liaozhaorun/NewStock
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
a567d86434c2a833649887d4b0faa93bc44af385
NewStock/code/train/UpdateRegression.ipynb

1134 lines
212 KiB
Plaintext
Raw Normal View History

(data leak)回归模型+滚动训练+l2_code+index+industry，sharpe-1.67
2025-02-23 23:47:19 +08:00
{
"cells": [
{
"cell_type": "code",
"id": "79a7758178bafdd3",
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:46:25.278301Z",
"start_time": "2025-02-23T15:46:25.186393Z"
}
},
"source": [
"%load_ext autoreload\n",
"%autoreload 2\n",
"\n",
"import pandas as pd\n",
" \n",
"\n",
" \n",
"pd.set_option('display.max_columns', None)\n"
],
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"The autoreload extension is already loaded. To reload it, use:\n",
" %reload_ext autoreload\n"
]
}
],
"execution_count": 22
},
{
"cell_type": "code",
"id": "a79cafb06a7e0e43",
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:47:12.994735Z",
"start_time": "2025-02-23T15:46:25.278301Z"
}
},
"source": [
"from utils.utils import read_and_merge_h5_data\n",
"\n",
"print('daily data')\n",
"df = read_and_merge_h5_data('../../data/daily_data.h5', key='daily_data',\n",
" columns=['ts_code', 'trade_date', 'open', 'close', 'high', 'low', 'vol'],\n",
" df=None)\n",
"\n",
"print('daily basic')\n",
"df = read_and_merge_h5_data('../../data/daily_basic.h5', key='daily_basic',\n",
" columns=['ts_code', 'trade_date', 'turnover_rate', 'pe_ttm', 'circ_mv', 'volume_ratio',\n",
" 'is_st'], df=df, join='inner')\n",
"\n",
"print('stk limit')\n",
"df = read_and_merge_h5_data('../../data/stk_limit.h5', key='stk_limit',\n",
" columns=['ts_code', 'trade_date', 'pre_close', 'up_limit', 'down_limit'],\n",
" df=df)\n",
"print('money flow')\n",
"df = read_and_merge_h5_data('../../data/money_flow.h5', key='money_flow',\n",
" columns=['ts_code', 'trade_date', 'buy_sm_vol', 'sell_sm_vol', 'buy_lg_vol', 'sell_lg_vol',\n",
" 'buy_elg_vol', 'sell_elg_vol', 'net_mf_vol'],\n",
" df=df)\n",
"print(df.info())"
],
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"daily data\n",
"daily basic\n",
"inner merge on ['ts_code', 'trade_date']\n",
"stk limit\n",
"left merge on ['ts_code', 'trade_date']\n",
"money flow\n",
"left merge on ['ts_code', 'trade_date']\n",
"<class 'pandas.core.frame.DataFrame'>\n",
"RangeIndex: 8296325 entries, 0 to 8296324\n",
"Data columns (total 21 columns):\n",
" # Column Dtype \n",
"--- ------ ----- \n",
" 0 ts_code object \n",
" 1 trade_date datetime64[ns]\n",
" 2 open float64 \n",
" 3 close float64 \n",
" 4 high float64 \n",
" 5 low float64 \n",
" 6 vol float64 \n",
" 7 turnover_rate float64 \n",
" 8 pe_ttm float64 \n",
" 9 circ_mv float64 \n",
" 10 volume_ratio float64 \n",
" 11 is_st bool \n",
" 12 up_limit float64 \n",
" 13 down_limit float64 \n",
" 14 buy_sm_vol float64 \n",
" 15 sell_sm_vol float64 \n",
" 16 buy_lg_vol float64 \n",
" 17 sell_lg_vol float64 \n",
" 18 buy_elg_vol float64 \n",
" 19 sell_elg_vol float64 \n",
" 20 net_mf_vol float64 \n",
"dtypes: bool(1), datetime64[ns](1), float64(18), object(1)\n",
"memory usage: 1.2+ GB\n",
"None\n"
]
}
],
"execution_count": 23
},
{
"cell_type": "code",
"id": "f7a55c19-b7dc-4d2f-a478-cffab11690df",
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:47:14.883651Z",
"start_time": "2025-02-23T15:47:13.089165Z"
}
},
"source": [
"print('industry')\n",
"df = read_and_merge_h5_data('../../data/industry_data.h5', key='industry_data',\n",
" columns=['ts_code', 'l2_code'],\n",
" df=df, on=['ts_code'], join='left')\n"
],
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"industry\n",
"left merge on ['ts_code']\n"
]
}
],
"execution_count": 24
},
{
"cell_type": "code",
"id": "4077d4449d406c86",
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:47:15.140211Z",
"start_time": "2025-02-23T15:47:14.978172Z"
}
},
"source": [
"import pandas as pd\n",
"import numpy as np\n",
"\n",
"def calculate_indicators(df):\n",
" \"\"\"\n",
" 计算四个指标：当日涨跌幅、5日移动平均、RSI、MACD。\n",
