MyAns.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 沪深300股回测分析\n",
    "\n",
    "通过tushare Pro取沪深300成份股最近一个月的行情数据，使用RNN算法根据每股收益和市值排序做分组收益率回测。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 导入需要的库"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tushare as ts\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import r2_score\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, LSTM, Dropout"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  设置我的token，获取从20230301-20230331的所有数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "pro = ts.pro_api(token='2cb7b30105056801542e9ff02a8370c509b84397686c94376801c170')\n",
    "hs300 = pro.index_weight(index_code='000300.SH', start_date='20230301', end_date='20230331')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "         ts_code trade_date     open     high      low    close  pre_close  \\\n",
      "13432  600519.SH 2023-03-31  1825.00  1848.00  1819.00  1820.00    1800.00   \n",
      "13620  600519.SH 2023-03-31  1825.00  1848.00  1819.00  1820.00    1800.00   \n",
      "13270  002756.SZ 2023-03-31    83.99    85.17    83.79    83.90      83.71   \n",
      "13754  002756.SZ 2023-03-31    83.99    85.17    83.79    83.90      83.71   \n",
      "13391  688169.SH 2023-03-31   370.00   379.99   364.00   365.00     371.48   \n",
      "...          ...        ...      ...      ...      ...      ...        ...   \n",
      "13539  000100.SZ 2023-03-31     4.44     4.45     4.36     4.43       4.46   \n",
      "13372  601989.SH 2023-03-31     3.75     3.78     3.74     3.78       3.77   \n",
      "13747  601989.SH 2023-03-31     3.75     3.78     3.74     3.78       3.77   \n",
      "13695  601236.SH 2023-03-31     8.61     8.67     8.50     8.63       8.82   \n",
      "13732  601236.SH 2023-03-31     8.61     8.67     8.50     8.63       8.82   \n",
      "\n",
      "       change  pct_chg         vol       amount         pe       pb  \\\n",
      "13432   20.00   1.1111    27446.33  5024790.356    36.4542  11.5757   \n",
      "13620   20.00   1.1111    27446.33  5024790.356    36.4542  11.5757   \n",
      "13270    0.19   0.2270    53383.15   450279.069     5.5054   2.8335   \n",
      "13754    0.19   0.2270    53383.15   450279.069     5.5054   2.8335   \n",
      "13391   -6.48  -1.7444     5651.29   209848.144    24.3836   3.7196   \n",
      "...       ...      ...         ...          ...        ...      ...   \n",
      "13539   -0.03  -0.6726  1555884.88   684915.539   289.4106   1.4923   \n",
      "13372    0.01   0.2653   585491.07   220498.298   391.9430   1.0235   \n",
      "13747    0.01   0.2653   585491.07   220498.298   391.9430   1.0235   \n",
      "13695   -0.19  -2.1542   431869.42   371378.511  1056.6077   1.7797   \n",
      "13732   -0.19  -2.1542   431869.42   371378.511  1056.6077   1.7797   \n",
      "\n",
      "           total_mv        eps       circ_mv  \n",
      "13432  2.286280e+08  49.925660  1.975068e+07  \n",
      "13620  2.286280e+08  49.925660  1.975068e+07  \n",
      "13270  3.479278e+06  15.239583  1.227908e+06  \n",
      "13754  3.479278e+06  15.239583  1.227908e+06  \n",
      "13391  3.419744e+06  14.969078  9.193849e+05  \n",
      "...             ...        ...           ...  \n",
      "13539  7.562848e+06   0.015307  5.067914e+06  \n",
      "13372  8.619169e+06   0.009644  8.421270e+06  \n",
      "13747  8.619169e+06   0.009644  8.421270e+06  \n",
      "13695  4.070588e+06   0.008168  2.287233e+06  \n",
      "13732  4.070588e+06   0.008168  2.287233e+06  \n",
      "\n",
      "[580 rows x 16 columns]\n"
     ]
    }
   ],
   "source": [
    "df = pd.DataFrame()\n",
    "for ts_code in hs300.con_code.tolist():\n",
    "    data = pro.daily(ts_code=ts_code, start_date='20230301', end_date='20230331')\n",
    "    df = pd.concat([df, data], ignore_index=True)\n",
    "\n",
    "df['trade_date'] = pd.to_datetime(df['trade_date'], format='%Y%m%d')\n",
    "df = df.sort_values(by='trade_date', ascending=True)\n",
    "\n",
    "df_basic = pro.daily_basic(ts_code=','.join(hs300.con_code.tolist()), trade_date='20230331',\n",
    "                            fields='ts_code,trade_date,pe,pb,total_mv')\n",
    "\n",
    "df_basic['trade_date'] = pd.to_datetime(df_basic['trade_date'])\n",
    "df_merged = pd.merge(df, df_basic, on=['ts_code', 'trade_date'], how='left')\n",
    "df_merged = df_merged.dropna(subset=['pe', 'pb', 'total_mv'])\n",
    "df_merged['eps'] = df_merged['close'] / df_merged['pe']\n",
    "df_merged['circ_mv'] = df_merged['total_mv'] / df_merged['pb']\n",
    "df_merged = df_merged.sort_values(by=['eps', 'circ_mv'], ascending=False)\n",
    "print(df_merged)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 设置分组大小，并进行分组"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "group_num = 10\n",
