rao851FfmPmC8qu7zAauqOkFV81U1v02bNnGElKJ6XQYX/hEpfI8/tZ5Oh5aZ/HTyPDaX2MPGjTGHTzyJvgjoELOcB6yrr4yIBIBcYCtwEvCIiHwD3Ab8UkTGHmLMqS3/Rsi/kfQvHmfSSd+ybU8lt7+6wOaxN8YcNvEk+jlANxHpLCJpwJXA9FplpgPXue+HAzPVcbqqdlLVTsCfgd+r6pMJij11DX0YOp5C3id38fiZ8J+vNvP0h4XJjsoY41H7TfTumPtYYAawHJiqqktF5H4RucQt9jzOmHwhcAew1y2YJkYgDf7nJWjWmvMW38mInhk89sFXLPluR7IjM8Z4kDS2ybby8/O1oKAg2WE0jHULYOJQQkf25vSNd5Cb1ZzpY08jLWDfYzPGHBgRmauq+XVts4ySTEf3hWFPEvhuNlM7vcWXG3bx5Myvkh2VMcZjLNEnW+/hMGAMHQpf4bbuO3nqo1UsLrIhHGNM4liibwwG/QqyjuSne56mbXM/d762gPKQPXPWGJMYlugbg4wcOO9B/BsX8bcTlrFyYwmPf2BDOMaYxLBE31j0+iF0Op3jlo7nhr7N+ctHq2yWS2NMQliibyxE4MI/QcUe7glM5sicDH4xbSGhsD1v1hhzaCzRNyZtjoOTf0LakimMP7mUlRtLmFpQlOyojDEpzhJ9Y3PmXZCTx0nLHmRAxxwefW8lu8tDyY7KGJPCLNE3NmnNYej/IpuW8WinL9hcUs4Em87YGHMILNE3Rt0vhq5DyFv0BJf1bMGET1azcWdZsqMyxqQoS/SNkQicOQ7KdvCr9gWEIhHGv7cy2VEZY1KUJfrGqkN/6HASRyx6nmtP6sDUgrWs2LAr2VEZY1KQJfrG7OSfwPY13J63kubpAf733eXJjsgYk4Is0Tdmx18ELTuRNe8Zxp7dlY9WFPNp4eZkR2WMSTGW6Bsznx8G/hiKZnN9x020b5HJg28vJ2JPozLGHABL9I1d36shI5f02U/z8/OOZdn6nfxr6YZkR2WMSSGW6Bu79CznObNfvsUlHSvp0qY5f35/pfXqjTFxs0SfCgaMAfHjn/0MPxvcjZUbS3h78fpkR2WMSRGW6FNBzlHO7JbzXuaibs3o2jaLxz74irD16o0xcbBEnypOGQuVu/HP/xu3DelG4aYS3lq0LtlRGWNSgCX6VNGuN3Q+E2Y9ywXdW3Pckdk8br16Y0wcLNGnkpPGwK71+L7+iFuHdGNV8W7+udB69caYfbNEn0q6ngOZLWHRqwzt2Y7j2zm9ens4iTFmXyzRp5JAGvS8DL58G19lCbcN6cbqzbuZbr16Y8w+WKJPNX2uhFApLJvOuT3a0f2oHOvVG2P2yRJ9qsnrDy07w6JX8fmE24Z045ste3hzgfXqjTF1s0SfakTghCvg609gx3ec2+NIeh6dwxMzrVdvjKmbJfpUdML/AAqLX0NEuG3IsazZsoe/z/8u2ZEZYxohS/SpqFUXyBsAi14FVYZ0b0uv9jk8ObOQSuvVG2NqsUSfqvpcAZuWwcYlTq9+8LF8u3UPb8yzXr0xpiZL9Kmq52XgC8LCKQAM7t6WE/JyeeLDr6xXb4ypIa5ELyJDRWSFiBSKyLg6tqeLyKvu9lki0sldf46IzBWRxe7PQYkNvwlrdgR0OxcWT4NI2B2r78baraX8fV5RsqMzxjQi+030IuIHngLOB3oAV4lIj1rFRgHbVLUrMB542F2/GbhYVXsD1wEvJypwgzN8U7IBvv4YgLOPa0ufvFyemFlIRch69cYYRzw9+gFAoaquVtUKYAowrFaZYcCL7vtpwGAREVWdr6rRG7yXAhkikp6IwA3Q7TxIz4WFrwJU3YFTtK2U161Xb4xxxZPo2wNrY5aL3HV1llHVELADaFWrzA+B+apaXvsDRGS0iBSISEFxcXG8sZtgBvT8ASz/J5SXAHDWcW3o06EFT1qv3hjjiifRSx3ras+Nu88yItITZzhnTF0foKoTVDVfVfPbtGkTR0imSp8roXI3rHgXoGqs/rvtpUyba716Y0x8ib4I6BCznAfU/r59VRkRCQC5wFZ3OQ94A7hWVVcdasCmlg4DIScPFr9WteqsY51e/dMf2X31xpj4Ev0coJuIdBaRNOBKYHqtMtNxLrYCDAdmqqqKSAvgbeAeVf00UUGbGD4f9B4Oqz6A3VsAp1d/6+CuFG0rtWP3UTAAABCCSURBVPvqjTH7T/TumPtYYAawHJiqqktF5H4RucQt9jzQSkQKgTuA6C2YY4GuwK9FZIH7apvwWjR1vS+HSAiWvVG16uzj2tK7fS5Pflhoc+AY08SJauN6FF1+fr4WFBQkO4zU8/TJkJ4Do2ZUrXpv2UZ+9FIBf7q8Dz88MS+JwRljDjcRmauq+XVts2/GekXv4bD2C9i2pmrVkO5t6XFUjvXqjWniLNF7Ra/hzs8l06pWiQg/G9yNrzfv5q1F65MUmDEm2SzRe0XL7zl34CyeVmP1uT2O5Ph22Twx8yvCkcY1TGeMaRiW6L3khMudGS03LKla5fMJPx3UjVXFu3lnsfXqjWmKLNF7SY9LwReocU89wPm92tGtbRZPzPyKiPXqjWlyLNF7SfNW0GWwO6Nl9cVXn0/46eBurNxYwtvWqzemybFE7zW9L4edRc4dODEu7H0Ux7fL5ndvL2NHaWWSgjPGJIMleq85/gIINoNFU2us9vuEh394AsW7ynno3eVJCs4YkwyW6L0mrTkcfyEsexNCFTU29enQgh+dfgyTZ6/ls8LNSQrQGNPQLNF7Ue/LoXQbFL6316bbzzmWzq2bc/ffF7GnIpSE4IwxDc0SvRd1GQTZR8Gc5/falBH089BlvVm7tZQ//XtlEoIzxjQ0S/Re5A9C/o3OjJZb9p4Z+qRjWjFyYEcmfvo1877dloQAjTENyRK9V/W7DnzBOnv1AHcPPZ6jcjK4a9oiykPhBg7OGNOQLNF7VfaR0OMSmD8JKnbvvTkjyO8v603hphL+OGNFEgI0xjQUS/Re1v9HUL5jr2/KRp11XFtGDuzIX//zNc98bA//MsarLNF7WceBcGRvmP0c1PPcgfsu6cVFJxzFQ+9+ycuff9Og4RljGoYlei8TgQE3wcbFsHZWnUX8PmH8FX0Z0r0tv/7HUnuguDEeZIne63pfDum5MHtCvUWCfh9PjujHaV1bc9e0hbxtc9cb4ymW6L0urTl8/2pYNh12bay3WEbQz4RrT6Rfx5bcOmU+/166oQGDNMYcTpbom4L+N0GkEua9uM9izdICTLyhPz2OzmH0y3MZ9/oiduyxCdCMSXWW6JuCVl2c6YsLJkJ434k7JyPIlNEDGXPGMbw2t4jBj37E9IXraGwPkTfGxM8SfVMx4Eewaz0s+8d+izZLC3DPBd2ZPvZUjm6Ryc8mz+f6F+awduueBgjUGJNo0th6avn5+VpQUJDsMLwnEoZnTneS/ZiPoUXHuHYLR5QXP/uGP/57BWWVYfp1bMmg7m0ZfPyRHHtkFiJymAM3xsRDROaqan6d2yzRNyGbC+GvZ8MRx8CNMyCYEfeu67aXMmXOWmZ+uZEl3+0EoH2LTM46rg09j87luHZZdDsym5yM4OGK3hizD5boTbUv34EpV0HfkTDsSede+wO0YUcZH67YxAfLN/H5qs3srqieK+eo3Ay6ts2idVY6ORkBcjOD5GQGyckIkp0RICsjQHZGkKz0ANnu9oygP5E1NKZJ2leiDzR0MCbJjr8AzrgLPnkE8k50Zrk8QO1yM7hqQEeuGtCRSET5bnspKzfuYuXGElZu3MWq4hK+2bKbnaUhdpZV1vel3CrpAR8tm6XRolmQFs2CpAf8VIYjhMJKRThCZTiCCDRPC5CV7jQWzd2Gok1WOm2ynVfb7HRaNksjrEp5ZYTyUITyUJiKUIT0gJ9sd7/m6X7SA9a4mKbDEn1TdNY4WDcP3rnLmSKhQ/+DPpTPJ3Q4ohkdjmjG4O5H7rU9ElFKKkLsLK1kV1mIkvIQu8qc97vKQuworWT7ngq276lke9X7CoJ+H0G/j+xggKDfh6qyuyLM+h1l7C4Osbs8xM7SEBXhSB1R7V/QLwT9PgQQEQRAIM3vIzPNT7M0P5lpAZoF/QQDPkJuwxOKRAhFnJYrelbinK0EycoIkB7wkR7wkRZw4k/zO+/TYtanB3z4fT78Ivh8zreT/SJkBJ3GKCs9QMBv90mYxLGhm6Zqz1aYcBaEK2DMJ5DVNtkRHTBVZWdZiOJd5c6rpJytJeUE/E4yTQ/63UQrlFdGKCl3GoiS8hAl5WHCkQiqoOD+VCrDEfZUhCmtCLO7IkxpRYjKsBL0C36f0zj4fc5wV4nbWO0sq27EEiXTTfpBv4/KcISKcISKkPNSoFman6z06BlKgOZpfgJ+HwGfOC+/EPD5yAz6aZbuNFzN0gJkBv2IOBfZI6qEIxBRJeBzGpqMoM/96Xf2dfdr5jZ+Ab+P0krn91NWGaa0MkxlKEK6Wz4zzf0Z9JOVEaj6XZnDz4ZuzN6aHQFXvgLPnePcjTPwFsi/ATJykx1Z3ESE3MwguZlBurbNSnY4qGqNhBz7vtx9RdeHIxHCkdiEq5RVhqvOdHaVVVJS7pyxpNc6OxCB3eVhdpeH2F3hNFp7ykPsrnAaL+fMw2m0St1Ga09lmHCk4Tt1WekBctzrMtkZAdKDztlMbIPk/O6qG1sAwdkWbViDbsNV9d5t1Pw+QRB84lxuEhF8ImQEfVUNT7ThEpxGTd1/K1XnbCo94K8603JeftKDzvsMt7Pg8wnq/juFIs5PxRl2DPik0d99Zj36pu7bWfDhg/D1x5CW7ST7gbdAztHJjswkULQRKq0Io+oMuUWHjHw+CIWdhqYsFKnqrZdVhtlTEWZPRcj9GSYUjji9dvfsIDPoJ+AXymP2i+5T4g6v7SqrrDrrqQxHqAw7ibIyXD0M5gyhOQlexEnI0QYrFI5Q6f50lt393TOyhuATqK+d9AluQ+E0GNGUH60POA2KzwcBnw+fuMtuo+S8d/5NTu/amjvOPe6gYrQevalfx5PguumwbgF89jh8/iR88Rfodg607Q6tj4XW3aBVN8jISXa05iCJOD3X+i5CpwegeXrqpYNIpLqHHnHPCCIRKA+Fq4aYSt1GyxHt/TspOBRR94wrXHXmVRGKUBYKV13QL3PPhvzuWYjf7/wE9jpbiz6tLXqGAk5M0SGycKT6FdHoq/rMLv0w3YEWV49eRIYCjwF+4DlVfajW9nTgJeBEYAtwhap+4267BxgFhIGfqeqMfX2W9eiTbNs38PnTsPpD2LoaIjHjzplHQFoWpDWDYDNnwrRgMwhmusvu+4B7f34kDBp2fyr4fOALxLz8TtlgJgSbVx8nkO5sE79bzr0wGQ45c/aEK524IuHq48QeN5AG/phXIB3E57zA7Wr5INr3EnHei8R8pv+gbj01JlkOqUcvIn7gKeAcoAiYIyLTVXVZTLFRwDZV7SoiVwIPA1eISA/gSqAncDTwvogcq6r2kNLGqmUnuOAR53240kn8m1c6r+1roXKP82jCyj1QsQdKNkBlGVSWQuVu92dpdTIVv5s0faCR6gSdCn8CvqDzoHV/MKbhcN+L323EQm4DFHLrJNWNiriNhz8N/OnOvoF0ZznakImv+vdT1WC5nxldrjqer7qBktqNlM/dJwj+mGOA27WM6WL6AjVj8QedY2jEfbnlYe9G1Oev/l3E/n6q6i11xEf1uhp/EzF/G1WNa8w6a2gTJp5ztQFAoaquBhCRKcAwIDbRDwN+676fBjwpztWJYcAUVS0HvhaRQvd4nycmfHNY+YPOsE3rbsCFiT22qpMcQ+VOoxFtOCpLIVRWfSYQbRRUayYwX9Dp6UeijYf7Clc6vf5QuXNHUbjCeV+V7NxEptHbMrVmIow2RtHjhCtj3le4yxVOmRoJMFB9xhD7GRqpjqMqnmhMEbduEaceGq7+rEi4+swFrZmEY5NxbL3ClVQl6KakrsYDqTphq254ajdGtRsmiaNxiR63rn1qN2619qt6K3uvi+p2Dpz34P5qfMDiSfTtgbUxy0XASfWVUdWQiOwAWrnrv6i1b/vaHyAio4HRAB07xjcHi0lxItW9wfTk3zHjGZFwzUYK9k5IkXBMw+M2ilDrTMTnNirhWg1pKKbRC7mNXmXNxrPqvVYvxzakVcN5kZrDe9GzouhQX1y01hChe8y6Pjs2Bmo3mFRvry/Z13WsGp9D9fJe+1Fze331y9krPSZEPIm+rlrXjrK+MvHsi6pOACaAM0YfR0zGmLr4oj3a+OcxMt4Xz9fvioAOMct5wLr6yohIAMgFtsa5rzHGmMMonkQ/B+gmIp1FJA3n4ur0WmWmA9e574cDM9W5nWc6cKWIpItIZ6AbMDsxoRtjjInHfodu3DH3scAMnNsrJ6rqUhG5HyhQ1enA88DL7sXWrTiNAW65qTgXbkPAT+yOG2OMaVj2zVhjjPGAfd1Hb1PkGWOMx1miN8YYj7NEb4wxHmeJ3hhjPK7RXYwVkWJgzSEcojWwOUHhJJuX6gLeqo+X6gLeqo+X6gLx1+d7qtqmrg2NLtEfKhEpqO/Kc6rxUl3AW/XxUl3AW/XxUl0gMfWxoRtjjPE4S/TGGONxXkz0E5IdQAJ5qS7grfp4qS7grfp4qS6QgPp4bozeGGNMTV7s0RtjjIlhid4YYzzOM4leRIaKyAoRKRSRccmO50CJyEQR2SQiS2LWHSEi74nIV+7PlsmMMV4i0kFEPhSR5SKyVERuddenan0yRGS2iCx063Ofu76ziMxy6/OqO413ShARv4jMF5G33OVUrss3IrJYRBaISIG7LiX/1gBEpIWITBORL93/Qycfan08kehjHmB+PtADuMp9MHkq+RswtNa6ccAHqtoN+MBdTgUh4E5V7Q4MBH7i/nukan3KgUGq2gfoCwwVkYHAw8B4tz7bgFFJjPFA3Qosj1lO5boAnK2qfWPuN0/VvzWAx4B/qerxQB+cf6dDq4+qpvwLOBmYEbN8D3BPsuM6iHp0ApbELK8AjnLfHwWsSHaMB1mvfwDneKE+QDNgHs5zkzcDAXd9jb/BxvzCedLbB8Ag4C2cR36mZF3ceL8BWtdal5J/a0AO8DXujTKJqo8nevTU/QDzw/OU3YZ1pKquB3B/tk1yPAdMRDoB3wdmkcL1cYc6FgCbgPeAVcB2VQ25RVLpb+7PwF2A+zRvWpG6dQHnOdT/FpG5IjLaXZeqf2vHAMXAC+7Q2nMi0pxDrI9XEn1cDyE3DUtEsoDXgdtUdWey4zkUqhpW1b44veEBQPe6ijVsVAdORC4CNqnq3NjVdRRt9HWJcaqq9sMZuv2JiJyR7IAOQQDoB/xFVb8P7CYBw05eSfRefQj5RhE5CsD9uSnJ8cRNRII4Sf4VVf27uzpl6xOlqtuBj3CuPbQQkejjOFPlb+5U4BIR+QaYgjN882dSsy4AqOo69+cm4A2chjhV/9aKgCJVneUuT8NJ/IdUH68k+ngeYJ6KYh+6fh3OWHejJyKC8xzh5ar6aMymVK1PGxFp4b7PBIbgXCD7EBjuFkuJ+qjqPaqap6qdcP6fzFTVq0nBugCISHMRyY6+B84FlpCif2uqugFYKyLHuasG4zxz+9Dqk+yLDwm8iHEBsBJn7PT/JTueg4h/MrAeqMRp1UfhjJ1+AHzl/jwi2XHGWZfTcE79FwEL3NcFKVyfE4D5bn2WAL9x1x8DzAYKgdeA9GTHeoD1Ogt4K5Xr4sa90H0tjf7fT9W/NTf2vkCB+/f2JtDyUOtjUyAYY4zHeWXoxhhjTD0s0RtjjMdZojfGGI+zRG+MMR5nid4YYzzOEr0xxnicJXpjjPG4/w/9jxdam4iLvgAAAABJRU5ErkJggg==\n",
389 |       "text/plain": [
390 |        "<Figure size 432x288 with 1 Axes>"
391 |       ]
392 |      },
393 |      "metadata": {
394 |       "needs_background": "light"
395 |      },
396 |      "output_type": "display_data"
397 |     }
398 |    ],
399 |    "source": [
400 |     "#迭代图像\n",
401 |     "loss = history.history['loss']\n",
402 |     "val_loss = history.history['val_loss']\n",
403 |     "epochs_range = range(epoch)\n",
404 |     "plt.plot(epochs_range, loss, label='Train Loss')\n",
405 |     "plt.plot(epochs_range, val_loss, label='Test Loss')\n",
406 |     "plt.legend(loc='upper right')\n",
407 |     "plt.title('Train and Val Loss')\n",
408 |     "plt.show()"
409 |    ]
410 |   },
411 |   {
412 |    "cell_type": "code",
413 |    "execution_count": 22,
414 |    "metadata": {},
415 |    "outputs": [
416 |     {
417 |      "data": {
418 |       "text/plain": [
419 |        "Text(0.5, 1.0, 'Train Data')"
420 |       ]
421 |      },
422 |      "execution_count": 22,
423 |      "metadata": {},
424 |      "output_type": "execute_result"
425 |     },
426 |     {
427 |      "data": {
428 |       "image/png": "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\n",
429 |       "text/plain": [
430 |        "<Figure size 864x432 with 1 Axes>"
431 |       ]
432 |      },
433 |      "metadata": {
434 |       "needs_background": "light"
435 |      },
436 |      "output_type": "display_data"
437 |     }
438 |    ],
439 |    "source": [
440 |     "#在训练集上的拟合结果\n",
441 |     "y_train_predict=model.predict(X_train)\n",
442 |     "y_train_predict=y_train_predict[:,0]\n",
443 |     "draw=pd.concat([pd.DataFrame(y_train),pd.DataFrame(y_train_predict)],axis=1)\n",