    "group_size = len(df_merged) // group_num"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 将时间序列数据转换为输入特征"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def time_series_to_features(df, time_steps):\n",
    "    features = []\n",
    "    for i in range(time_steps, len(df)):\n",
    "        features.append(df[i-time_steps:i, 0])\n",
    "    return np.array(features)\n",
    "X = df_merged[['eps', 'circ_mv']].values\n",
    "scaler = StandardScaler()\n",
    "X = scaler.fit_transform(X[:, [0]])\n",
    "time_steps = 30\n",
    "X_ts = time_series_to_features(X, time_steps)\n",
    "y_ts = df_merged['pct_chg'].values[time_steps:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 划分训练集和测试集"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_size = int(len(X_ts) * 0.8)\n",
    "X_train, X_test = X_ts[:train_size], X_ts[train_size:]\n",
    "y_train, y_test = y_ts[:train_size], y_ts[train_size:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 建立 RNN 模型"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(LSTM(units=64, return_sequences=True, input_shape=(X_ts.shape[1], 1)))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(LSTM(units=64, return_sequences=True))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(LSTM(units=64))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Dense(units=1))\n",
    "model.compile(optimizer='adam', loss='mean_squared_error')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 将输入数据reshape成LSTM模型需要的形状"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = np.reshape(X_train, (X_train.shape[0], X_train.shape[1], 1))\n",
    "X_test = np.reshape(X_test, (X_test.shape[0], X_test.shape[1], 1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 训练模型，迭代50次"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "14/14 [==============================] - 5s 92ms/step - loss: 7.0602 - val_loss: 4.7175\n",
      "Epoch 2/50\n",
      "14/14 [==============================] - 0s 25ms/step - loss: 6.9787 - val_loss: 4.4645\n",
      "Epoch 3/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 7.0295 - val_loss: 4.4330\n",
      "Epoch 4/50\n",
      "14/14 [==============================] - 0s 23ms/step - loss: 6.9848 - val_loss: 4.6607\n",
      "Epoch 5/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 7.0021 - val_loss: 4.6414\n",
      "Epoch 6/50\n",
      "14/14 [==============================] - 0s 26ms/step - loss: 6.9978 - val_loss: 4.5422\n",
      "Epoch 7/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.9813 - val_loss: 4.4928\n",
      "Epoch 8/50\n",
      "14/14 [==============================] - 0s 23ms/step - loss: 6.9935 - val_loss: 4.4456\n",
      "Epoch 9/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.9772 - val_loss: 4.5454\n",
      "Epoch 10/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.9664 - val_loss: 4.4430\n",
      "Epoch 11/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.9592 - val_loss: 4.3952\n",
      "Epoch 12/50\n",
      "14/14 [==============================] - 0s 25ms/step - loss: 6.9660 - val_loss: 4.4155\n",
      "Epoch 13/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.9136 - val_loss: 4.4314\n",
      "Epoch 14/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.9365 - val_loss: 4.4263\n",
      "Epoch 15/50\n",
      "14/14 [==============================] - 0s 26ms/step - loss: 7.0489 - val_loss: 4.4830\n",
      "Epoch 16/50\n",
      "14/14 [==============================] - 0s 25ms/step - loss: 6.9793 - val_loss: 4.4365\n",
      "Epoch 17/50\n",
      "14/14 [==============================] - 0s 28ms/step - loss: 6.9776 - val_loss: 4.4523\n",
      "Epoch 18/50\n",
      "14/14 [==============================] - 0s 25ms/step - loss: 6.9532 - val_loss: 4.3658\n",
      "Epoch 19/50\n",
      "14/14 [==============================] - 0s 22ms/step - loss: 6.9512 - val_loss: 4.4657\n",
      "Epoch 20/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.9251 - val_loss: 4.4046\n",
      "Epoch 21/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.9188 - val_loss: 4.3798\n",
      "Epoch 22/50\n",
      "14/14 [==============================] - 0s 23ms/step - loss: 6.9006 - val_loss: 4.4687\n",
      "Epoch 23/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.9053 - val_loss: 4.3931\n",
      "Epoch 24/50\n",
      "14/14 [==============================] - 0s 26ms/step - loss: 6.8813 - val_loss: 4.3564\n",
      "Epoch 25/50\n",
      "14/14 [==============================] - 0s 23ms/step - loss: 6.8811 - val_loss: 4.3843\n",
      "Epoch 26/50\n",
      "14/14 [==============================] - 0s 22ms/step - loss: 6.8795 - val_loss: 4.3933\n",
      "Epoch 27/50\n",
      "14/14 [==============================] - 0s 26ms/step - loss: 6.8911 - val_loss: 4.3426\n",
      "Epoch 28/50\n",
      "14/14 [==============================] - 0s 23ms/step - loss: 6.8891 - val_loss: 4.3597\n",
      "Epoch 29/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.8931 - val_loss: 4.4641\n",
      "Epoch 30/50\n",
      "14/14 [==============================] - 0s 25ms/step - loss: 6.8917 - val_loss: 4.4235\n",
      "Epoch 31/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.8837 - val_loss: 4.3802\n",
      "Epoch 32/50\n",
      "14/14 [==============================] - 0s 25ms/step - loss: 6.8717 - val_loss: 4.4146\n",
      "Epoch 33/50\n",
      "14/14 [==============================] - 0s 28ms/step - loss: 6.8687 - val_loss: 4.3752\n",
      "Epoch 34/50\n",
      "14/14 [==============================] - 0s 25ms/step - loss: 6.8786 - val_loss: 4.3610\n",
      "Epoch 35/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.8900 - val_loss: 4.4070\n",
      "Epoch 36/50\n",
      "14/14 [==============================] - 0s 26ms/step - loss: 6.8563 - val_loss: 4.3439\n",
      "Epoch 37/50\n",
      "14/14 [==============================] - 0s 25ms/step - loss: 6.8888 - val_loss: 4.3832\n",
      "Epoch 38/50\n",
      "14/14 [==============================] - 0s 28ms/step - loss: 6.8745 - val_loss: 4.3425\n",
      "Epoch 39/50\n",
      "14/14 [==============================] - 0s 32ms/step - loss: 6.8878 - val_loss: 4.3485\n",
      "Epoch 40/50\n",
      "14/14 [==============================] - 0s 30ms/step - loss: 6.9603 - val_loss: 4.3939\n",
      "Epoch 41/50\n",
      "14/14 [==============================] - 0s 30ms/step - loss: 6.8819 - val_loss: 4.3752\n",
      "Epoch 42/50\n",
      "14/14 [==============================] - 0s 25ms/step - loss: 6.8942 - val_loss: 4.3854\n",
      "Epoch 43/50\n",
      "14/14 [==============================] - 0s 24ms/step - loss: 6.9265 - val_loss: 4.3491\n",
      "Epoch 44/50\n",
      "14/14 [==============================] - 0s 23ms/step - loss: 6.9130 - val_loss: 4.3621\n",
      "Epoch 45/50\n",
      "14/14 [==============================] - 0s 25ms/step - loss: 6.9225 - val_loss: 4.3870\n",
      "Epoch 46/50\n",
      "14/14 [==============================] - 0s 28ms/step - loss: 6.9741 - val_loss: 4.3332\n",
      "Epoch 47/50\n",
      "14/14 [==============================] - 0s 31ms/step - loss: 6.9687 - val_loss: 4.3736\n",
      "Epoch 48/50\n",
      "14/14 [==============================] - 0s 29ms/step - loss: 6.9571 - val_loss: 4.4279\n",
      "Epoch 49/50\n",
      "14/14 [==============================] - 0s 26ms/step - loss: 6.9689 - val_loss: 4.3671\n",
      "Epoch 50/50\n",
      "14/14 [==============================] - 0s 25ms/step - loss: 6.9796 - val_loss: 4.3352\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train, y_train, epochs=50, batch_size=32, validation_data=(X_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 可视化训练集和验证集上的损失变化"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "epochs = range(1, len(loss) + 1)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(10,6))\n",
    "\n",
    "ax.plot(epochs, loss, 'bo', label='Training loss')\n",
    "ax.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "ax.set_title('Training and Validation Loss', fontsize=16)\n",
    "ax.set_xlabel('Epochs', fontsize=14)\n",
    "ax.set_ylabel('Loss', fontsize=14)\n",
    "ax.legend(fontsize=14)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 预测测试集数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "预测结果y_pre= [4.0711465 4.0779867 4.084519  4.090777  4.0968328 4.1027703 4.108698\n",
      " 4.114795  4.1211824 4.1279135 4.135009  4.142371  4.149884  4.157432\n",
      " 4.165022  4.1725755 4.180029  4.1872044 4.1941776 4.2011185 4.208091\n",
      " 4.2150555 4.221998  4.228931  4.2358313 4.2428107 4.2498875 4.2570815\n",
      " 4.264357  4.2718477 4.2795753 4.2876387 4.2961106 4.3048925 4.3138776\n",
      " 4.3229795 4.3321652 4.3412976 4.3503823 4.3594327 4.3684707 4.378201\n",
      " 4.3889046 4.400645  4.413359  4.4270906 4.441725  4.4572096 4.4733706\n",
      " 4.4899178 4.506709  4.523483  4.539993  4.5561714 4.5719633 4.5873613\n",
      " 4.602406  4.6171536 4.6316137 4.6458054 4.6598144 4.673525  4.686949\n",
      " 4.7001843 4.713225  4.7260485 4.7386556 4.751217  4.7637987 4.776452\n",
      " 4.789244  4.8018804 4.814666  4.827969  4.841929  4.856396  4.871357\n",
      " 4.886615  4.902076  4.9175887 4.933001  4.9484553 4.9639473 4.9798546\n",
      " 4.9964137 5.0136647 5.031618  5.0502653 5.0695724 5.0894747 5.109909\n",
      " 5.1308455 5.152147  5.1738496 5.195982  5.218333  5.2407    5.262802\n",
      " 5.2845683 5.30577   5.3263054 5.3460555 5.3649535 5.382808  5.3998723\n",
      " 5.4164834 5.4327855 5.4488683 5.4647846 5.4804635]\n"
     ]
    }
   ],
   "source": [
    "y_pred = model.predict(X_test)\n",
    "y_pred = scaler.inverse_transform(y_pred).ravel()\n",