444 |     "draw.iloc[200:500,0].plot(figsize=(12,6))\n",
445 |     "draw.iloc[200:500,1].plot(figsize=(12,6))\n",
446 |     "plt.legend(('real', 'predict'),fontsize='15')\n",
447 |     "plt.title(\"Train Data\",fontsize='30') #添加标题"
448 |    ]
449 |   },
450 |   {
451 |    "cell_type": "code",
452 |    "execution_count": 23,
453 |    "metadata": {},
454 |    "outputs": [
455 |     {
456 |      "data": {
457 |       "text/plain": [
458 |        "Text(0.5, 1.0, 'Test Data')"
459 |       ]
460 |      },
461 |      "execution_count": 23,
462 |      "metadata": {},
463 |      "output_type": "execute_result"
464 |     },
465 |     {
466 |      "data": {
467 |       "image/png": "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\n",
468 |       "text/plain": [
469 |        "<Figure size 864x432 with 1 Axes>"
470 |       ]
471 |      },
472 |      "metadata": {
473 |       "needs_background": "light"
474 |      },
475 |      "output_type": "display_data"
476 |     }
477 |    ],
478 |    "source": [
479 |     "#在测试集上的预测\n",
480 |     "y_test_predict=model.predict(X_test)\n",
481 |     "y_test_predict=y_test_predict[:,0]\n",
482 |     "draw=pd.concat([pd.DataFrame(y_test),pd.DataFrame(y_test_predict)],axis=1);\n",
483 |     "draw.iloc[200:500,0].plot(figsize=(12,6))\n",
484 |     "draw.iloc[200:500,1].plot(figsize=(12,6))\n",
485 |     "plt.legend(('real', 'predict'),loc='upper right',fontsize='15')\n",
486 |     "plt.title(\"Test Data\",fontsize='30') #添加标题"
487 |    ]
488 |   },
489 |   {
490 |    "cell_type": "code",
491 |    "execution_count": 24,
492 |    "metadata": {},
493 |    "outputs": [
494 |     {
495 |      "name": "stdout",
496 |      "output_type": "stream",
497 |      "text": [
498 |       "训练集上的MAE/MSE/MAPE/涨跌准确率\n",
499 |       "0.022145526823359286\n",
500 |       "0.0009907685205786546\n",
501 |       "4.949584197505637\n",
502 |       "0.42456541048090346\n",
503 |       "测试集上的MAE/MSE/MAPE/涨跌准确率\n",
504 |       "0.010079520278647018\n",
505 |       "0.00016799508854128877\n",
506 |       "1.404783570780278\n",
507 |       "0.38918507235338917\n"
508 |      ]
509 |     }
510 |    ],
511 |    "source": [
512 |     "#输出结果\n",
513 |     "def mape(y_true, y_pred):\n",
514 |     "    return np.mean(np.abs((y_pred - y_true) / y_true)) * 100\n",
515 |     "def up_down_accuracy(y_true, y_pred):\n",