    "print('预测结果y_pre=',sep=' ',end=' ')\n",
    "print(y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 计算R方值"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R方值: 0.008809590316457894\n"
     ]
    }
   ],
   "source": [
    "y_test = scaler.inverse_transform(y_test.reshape(-1, 1)).ravel()\n",
    "r2 = r2_score(y_test, y_pred)\n",
    "print('R方值:', r2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 计算每组收益率"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_merged['y_pred'] = np.nan\n",
    "df_merged.iloc[-len(y_pred):, df_merged.columns.get_loc('y_pred')] = y_pred.reshape(1, -1)[0]\n",
    "df_merged['group'] = pd.qcut(df_merged['eps']*df_merged['circ_mv'], group_num, labels=False)\n",
    "grouped = df_merged.groupby('group')\n",
    "group_ret = grouped['pct_chg'].mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 可视化回测结果"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "\n",
    "sns.set(style=\"whitegrid\")\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(10, 6))\n",
    "\n",
    "sns.barplot(x=group_ret.index, y=group_ret.values, ax=ax, color=\"lightblue\")\n",
    "\n",
    "ax.set_title(\"Grouped Returns\", fontsize=18)\n",
    "ax.set_xlabel(\"Group\", fontsize=14)\n",
    "ax.set_ylabel(\"Average Return\", fontsize=14)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 可视化模型预测结果的分布"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEXCAYAAACtTzM+AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAA+T0lEQVR4nO3daVyU9f7/8dcw7DuyCbiLiAluaO6YmpEdCQ1PmZb670SaxzR/ZVpqWrkvuXXqpI9Om1qamWV1FJdsEVc0ARV3VGQZFtlkG5jrf0OcI8kyIMOAfp53cuba3rM0b65r5vpeKkVRFIQQQjzwzEwdQAghRMMghSCEEAKQQhBCCFFGCkEIIQQghSCEEKKMFIIQQghACkGUSUxMpEOHDoSFhREWFkZoaChPPfUU27dv18+zevXqcrcr8sEHH7Bnz54Kp925fPv27cnMzKxRxpiYGN5++20AYmNjmTJlSo2Wr43S0lJefvllQkJC2LBhQ7lpa9eupVevXoSFhTF8+HBCQ0MZP348ly9fvqdtZmZm0r59ewD27t3L/Pnzq5x///79rF69usbbmTBhAtu2bSt337Vr1wgICCA1NfWu+UNDQ4mMjKx0fYMGDSI2NrbGOUTDYW7qAKLhsLa25vvvv9ffvn79OuPHj8fGxoaQkBCmTp1a7ToOHz6Mr69vhdMMWb4qFy5c0H9QBQYGsmbNmntanyFSU1P5448/+PPPP1Gr1XdNf+KJJ/QlBfDll1/y2muv3fVBW1uDBw9m8ODBVc4TGxtLdnZ2nWyvefPm9O3bl23btvHyyy/r7z9x4gS5ubnVZhGNmxSCqJSPjw9Tpkzhk08+ISQkhJkzZ9KuXTv+8Y9/sGbNGnbv3o2FhQUuLi4sWrSI3bt3ExcXx9KlS1Gr1ezdu5esrCyuXbvGI488QkZGhn55gFWrVhEbG4tOp+PVV19l4MCBbNu2jV27dvHxxx8D6G/PmzePNWvWkJuby5tvvsnw4cN57733+PHHH8nNzeWdd94hPj4elUpF//79+b//+z/Mzc0JDAzkpZde4sCBA2g0GsaOHcv48ePveqzHjh1j6dKlFBQUYGFhwauvvkq3bt148cUXKSkp4amnnmLt2rW0aNGiyuesd+/evP/++wA8//zzODk5cenSJZ599lmGDx/OggULOHfuHFqtlt69e/PGG29gbm5OZGQkK1euxMbGhoCAAP367nw+0tLSmDt3LpcuXcLMzIxRo0bRuXNnvv76a0pLS3FwcGDatGl88803fPXVV+h0OpydnZkzZw5t27YlNTWVmTNnotFo8Pb2JiMjo8LHMHr0aObPn8/EiRNRqVQAbNmyhWeeeYYbN27w9ttvk5GRQVpaGj4+PqxatQpXV1f98ocPH9a/NhXd/uijj4iMjESn0+Hj48PcuXPx9PQkMjKSjz76CJVKhVqt5o033qBHjx6GvFVFHZFDRqJK/v7+nDt3rtx9ycnJfP7553z77bds27aNvn37EhMTw5gxYwgICOCNN95gyJAhABQWFvLTTz8xffr0u9bdrFkzvvvuO5YtW8bMmTOrPITk5eXFlClT6N69O4sWLSo3bf78+Tg7O7Njxw6+/fZbzp49y3/+8x8AiouLcXFx4euvv2bNmjWsWLGCoqKicsvfuHGDKVOmMGvWLHbs2MGSJUuYPn06N27cYN26dfo9p+rKoKSkhK1bt9KzZ0/9fY6Ojvz88888//zzLFy4kI4dO7Jt2za2b9/OjRs3+PTTT0lPT+ett95i7dq1bNu2DR8fnwrX/84779CqVSt27tzJ5s2b2bJlC87OzowaNYonnniCadOmceTIEbZv387GjRvZvn07L774Iq+88goA7777Lp07d+ann35i9uzZlR7a6t+/P4qicOTIEQByc3PZu3cvTz/9ND/99BNdunRh8+bN7N279669yups376dc+fO8c033/D9998zYMAAZs+eDcDSpUuZO3cu27ZtY+rUqRw+fNjg9Yq6IXsIokoqlQpra+ty93l6euLv78+IESMIDg4mODiY3r17V7h8UFBQpet+9tlnAfDz86Nt27acOHGiVhl/++03vvrqK1QqFZaWlowaNYrPP/+cl156CUB/mKNjx44UFxeTn5+PlZWVfvmYmBhatGhB586dAWjXrh3dunXjyJEj5T7cK/Lzzz8THR0NgFarpWPHjrz33nv66d27d9f/e//+/cTGxrJ161bgVlkCREdH4+fnpz/U9swzz+j3Mu4UFRWlL1YHBwf9X9x32r9/P1euXGHUqFH6+7Kzs8nKyiIqKooZM2YA0LJly0of2+29j2+//ZaePXvyww8/EBwcjKurK+PGjePYsWN8+umnJCQkcP78ef3zZohffvmF2NhYwsPDAdDpdBQUFADwt7/9jcmTJzNgwAD69u1LRESEwesVdUMKQVQpNjYWPz+/cveZmZmxYcMGYmNjOXjwIAsXLqRnz576v/TuZGtrW+m6zcz+t4OqKArm5uaoVCruHF5Lq9VWm1Gn0911u6SkRH/79of/7cMffx2+66/L357nznVU5q/fIfzVnY9fp9OxevVq2rZtC0BOTg4qlYqDBw+Wy2RuXvH/lrefn9uuXbuGi4vLXY8lLCxMXxw6nQ6NRoOTk9Ndz21l2wEIDw/n8ccfJy8vjy1btvDOO+8AsGzZMmJiYggPD6dnz56UlJTc9XxW9RrqdDpefPFFRo8eDdzag7v9/ce0adMYOXIkf/zxB9u2bWPdunVs27at3PtEGJc806JSly9f5sMPP+SFF14od398fDzDhg2jbdu2TJgwgfHjx3P27FkA1Gq1QR+kAN999x0Ap06d4sqVK3Tu3JkmTZpw/vx5ioqKKCkp4ZdfftHPX9m6+/Xrx8aNG1EUheLiYrZs2UKfPn0MfpydO3fm8uXLxMTEAHD+/HmOHj3Kww8/bPA6DNGvXz8+++wzfc6XX36ZDRs20L17dy5cuEB8fDxApV9I9+7dm2+//Ra4dRhn3LhxJCQklHte+vbty08//YRGowHgq6++Yty4ccCtQ0GbN28GICkpqcpDMi4uLgwcOJA1a9agVqvp0qULAH/88Qfjxo1j+PDhuLq6EhUVRWlpabllmzRpQlJSEhkZGSiKUu5XZ/369WPr1q3k5eUBt3559sYbb1BSUsKgQYPIz8/n2WefZe7cuVy8eNHg95KoG7KHIPQKCwsJCwsDbv31bmVlxf/93//xyCOPlJvP39+foUOHEh4ejq2tLdbW1vq9g4EDB7JkyRKD/rK/du0aw4cPR6VS8f777+Ps7Ezfvn3p0aMHQ4cOxd3dnZ49e+rLpmvXrqxatYp//vOfjB07Vr+e2bNnM3/+fEJDQ9FqtfTv35+JEyca/LibNGnC6tWree+99ygsLESlUrFo0SJat25NYmKiweupzqxZs1iwYIE+Z58+fXjxxRexsLBg+fLlvP7661hYWFT6Rerbb7/NvHnzCA0NRVEUJkyYQEBAAFqtlldeeQULCwvmzJlDREQEL7zwAiqVCnt7ez744ANUKhVz587lzTffZOjQoTRt2hR/f/8q844ePZqnn36aBQsW6O/75z//ydKlS/nwww9Rq9V069aNq1evllvO19eXUaNGER4ejru7e7n3z9///ndSU1N5+umnUalUeHl5sXjxYszNzXnrrbd4/fXX9XtCCxcuxNLSsvZPuKgxlQx/LYQQAuSQkRBCiDJSCEIIIQApBCGEEGWMWgh5eXkMGzasyi/m9u/fz6BBg4wZQwghhAGMVggnT57k2WefJSEhodJ50tPTWbJkSY3XrSgKRUVFd/3+WQghRO0ZrRC2bNnC3Llz8fDwqHSe2bNnM3ny5Bqvu7i4mLi4OIqLi+8lYr05deqUqSMYrDFlhcaVV7Iah2StO0b/2emgQYP44osvaNasWbn7v/jiC7KysnjqqacYO3Ys+/btM3idRUVFxMXF1XVUIYS471U1nIxJTkw7d+4ckZGRfPbZZ6SkpNR6PQEBAeXGpGmooqOjq3wRGpLGlBUaV17JahySte6YpBB27txJWloa4eHhaLVaNBoNo0ePZtOmTaaII4QQAhMVwpQpU/RXu0pMTGTs2LFSBkIIYWL1eh5CRESEXGJPCCEaKKPvIdz5ZfH69evvmt6sWbMafaEshBDCOORMZSGEEIAUghBCiDJSCEIIIYAHtBBKS+++ZKIx1eR3x/WdTQghbnsgr5imVpuxaVd8vW0vKSkJb29vg+YdHVL1VayEEMJYHsg9BCGEEHeTQhBCCAFIIQghhCgjhSCEEAKQQhBCCFFGCkEIIQQghSCEEKKMFIIQQghACkEIIUQZKQQhhBCAFIIQQogyUghCCCEAKQQhhBBlpBCEEEIAUghCCCHKSCEIIYQA6qEQ8vLyGDZsGImJiXdN27NnD2FhYTz55JNMmjSJ7OxsY8cRQghRCaMWwsmTJ3n22WdJSEi4a1peXh7z5s1j3bp1/PDDD7Rv3561a9caM44QQogqGLUQtmzZwty5c/Hw8LhrmlarZd68eXh6egLQvn17kpOTjRlHCCFEFYx6TeUFCxZUOs3FxYVHH30UgMLCQtatW8fzzz9vzDhCCCGqYNRCMERubi6TJk3C39+fESNG1GjZuLi4Wm0zKCiIpKSkWi1bW4Zvz5/o6GijZqmOqbdfU40pr2Q1DslquKCgoEqnmbQQNBoN//jHP+jVqxdvvfVWjZcPCAjAysqqVtv29vau1XK1kZSUVKPtVfWCGVt0dLRJt19TjSmvZDUOyVp3TFYIpaWlTJw4kaFDhzJp0iRTxRBCCFGm3gshIiKCKVOmkJKSwunTpyktLWXXrl3Arb/4q/reQQghhPHUSyHs27dP/+/169cDEBgYSHx8fH1sXgghhAHkTGUhhBCAFIIQQogyUghCCCEAKQQhhBBlpBCEEEIAUghCCCHKSCEIIYQApBCEEEKUkUIQQggBSCEIIYQoI4UghBACkEIQQghRRgpBCCEEIIUghBCijBSCEEIIQApBCCFEGSkEIYQQgBSCEEKIMlIIQgghACkEIYQQZaQQhBBCAFIIQgghyhi9EPLy8hg2bBiJiYl3TTtz5gzh4eGEhIQwa9YsSkpKjB1HCCFEJYxaCCdPnuTZZ58lISGhwunTp09nzpw57Nq1C0VR2LJlizHjCCGEqIJRC2HLli3MnTsXDw+Pu6Zdv36dwsJCunTpAsBTTz3Fzp07jRlHCCFEFcyNufIFCxZUOk2j0eDu7q6/7e7uTmpqqjHjCCGEqIJRC6EqiqLcdZ9KparROuLi4mq17aCgIJKSkmq1bG0Zvj1/oqOjjZqlOqbefk01pryS1Tgkq+GCgoIqnWayQvD09CQ9PV1/Oy0trcJDS1UJCAjAysqqVtv39vau1XK1kZSUVKPtVfWCGVt0dLRJt19TjSmvZDUOyVp3TPazUx8fH6ysrPRtuX37doKDg00VRwghHnj1XggRERHExsYCsHz5chYtWsTQoUMpKChg7Nix9R2nwSkt1Zl0+5X99WLqXEII46uXQ0b79u3T/3v9+vX6f/v7+7N169b6iNBoqNVmbNoVb7LtV3Z4a3SIvwnSCCHqk5ypLIQQApBCEEIIUUYKQQghBCCFIIQQoowUghBCCMCEJ6aJqimKgqKASlXzM7iFEKI2pBAaAG1JKSkZ+WhuFBB7MZ2LiVkUFpVye3APFWBtZY6DrSWOdha4OtnQ1NUOW2t5+YQQdUc+UUxEURRSMvK5kJhFUvpNdDoFtZmKds2d8XKzx8bKHDMVKMqteW8WlpCbX8yVlFwuJGYD4GRvSXMPB9r4OGFnY2HiRySEaOykEOqZoigkp9/k5Pl0svKKsLJU49vMmeYe9rg62/D80A5VnpimKApZeUUkp98kOf0mcZcyiLuUgZerHe1aOOPtZieHmIQQtSKFUI9u5BZy4mwaqZn52NlY0CugKS2aOqA2M/y7fZVKhYuDNS4O1jzU2pW8Ai2Xrmdz6Xo2v524jouDFQFtXfFxt5diEELUiBRCPVAUhTMJmcScT8PCXE239h74NndGbXbvH9j2NhZ08nUjoI0rCck5nLqUwe9/JtHE0Ypu7T1wd7Gtg0cghHgQSCEYWX6hlpiEArJu5tHMw56HH2qKlaW6zrdjZqaijY8TrbwcSUjOIeZCOnuOXqNlUwc6+7ljZy3fMQghqiaFYESZOYX8ejyR4pJSHu7YlDbejkY/jHO7GFp4OnA6IZP4hEyup+XRuZ077Zo7y2EkIUSl5MQ0I0lKy2PPkauozVR0a2NLWx+nev0wNjc3o5OvG0/0bY2bsw3R8Rr2Hr1Gzs3iessghGhcpBCM4GJiFr/9eR1HO0uG9GyJnXXdHyIylL2NBY90a0bPjk3Jziti58EEzl+9UeElTIUQDzY5ZFTHLl3P5sjpVJq62tKvsw8W5mbcMHEmlerWYSQvNzsOx6VwLF5DcsZNHu7YFGtLeQsIIW6RPYQ6lJCcw+FTKTR1tSW4y60yaEhsrMwZ0M2Hbu09SE7P579RCSSn3zR1LCFEA9GwPrEasURNLofiknF3saF/Fx/U6ob51KpUKtq3dOGxXi2wslCz/3giJ85q0OnkEJIQD7qG+anVyKRnFXAgJpkmjtYM6NoM8wZaBndycbDmsV4t8W3mTPyVG+w7do2CohJTxxJCmFDD/+Rq4G4Wavn9z+vYWpkzoGvDO0xUFXO1GT0e8qR3oBeZOYXsPJhAVp6UghAPqsbz6dUAlZTo+O3EdUp1CsFdfbBqpF/QtvJyJKRXSyzM1ZxMKOD05Qz5FZIQDyAphFpSFIWDcclk5xbRJ9ALJ3srU0e6J072VoT0aom7kzknz6fz+59JFGtLTR1LCFGPjFoIO3bs4IknnmDIkCFs3LjxrumnTp0iPDycJ598kgkTJpCTk2PMOHXq7NUbJGry6OLnjre7vanj1AkLczM6NLMmyN+D5PQ8dh66QmZOoaljCSHqiUGF8MorrxAVFVWjFaemprJy5Uo2bdrE999/z+bNm7lw4UK5eRYsWMCUKVP44YcfaN26NZ988kmNtmEqGdkFnDyXho+7Pe1bupg6Tp1SqVT4tXBhcI8WKIrC7iNXuZCYJYeQhHgAGFQIjz32GB9++CEhISF88sknZGVlVbtMVFQUvXr1wtnZGVtbW0JCQti5c2e5eXQ6HTdv3vodfEFBAdbW1jV/BPWsWFvKgZhkrK3M6RnQ9L4dG8jN2YbHe7XEw8WGo6dTWfX1CQrlV0hC3NcMKoTQ0FA2bNjAhx9+SEZGBn//+9+ZPn06MTExlS6j0Whwd3fX3/bw8CA1NbXcPDNnzmTWrFn069ePqKgoRo0aVcuHUT8UReHI6VTyC7X07eSNlYXphqSoD1aW5gzo1ozAtq78En2N19b8RqIm19SxhBBGYvDPYnQ6HVeuXCEhIYGSkhJcXV2ZN28evXv3Zvr06XfNX9Ehhjv/mi4sLGTWrFl8/vnndOrUiU8//ZQZM2awbt06g8PHxcUZPO+dgoKCSEpKqvFyqVlarqUW0trTkuL8GyTlGz4oheHb869Vtrr01+03sYF3Inqz+PPDTF3xC0/2dCGgZcO5zkJ0dLSpIxhMshqHZDVcUFBQpdMMKoSVK1eybds2mjdvzujRo1m9ejUWFhbk5+czcODACgvB09OTY8eO6W9rNBo8PDz0t8+dO4eVlRWdOnUC4JlnnmH16tUGPyiAgIAArKxq9+seb2/vGs2fX1hCVPxl3JytebhTC8xqcKgoKSmpRturaba6VFnWru09+Ncbj7L0y2NsPZBJIU688GRHLMxNu5cUHR1d5Ru8IZGsxiFZ645Bh4wyMzNZv349mzZtYtiwYVhY3LrYiq2tLStWrKhwmT59+nDw4EEyMzMpKCggMjKS4OBg/fSWLVuSkpLCpUuXANi7dy+BgYH3+niMQlEUjp5OQadT6NXRq0ZlcD9xc7Zh4aS+DB/Qlh8PXGbmv/5Ak5lv6lhCiDpiUCGUlpbi7+9f7r5XXnkFgH79+lW4jKenJ9OmTWPs2LEMHz6cYcOG0alTJyIiIoiNjcXJyYlFixbx6quvEhoayrfffsvChQvv8eEYx6WkHJLSb9K5nTsOdpamjmNS5moz/vFkAG+O60GiJo8p7+/nj5PXTR1LCFEHqjxkNHfuXFJTU4mOjiYzM1N/f0lJif4v+6qEhoYSGhpa7r7169fr/z1gwAAGDBhQ08z1Kr9Qy4mzGjxcbPBr4WzqOA1Gn07etPZ2YvnGYyz54hjRPTS8NCIQG6vGeba2EKKaQhg5ciTnz5/n7NmzhISE6O9Xq9V07drV6OEaguNlI4E+3PH+/YlpbXm52bFkcn++ijzLN3vPcepyBq+PCcKvxf11boYQD4oqCyEwMJDAwED69u2Lp6dnfWVqMK6n5XEtNY9Ovm442D7Yh4oqY6424/mhHejq586KTcd5Y+3vPPOoHyMH+zWqgf6EENUUwtSpU1m9ejUvvvhihdN37NhhlFANQUmJjugzqTjaWeLfqomp4zR4AW3dWPvaI/x7WyybIs8SFZvM1Ge64tvc2dTRhBAGqrIQIiIiAJgzZ069hGlI4i5lcLOwhME9mqM2k0NFhrC3teT154Lo38WbD789yWtrfiN8oC+jhrTH8j4/iU+I+0GV+/QBAQEAPPzww3h5efHwww+TlZXFkSNH6NChQ70ENIXsvCLir2TSxscJD5eGcwJWY9EzwIt/vTGYwd2b883e80xZsZ/jZzWmjiWEqIZBB3nffvtt1q9fz8WLF3n33Xe5fv06s2bNMnY2k1AUheNnNZirzejSzs3UcRotexsLpjzTlXde6o1OUZi77iDz/3OYlAy5hrMQDZVBhRAXF8e8efPYvXs3I0aMYNGiRVy/fn/+9jwp7SYpGfkEtnVrtBe8aUi6tffgX9MHMvaJDpw8n8akpfv48r9n5HKdQjRABhWCoiiYmZlx4MABevXqBdwanfR+U6pTOH5Og6OdJe3ky9A6Y2Gu5u+D/fj3zMH07eTNlj3neGnhHrb/epEiuQiPEA2GQYXQokULIiIiSExM5OGHH+a1116jffv2xs5W785dvUFevpau7T0wky+S65yrkw2vjQli2ZT+tPJy5JMf4nhp4W5+/OMS2hIpBiFMzaBjIosWLWL37t0EBQVhYWFB9+7dGT58uJGj1a+CohJOXcrA280Obzc7U8e5r/m3bMJ7E/sQezGdjTvj+fi7WLbuO8+wfm0I6dVSzvkQwkQM2kOwtbWle/fu5OTkcOrUKTp16mTQ0BWNSeyFdEpKdXRt71H9zKJOBLZ1Y9Gkvsyf0Acfd3s+/+k049+N5INv/uRKSuO5nKoQ9wuD9hCWLVvGhg0bcHV11d+nUqnYu3ev0YLVp8ycQi5ez6Z9SxccH/DB6+qbSqWis587nf3cSUjOYcfvl/jl2DV2HbpCxzauDAxqTt/O3tjbWJg6qhD3PYMK4b///S+RkZH35fAViqJw4qwGKws1AW1cq19AGE0rL0deeboLY5/oQOThK+w9epUPvvmTf2+L4eGOnjzSrRld23tgLb/+EsIoDPo/y8vL674sA4Dk9JtobhQQ5O8hZ9M2EE72Vvx9sB8jB7XjQmIW+6MT+e3P60TFJGNpbkagrxvdO3jSvcP9+Z4UwlQMKoTevXuzdOlSBg8ejLW1tf7+jh07Gi1YfdApCn+eT8PexoK2zZxNHUf8hUqlol1zF9o1d+GF0I7EXkzn6JlUos+k8vF3sXz8XSxN7M3pdv4EHVo3oUOrJjTzsJdRaYWoJYMKYdu2bQDs3LlTf9/98B3CleQcsvOK6dPJS8YrauDUajO6+HnQxc+DiLBAktLzOHYmlRNnNRw+lcKeo1cBcLC9dQ5Ja29HWns70drbER93e9Tq+h15taRUh/lfttkQLp1YWqqr9+dCNB4GFcK+ffuMnaPelZbqiLmQThNHK1p4Opg6jqghbzd7nuxvz5P927Jx5xly87Wk3cgnLauAS0nZ/HlOg065Na+ZSoW9rQWOdpa3/mtriYOtJQ52llhbqo2yRzE6xJ9Nu+LL3VfTa2sbw+gQ/+pnEg8sgwrh5s2brFixgosXL7J69Wref/99ZsyYgZ1d4/29/vnELPILS+gpF75p9FQqFY52ljjaWeoP/ZXqFHJuFpGVW0RWXhF5+VpybhaTlHYTnaLol1WbqbC3scDe1gI7G0v9v+1tLLCzsbjrr3wh7mcGFcL8+fPx8PAgIyMDKysr8vLyePvtt1mxYoWx8xnFzQItpy5l0NTVlqaujbfUROXUZipcHKxxcbAud79OUcgvLCH3ZjG5+cXkFWi5WaAlL19LamY+JaVKufmtLdVlBWGpL4nbhWFjZS5/TIj7ikGFcObMGRYtWsSvv/6KjY0Ny5cvZ9iwYcbOZjTf/nKeYq2Ozu3cTR1F1DMzVdkegY0FXpT/Y0BRFIq0peTla/9XFGVlkXYjnyvJJdxZF2Zm/1vXnUXhaGdFqa58sQjRGBhUCGZm5XebS0tL77qvscjOK+L73y7RsqkDTRytq19APDBUKhXWluZYW5rj5mxz1/RSnUJ+oVZfGHkFxdws+3daVgHaEp1+3j1Hr2JvY4GzgxUuDlY4O1hLSYgGz6BC6NGjB8uWLaOwsJDff/+dDRs20LNnT2NnM4qSUh0dWzehmXyRLGpIbaa69WV0BWMtKYpCcYmOvPxisvOK8XS15VBsMtdScrmYmK2fz+VaAm7ONrg52+DubIOttRx2Eg2HQYXw+uuvs27dOhwcHFi1ahX9+vVj0qRJ1S63Y8cOPvroI7RaLePHj2fMmDHlpl+6dIm5c+eSnZ2Nu7s777//Pk5OTrV7JAZydbLh3Ql97voFiBD3QqVSYWWhxsrJBlcnG0aH+GNnbYFS9p3FjdwirlxPo6hUzeWkbM5fywLAxkqNu7Mtnq62NG1ii70M7CdMqNpC2L17N5988glnz57F2tqa9u3b061bN6ysrKpcLjU1lZUrV7Jt2zYsLS0ZNWoUPXv2xNfXF7j1F9XLL7/MrFmzCA4OZvny5axbt47p06fXzSMTogFQqVTYlX3HYFaSg7e3NzqdQlZeEelZBaRnFaC5UcDV1Fzg1pXmmrra4tnEDk9XW6zk7HlRj6oshO3bt/Phhx8yZcoU/P39UalUxMbGsmDBAoqKinjssccqXTYqKopevXrh7OwMQEhICDt37mTy5MkAnDp1CltbW4KDgwGYOHEiOTkywqW4/5mZqWjiaE0TR2v8WrigKAq5+cWkZOSTkpFPQnIuFxKzUQFuLjb4uNvTzN0eBxl4URhZlYXw5Zdf8tlnn5U7maZt27Z07tyZt956q8pC0Gg0uLv/71c8Hh4exMTE6G9fvXoVNzc3ZsyYwenTp/Hz82POnDn38liEaJRunUdhhaOdFX4tXNDpFDJzCklKv8l1TR5/nkvjz3NpONpZ4uNuj4+HPW5O1vLdg6hzVRaCVqut8MzK1q1bU1RUVOWKFeXuX1Tc+QYuKSnhyJEjbNiwgcDAQFatWsXixYtZvHixodmJi4szeN47BQUFkZSUVKtla8vw7fnXe7a/qnj7/kRHR9d7luqY4rU0TMWvY02yutmCWytLCovNycgtIT2nhPiETM4kZGJlrsLdyRx3JwscbMxqUA6Gv44N8fWujGQ1XFVDqFRZCGp15ccvK/rAv5OnpyfHjh3T39ZoNHh4/O/iM+7u7rRs2ZLAwEAAhg0bxpQpU6pc518FBARU+11GZepzCIGaDllgyuENqsraEMbiqYiph4OozF9z3cvQFW3K/lusLSUp/SZXU3JISr9JYoYWexsLmjd1oKWnA84OVtWWgyGvY3R0dIN9vf9KstYdo51M0KdPHw4ePEhmZiYFBQVERkbqvy8A6Nq1K5mZmcTH3/q1z759+xr96KlCGJulhZpWXo4Ed23GiEd86dmxKfa2FsQnZLLz0BV+OpBA7MV08vKLTR1VNEJV7iGcPXuWbt263XW/oigUF1f9hvP09GTatGmMHTsWrVbLyJEj6dSpExEREUyZMoXAwED+9a9/MXv2bAoKCmjatClLly69t0cjxAPE0kJNGx8n2vg4UVhcQmJqHldScoi7mEHcxQzcnW1o7e1Ic08HudaHMEiVhbB79+57WnloaCihoaHl7lu/fr3+3507d2br1q33tA0hBFhbmuPb3Bnf5s7cLNCSkJzD5aRsjpxOJTpeQzMPe1p7O1GqU2Sod1GpKgvBx8envnKIBk7G0W887Gws6NjGlYdaNyEju5DLSTlcTcnhSkouMRfSeKRbcwb1aE7Lpo6mjtrg1fX7vq6+PzDW/49ycVphELXarEGe3S3j+1dOpVLph8no5u9OUtpNirU6tv92kW37L+DbzIlB3VsQ3NUHJ/va/TjjflfX7/u6uiaGsd73UghCPADUZmY093RgdIg/WblF/HYikb3HrrFueyz/2RFH9w6eDOregu4dPLEwlz3BB5UUghAPGGcHK54MbsuTwW1JSM5h79Gr7D+eyKG4FBxsLRnQzQdv+2K6KYqc/PaAkUIQ4gHWysuRfzwZwPi/PcSJc2nsPXqVXYeuoC3RsfPPXxjcvTkDujXD1enu4cDF/UcKQQiBWm1G9w6edO/gSV5+MRt/OMgFjYpPfzzN5z+dpoufB4O6N6dXoJcMuHcfk0IQQpRjb2tJ93b2TBgVxPW0PPYdu8Yv0ddYvjEaW2tz+nfxYVD35nRo1UQOKd1npBCEEJXycbfn+aEdGBPiT+zFdPYdu8avxxPZdegKnk1s6d/Fh/5dfGjt7SjlcB+QQhBCVMvMTEXndu50bufOxKc6cTA2iV9PXGfb/gts3XceH3f7snLwpoWc39BoSSEIIWrExsqcQd1bMKh7C7LzioiKTeaPP6+zec9Zvt59lpZNHejf1Yc+gd40l0vVNipSCEKIWnOyt2Jo71YM7d2KzJxCDpxM4vc/r7Phv/Fs+G88Pu529ArwomdHL9q3dMFMhs1o0KQQhBB1oomjNaH92xDavw1pNwo4ciqZQ6dS2P7rRb795QLODlb07NiUnh2b0rmduwy41wBJIQgh6py7iw1/69eGv/VrQ16BlugzqRyKS+a3E9fZdegKVpZqAtu6EeTvQTd/D7zd7E0dWSCFIIQwMnsbCwZ0a8aAbs3QlpQSeyGDI6dTOH5Ww7EzqQA0dbWlW3sPurX3INDXDVtrCxOnfjBJIQgh6o2