516 |     "    y_var_test=y_true[1:]-y_true[:len(y_true)-1]#实际涨跌\n",
517 |     "    y_var_predict=y_pred[1:]-y_pred[:len(y_pred)-1]#原始涨跌\n",
518 |     "    txt=np.zeros(len(y_var_test))\n",
519 |     "    for i in range(len(y_var_test-1)):#计算数量\n",
520 |     "        txt[i]=np.sign(y_var_test[i])==np.sign(y_var_predict[i])\n",
521 |     "    result=sum(txt)/len(txt)\n",
522 |     "    return result\n",
523 |     "print('训练集上的MAE/MSE/MAPE/涨跌准确率')\n",
524 |     "print(mean_absolute_error(y_train_predict, y_train))\n",
525 |     "print(mean_squared_error(y_train_predict, y_train) )\n",
526 |     "print(mape(y_train_predict, y_train) )\n",
527 |     "print(up_down_accuracy(y_train_predict,y_train))\n",
528 |     "print('测试集上的MAE/MSE/MAPE/涨跌准确率')\n",
529 |     "print(mean_absolute_error(y_test_predict, y_test))\n",
530 |     "print(mean_squared_error(y_test_predict, y_test) )\n",
531 |     "print(mape(y_test_predict,  y_test) )\n",
532 |     "print(up_down_accuracy(y_test_predict,y_test))"
533 |    ]
534 |   },
535 |   {
536 |    "cell_type": "code",
537 |    "execution_count": null,
538 |    "metadata": {},
539 |    "outputs": [],
540 |    "source": []
541 |   },
542 |   {
543 |    "cell_type": "code",
544 |    "execution_count": null,
545 |    "metadata": {},
546 |    "outputs": [],
547 |    "source": []
548 |   },
549 |   {
550 |    "cell_type": "code",
551 |    "execution_count": null,
552 |    "metadata": {},
553 |    "outputs": [],
554 |    "source": []
555 |   }
556 |  ],
557 |  "metadata": {
558 |   "kernelspec": {
559 |    "display_name": "Python 3",
560 |    "language": "python",
561 |    "name": "python3"
562 |   },
563 |   "language_info": {
564 |    "codemirror_mode": {
565 |     "name": "ipython",
566 |     "version": 3
567 |    },
568 |    "file_extension": ".py",
569 |    "mimetype": "text/x-python",
570 |    "name": "python",
571 |    "nbconvert_exporter": "python",
572 |    "pygments_lexer": "ipython3",
573 |    "version": "3.7.3"
574 |   }
575 |  },
576 |  "nbformat": 4,
577 |  "nbformat_minor": 2
578 | }
579 | 


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
1 | # CNN-LSTM-Attention
2 | 使用卷积神经网络-长短期记忆网络（bi-LSTM）-注意力机制对股票收盘价进行回归预测。The convolution neural network, short-term memory network and attention mechanism are used to predict the closing price.
3 | 


--------------------------------------------------------------------------------