FuZpuZXsFAMnpNzl+VsPxeA37jl3j56gEzNUq2rdsQidfNwJ93fBv6YKFuRxeqg9SCEIIk/Fys+Nvbq35W9/WaEtKOZOQyfF4DSfPp/H17rN8FXkWS3MzOrRuQqCvG53autOuhTPmMhS7UUghCPEAMXQc/fq+7m9pqQ4LczWdfN3p5OsOQF6BllMX04m5mE7M+XQ2/DceiMfaUs1DrV1vFYSvG6W6qq/vLgwnhSDEA8TQ8f3ratx+Q1U0vr+9jQU9A7zoGeAFQHZeEXGXMoi9kE7MhXQ+/+k0AFYWKjrHHKZTu1sF0bKpo/y8tZakEIQQjYKTvRV9O3nTt9OtorqRW0jshXT2HYrnmiaXI6dTAHC0syTQ143Ovm50aueOt5udDKthICkEIUSj5OJgTXDXZtjpUgkKCkJzI1+/93DyfBoHTiYB4OpkTSdfNzq3c6ebvwcuDtYmTt5wSSEIIe4LHi62DO7RgsE9WqAoCsnpNzl5IZ2Y82lEx2v4JToRAN/mzvQoG+rbt5mzHF66gxSCEOK+o1Kp8Ha3x9vdnqG9W6HTKVxOyubYmVSOnknV/4LJ2cGK7v6e9OnkRRc/jwf+8qFGLYQdO3bw0UcfodVqGT9+PGPGjKlwvv379/Puu++yb98+Y8YRQjygzMxUtG3mTNtmzjwzpD3ZeUVEx986Me5gbBJ7jl7FztqcngFe9Ovs/cCWg9EKITU1lZUrV7Jt2zYsLS0ZNWoUPXv2xNfXt9x86enpLFmyxFgxhBDiLk72Vgzq3pxB3ZujLdHx5zkNf5xM4nBcMvuOXcPB1oIBXZvpDz89KF9KG60QoqKi6NWrF87OzgCEhISwc+dOJk+eXG6+2bNnM3nyZFasWGGsKEKIBs7Q8yMqcq/nTFiYm9Hjoab0eKipvhz2Ryey6/AVfjxwGSd7S9p4O9Ha2wkry/v7jGmjFYJGo8Hd3V1/28PDg5iYmHLzfPHFFzz00EN07tzZWDGEEI2AoedHVMQY50z4eNjzZP82eDSxZfPuc5w4l0bMhXRaeTni18IFZwerOt1eQ2G0QlCUu88evHO369y5c0RGRvLZZ5+RkpJSq23ExcXVarmgoCCSkpJqtWxtGb49/3rP9lcVb9/0uSrWuHKZPqvhz1f9Zr2319FYWccPG8Tps5fJKzQjKUPL5aRsLl7PxtlOjY+rBa4O5jU+nFQ3Wf2Jjo6u1ZJV7VEZrRA8PT05duyY/rZGo8HDw0N/e+fOnaSlpREeHo5Wq0Wj0TB69Gg2bdpk8DYCAgKwsqpdU9fnWZg1/QumPrP9VVVZTZmrKo0lV32f/VsZQzKYImttt2fsrLfX7dcGiopLuXg9i/PXsjh1tRAne0s6tnGluacDZgYUQ11mNcbwIkb7Gr1Pnz4cPHiQzMxMCgoKiIyMJDg4WD99ypQp7Nq1i++//55169bh4eFRozIQQoj6ZlU2jlJovzb0DvRCUSAqJpmfDyRwOSkbXSMfV8loheDp6cm0adMYO3Ysw4cPZ9iwYXTq1ImIiAhiY2ONtVkhhDA6MzMVrbwceaJPK/p28katVnEoLoWfoxJI1ORWeMi8MTDqeQihoaGEhoaWu2/9+vV3zdesWTM5B0EI0eioVCpaNHWguac919PyOHk+nd//TMLd2YYufu64OduYOmKNyJnKQghxj1QqFc08HPB2s+fS9WxiL6az+8hVmns60NXPHTubxnEFOCkEIYSoI2ZmKnybO9PSy5H4hEzOJGSSlJZHQFtX2rdsYup41ZJCEEKIOmZhbkagrxttfJyIjtdw8nw6l5NyaO2hxvS/M6vcgzdYhxBC1BM7GwuCu/oQ3NWH0lKFk5cLOBibTLG21NTRKiR7CEIIYWQ+7vZ4NrHlcMwVrqTkkJp5kx4PNcXH3d7U0cqRPQQhhKgH5mozWnta8VjPllhaqPntxPUGt7cghSCEEPWoiaM1Ib1a0bGNK1dScvg56jJJ6TdNHQuQQhBCiHqnNlPRydft1t6CuZpfjydyPF5DqU5n0lxSCEIIYSJNHK15rFdL2jV35uzVG0Qevkp2XpHJ8kghCCGECZmrzejewZPgrj4UFJaw69AVLlzLMsnwF1IIQgjRAPi42zO0TyvcXWw4eiaVg7HJaEvq9xCSFIIQQjQQNlbmPNKtGZ183biakkvk4Sv1eghJCkEIIRoQlUpFxzauDOzenCJtKbsOXeFyUna9bFsKQQghGiDPJrY83qsVTZysORSXwtHTKZSWGvcQkhSCEEI0ULbW5gwKak6HVk24kJjN7iNXySvQGm17UghCCNGAmZmp6OLnTnAXH/IKtOw6lMCFxCzjbMsoaxVCCFGnfDzsCenVEhcHa5LTjHNmsxSCEEI0Eg62lgzq3pz+XX2Msn4pBCGEEIAUghBCiDJSCEIIIQApBCGEEGWMWgg7duzgiSeeYMiQIWzcuPGu6Xv27CEsLIwnn3ySSZMmkZ1dP2fjCSGEuJvRCiE1NZWVK1eyadMmvv/+ezZv3syFCxf00/Py8pg3bx7r1q3jhx9+oH379qxdu9ZYcYQQQlTDaIUQFRVFr169cHZ2xtbWlpCQEHbu3KmfrtVqmTdvHp6engC0b9+e5ORkY8URQghRDXNjrVij0eDu7q6/7eHhQUxMjP62i4sLjz76KACFhYWsW7eO559/vkbbiIuLq1W2oKAgkpKSarVsbRm+Pf96z/ZXFW/f9Lkq1rhymT6r4c9X/Wa9t9fReFnr/v1VN+vzJzo6ulZLBgUFVTrNaIVQ0cUdVCrVXffl5uYyadIk/P39GTFiRI22ERAQgJWVVa3yeXt712q52khKSqrR9uoz219VldWUuarSWHLV9H1gLIZkMEXW2m7P2Fnrct11mbWqD/baMtohI09PT9LT0/W3NRoNHh4e5ebRaDSMHj0af39/FixYYKwoQgghDGC0QujTpw8HDx4kMzOTgoICIiMjCQ4O1k8vLS1l4sSJDB06lFmzZlW49yCEEKL+GO2QkaenJ9OmTWPs2LFotVpGjhxJp06diIiIYMqUKaSkpHD69GlKS0vZtWsXcOsQkOwpCCGEaRitEABCQ0MJDQ0td9/69esBCAwMJD4+3pibF0IIUQNyprIQQghACkEIIUQZKQQhhBCAFIIQQogyUghCCCEAKQQhhBBlpBCEEEIAUghCCCHKSCEIIYQApBCEEEKUkUIQQggBSCEIIYQoI4UghBACkEIQQghRRgpBCCEEIIUghBCijBSCEEIIQApBCCFEGSkEIYQQgBSCEEKIMlIIQgghACkEIYQQZYxaCDt27OCJJ55gyJAhbNy48a7pZ86cITw8nJCQEGbNmkVJSYkx4wghhKiC0QohNTWVlStXsmnTJr7//ns2b97MhQsXys0zffp05syZw65du1AUhS1bthgrjhBCiGqYG2vFUVFR9OrVC2dnZwBCQkLYuXMnkydPBuD69esUFhbSpUsXAJ566inWrFnD6NGjq123oigAFBcX1zqfuZmu1svWlLWFyuDtFRUV1Wu2v6osq6lzVaYx5arJ+8BYDH2+6jvrvbyOxsxa1++vuspaVFR0T8tbWlqiUqnuut9ohaDRaHB3d9ff9vDwICYmptLp7u7upKamGrRurVYLwLlz52qdr22TWi9ai205AgUGzRsXF1ev2f6qsqymzlWZxpSrJu8DYzH0+arvrPfyOhoza12/v+oqa1xc3D0tHxAQgJWV1V33G60Qbv8Vf6c7G6m66VWxs7PDz88PCwsLg5cRQghxi6WlZYX3G60QPD09OXbsmP62RqPBw8Oj3PT09HT97bS0tHLTq2JmZoaDg0PdhRVCCGG8L5X79OnDwYMHyczMpKCggMjISIKDg/XTfXx8sLKyIjo6GoDt27eXmy6EEKJ+qZSKjt3UkR07dvDxxx+j1WoZOXIkERERREREMGXKFAIDA4mPj2f27NncvHmThx56iEWLFlW6KyOEEMK4jFoIQgghGg85U1kIIQQghSCEEKKMFIIQQghACkEIIUQZKQQjWLJkCTNnzrzr/ujoaMLDwwkLC2PcuHFcv37dBOnKqyzrbadPnyYgIKAeE1WusqwajYaXXnqJ4cOHM2rUKBITE02QrrzKsiYmJjJmzBjCwsJ4/vnnTf4eGDt2LH/7298ICwsjLCyMkydPlpseFRVFaGgojz32GCtXrjRRyluqy7p582aGDRtGaGgob7755j0NbXOvqst628aNG3n++efrOV0VFFGnoqKilJ49eyozZsy4a9rAgQOVM2fOKIqiKN98840yceLE+o5XTlVZFUVR8vPzlWeeeUbx8/Or52R3qyrruHHjlE2bNimKoiibNm1Spk6dWs/pyqsq6+uvv65s3LhRURRF+eKLL5TXXnutvuPp6XQ6pW/fvopWq61wekFBgTJgwADl6tWrilarVV544QVl//799ZzyluqyXrp0SRkyZIiSm5ur6HQ65Y033lA+/fTT+g1Zprqst50/f17p37+/8txzz9VTsurJHkIdysrKYuXKlUycOPGuacXFxUydOhV/f38A2rdvT3Jycn1H1Ksq622LFy9m/Pjx9ReqElVlzczMJD4+nlGjRgEQHh7Oq6++Ws8J/6e651Wn05GXlwdAQUEB1tbW9RmvnEuXLqFSqYiIiODJJ59kw4YN5abHxMTQsmVLmjdvjrm5OaGhoezcubNBZrW0tGTevHnY29ujUqnw8/MjKSmpQWaFW58Hb7/9NlOnTjVBwsoZbeiKB9Hbb7/NtGnTKvygt7S0JCwsDLj1ofDBBx/w6KOP1ndEvaqyAuzdu5fCwkIef/zxek52t6qyXrt2DW9vbxYuXMjhw4fx9vZmzpw5Jkh5S3XP69SpUxk1ahRffvklWq2WzZs313PC/8nJyaF3797MmzePwsJCxo4dS+vWrenbty9Q8QCVhg5AWd9ZfXx88PHxAW79kbBx40YWLVrUILMCrFixgvDwcJo1a2aSjJWRPYQ68s033+Dl5UXv3r2rnK+4uJjXX3+dkpISJkyYUE/pyqsua1paGh999JFJP1hvqy5rSUkJp0+fpk+fPvzwww8MHjy4yu9EjMmQ98CMGTN49913+f3333nnnXeYPHlyhQM91oeuXbuydOlSbG1tadKkCSNHjuTXX3/VT68ol6kGk6wu622pqamMGzeO8PBwevbsaYKk1Wc9cOAAycnJhIeHmyRfVWQPoY78/PPPpKWlERYWRnZ2Nvn5+SxcuJC33npLP8/Nmzd5+eWXcXZ25qOPPsLCwqJBZt2/fz9ZWVmMGTNGv0xYWBgbN27E3t6+QWV1d3fHzs6OgQMHAjBs2DDmz59frxkNzZqZmcmlS5f0e4YhISHMnTuXGzdu0KRJ/Y/hfezYMbRarb7AFEXB3Px/Hwl/HYDyrwNU1qfqsgJcvHiRiIgInnvuOV544QVTxASqz/rjjz9y/vx5wsLCyM/PJz09nVdffZVVq1aZKPEdTPoNxn3q22+/rfALxZdfflmZPXu2otPpTJCqYpVlvVND+FJZUSrPOnToUP2XnT/99JPy7LPP1ne0u1SUVafTKf3791eOHj2qKIqiHDt2TBk0aJAp4imKoij79u1Thg8frhQWFiq5ublKaGiocvz4cf30wsJCJTg4WElISFBKSkqUf/zjH8rPP//cILPm5uYqAwYMULZv326SfHeqLuudDh061KC+VJY9BCObNWsWgwYNwsvLi7179+Lr68vw4cOBW8dk169fb9qAd7iddfDgwaaOUq07s37wwQfMnTuXZcuWYW9vz+LFi00dr5y/Zn3vvfcoLCzEzs6OtWvXmizXwIEDOXnyJMOHD0en0zF69Gi6du1KWFgY69atw9PTk8WLF/PKK69QVFTEgAEDTPadUnVZ//vf/5Kens5//vMf/vOf/wAwaNAgk3xpa8jz2lDJ4HZCCCEA+VJZCCFEGSkEIYQQgBSCEEKIMlIIQgghACkEIYQQZeRnp8KkEhMTGTJkCH5+fvr7FEVh7NixjBw50oTJ4IUXXmD58uX1ctJYTk4Oa9as4fDhw5iZmaFSqRgzZgx///vfgVs/oVy9ejWBgYFGzyIeXFIIwuSsra35/vvv9bdTU1MZNmwYAQEB+sEATeHAgQP1sp2ioiKee+45QkND+e677zA3N+f69ev6gQVvl4IQxiaFIBocT09PWrZsSUJCAv7+/nzzzTd89dVX6HQ6nJ2dmTNnDm3btmXmzJlkZWVx7do1HnnkESZNmsT8+fM5fvw4arWaRx99lGnTpqHValm+fDlHjx6ltLSUhx56iNmzZ2Nvb8+gQYMYMWIEBw8eJDk5maFDh/LGG2/w5ptvAjBu3DjWrVtHfHw8H3/8McXFxWRmZjJ8+HD9qKrr1q1j69at2NnZ0b17d/bu3cu+ffsoLi6udLt3+vnnn7G1tSUiIkJ/n4+PD6tWrUKr1erv27x5M3PnziUzM5OwsDCmTZuGTqdj4cKFnDx5kps3b6IoCvPnzycoKIiZM2dib2/P2bNnSUlJoU2bNrz//vvY2dnx66+/snz5cszMzOjQoQNRUVFs2rSJZs2aVfp8iweAaU+UFg+6a9euKV26dCl33/Hjx5UePXooSUlJyuHDh5XRo0cr+fn5iqIoyu+//64MHTpUURRFmTFjhjJu3Dj9cgsXLlSmTZumlJSUKEVFRcqYMWOUQ4cOKWvXrlUWL16sHzJkxYoVyty5cxVFuXWNisWLFyuKoigpKSlKYGCgcvXqVUVRbg3ZkZGRoeh0OuW5555TLl++rJ+vQ4cOSkZGhvLbb78pISEhSnZ2tqLT6ZQ333xTGThwoKIoSpXbvdO7776rLFmypMrnaeDAgcq7776rKIqiaDQaJSAgQElKSlKOHz+uvPLKK0ppaamiKIry8ccfKxMmTNA/P88884xSVFSkFBcXK8OHD1e2bt2qZGZmKg8//LD+2hzbtm1T/Pz8lGvXrlX5fIv7n+whCJMrLCzUDw1eWlqKi4sLy5Ytw8vLiy+//JIrV67or3cAkJ2dTVZWFgBBQUH6+6OionjzzTdRq9Wo1Wr9OPTLli0jNzeXqKgoALRaLa6urvrlbg/V4enpiaurK9nZ2TRv3lw/XaVS8e9//5v9+/fz448/cvHiRRRFoaCggF9//ZXHH38cR0dHAMaMGcOhQ4eAW4MEVrXdO9evGDBgwLBhw4BbA/q5ubmRkZFB165dcXJy4uuvv+batWscPnwYOzs7/TL9+/fH0tISAD8/P7Kzszl27Bht27bVH44bMWKEfkDA/fv3V/p8Ozs7V5tRNG5SCMLk/vodwp10Oh1hYWFMnz5df1uj0eDk5ASAra2tfl5zc/NywzMnJydjbW2NTqfjrbfeYsCAAcCtUWeLior081lZWen/XdGHc35+PiNGjODRRx+le/fuhIeHs2fPHv0olnfOr1ary2Wvaru3denShY0bN951/969ezl27BgzZszQP76/5ty/fz8LFizg//2//8fgwYNp06YNP/zwg36+Oy/Ac3sZtVp912M0MzPTZ67q+Rb3N/nZqWjQ+vbty08//YRGowHgq6++Yty4cRXO27t3b7777jt0Oh3FxcVMmTKFo0eP0q9fPzZu3EhxcTE6nY45c+bw/vvvV7tttVpNSUkJV65cIS8vj1dffZVBgwZx5MgR/boGDBhAZGQkubm5AGzdulW/vKHbfeyxx8jLy2P9+vWUlpYCty78s3jx4mqP3R84cICBAwcyevRoAgMD2bNnj34dlenWrRsJCQnEx8cDsGvXLnJyclCpVDV6vsX9R/YQRIPWv39/IiIieOGFF1CpVNjb2/PBBx9UeKGWyZMns2DBAsLCwigtLeWJJ57gscceIzg4mCVLljBixAhKS0vp0KGDQRfRGTJkCKNHj+aDDz7gkUceYejQoTg6OtKiRQt8fX25cuUK/fv35+mnn+aZZ57B2tqadu3aYWNjA8CkSZMM2q6lpSWffvopy5YtIzQ0VH/I6+WXX+app56qMuOoUaN4/fXX9ct1796dyMhIdDpdpcs4Ozvz/vvvM2PGDMzMzAgICMDc3BwbG5saPd/i/iOjnQpxD2JjYzlx4gRjx44F4NNPP+XkyZMN42InlcjLy+PDDz/klVdewcbGhlOnTjFhwgR+//13+eB/wMkeghD3oHXr1qxfv54tW7agUqnw8vLivffeM3WsKtnb22NhYcHIkSMxNzfH3NycVatWSRkI2UMQQghxi3ypLIQQApBCEEIIUUYKQQghBCCFIIQQoowUghBCCEAKQQghRJn/DwjsZ4N9cGs0AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_style('whitegrid')\n",
    "sns.histplot(y_pred, kde=True, stat='density')\n",
    "plt.title('Distribution of Predicted Values')\n",
    "plt.xlabel('Percentage Change')\n",
    "sns.despine()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 可视化预测结果和真实值的对比"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(y_test, label='True')\n",
    "plt.plot(y_pred, label='Predicted')\n",
    "plt.title('Test set: True vs predicted')\n",
    "plt.xlabel('Index')\n",
    "plt.ylabel('Percentage change')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 可视化预测结果与输入特征之间的关系"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots(1, 2, figsize=(15, 5))\n",
    "axs[0].scatter(X_test[:, 0], y_test, alpha=0.5)\n",
    "axs[0].scatter(X_test[:, 0], y_pred, alpha=0.5)\n",
    "axs[0].set_xlabel('EPS')\n",
    "axs[0].set_ylabel('Percentage change')\n",
    "axs[0].legend(['True', 'Predicted'])\n",
    "axs[1].scatter(X_test[:, 1], y_test, alpha=0.5)\n",
    "axs[1].scatter(X_test[:, 1], y_pred, alpha=0.5)\n",
    "axs[1].set_xlabel('Circulating market value')\n",
    "axs[1].set_ylabel('Percentage change')\n",
    "axs[1].legend(['True', 'Predicted'])\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 自我评价\n",
    "整个模型跑下来的效果一般般，参数例如学习率、批次大小、神经元数等是手动设定的，可以使用网格搜索或贝叶斯优化来自动化寻找最优的超参数组合，以进一步提高模型性能。"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}