" \"\"\"\n",
" df['daily_return'] = (df['close'] - df['pre_close']) / df['pre_close'] * 100\n",
" # df['5_day_ma'] = df['close'].rolling(window=5).mean()\n",
" delta = df['close'].diff()\n",
" gain = delta.where(delta > 0, 0)\n",
" loss = -delta.where(delta < 0, 0)\n",
" avg_gain = gain.rolling(window=14).mean()\n",
" avg_loss = loss.rolling(window=14).mean()\n",
" rs = avg_gain / avg_loss\n",
" df['RSI'] = 100 - (100 / (1 + rs))\n",
"\n",
" # 计算MACD\n",
" ema12 = df['close'].ewm(span=12, adjust=False).mean()\n",
" ema26 = df['close'].ewm(span=26, adjust=False).mean()\n",
" df['MACD'] = ema12 - ema26\n",
" df['Signal_line'] = df['MACD'].ewm(span=9, adjust=False).mean()\n",
" df['MACD_hist'] = df['MACD'] - df['Signal_line']\n",
"\n",
" return df\n",
"\n",
"def generate_index_indicators(h5_filename):\n",
" df = pd.read_hdf(h5_filename, key='index_data')\n",
" df['trade_date'] = pd.to_datetime(df['trade_date'], format='%Y%m%d')\n",
" df = df.sort_values('trade_date')\n",
"\n",
" # 计算每个ts_code的相关指标\n",
" df_indicators = []\n",
" for ts_code in df['ts_code'].unique():\n",
" df_index = df[df['ts_code'] == ts_code].copy()\n",
" df_index = calculate_indicators(df_index)\n",
" df_indicators.append(df_index)\n",
"\n",
" # 合并所有指数的结果\n",
" df_all_indicators = pd.concat(df_indicators, ignore_index=True)\n",
"\n",
" # 保留trade_date列，并将同一天的数据按ts_code合并成一行\n",
" df_final = df_all_indicators.pivot_table(\n",
" index='trade_date',\n",
" columns='ts_code',\n",
" values=['daily_return', 'RSI', 'MACD', 'Signal_line', 'MACD_hist'],\n",
" aggfunc='last'\n",
" )\n",
"\n",
" df_final.columns = [f\"{col[1]}_{col[0]}\" for col in df_final.columns]\n",
" df_final = df_final.reset_index()\n",
"\n",
" return df_final\n",
"\n",
"\n",
"# 使用函数\n",
"h5_filename = '../../data/index_data.h5'\n",
"index_data = generate_index_indicators(h5_filename)\n",
"index_data = index_data.dropna()\n"
],
"outputs": [],
"execution_count": 25
},
{
"cell_type": "code",
"id": "c4e9e1d31da6dba6",
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:47:15.338283Z",
"start_time": "2025-02-23T15:47:15.234691Z"
}
},
"source": [
"import numpy as np\n",
"import talib\n",
"\n",
"def get_technical_factor(df):\n",
" # 按股票和日期排序\n",
" df = df.sort_values(by=['ts_code', 'trade_date'])\n",
" grouped = df.groupby('ts_code', group_keys=False)\n",
"\n",
" # 计算 up 和 down\n",
" df['log_close'] = np.log(df['close'])\n",
"\n",
" df['up'] = (df['high'] - df[['close', 'open']].max(axis=1)) / df['close']\n",
" df['down'] = (df[['close', 'open']].min(axis=1) - df['low']) / df['close']\n",
"\n",
" # 计算 ATR\n",
" df['atr_14'] = grouped.apply(\n",
" lambda x: pd.Series(talib.ATR(x['high'].values, x['low'].values, x['close'].values, timeperiod=14), index=x.index)\n",
" )\n",
" df['atr_6'] = grouped.apply(\n",
" lambda x: pd.Series(talib.ATR(x['high'].values, x['low'].values, x['close'].values, timeperiod=6), index=x.index)\n",
" )\n",
"\n",
" # 计算 OBV 及其均线\n",
" df['obv'] = grouped.apply(\n",
" lambda x: pd.Series(talib.OBV(x['close'].values, x['vol'].values), index=x.index)\n",
" )\n",
" df['maobv_6'] = grouped.apply(\n",
" lambda x: pd.Series(talib.SMA(x['obv'].values, timeperiod=6), index=x.index)\n",
" )\n",
" df['obv-maobv_6'] = df['obv'] - df['maobv_6']\n",
"\n",
" # 计算 RSI\n",
" df['rsi_3'] = grouped.apply(\n",
" lambda x: pd.Series(talib.RSI(x['close'].values, timeperiod=3), index=x.index)\n",
" )\n",
" df['rsi_6'] = grouped.apply(\n",
" lambda x: pd.Series(talib.RSI(x['close'].values, timeperiod=6), index=x.index)\n",
" )\n",
" df['rsi_9'] = grouped.apply(\n",
" lambda x: pd.Series(talib.RSI(x['close'].values, timeperiod=9), index=x.index)\n",
" )\n",
"\n",
" # 计算 return_10 和 return_20\n",
" df['return_5'] = grouped['close'].apply(lambda x: x / x.shift(5) - 1)\n",
" df['return_10'] = grouped['close'].apply(lambda x: x / x.shift(10) - 1)\n",
" df['return_20'] = grouped['close'].apply(lambda x: x / x.shift(20) - 1)\n",
"\n",
" # 计算 avg_close_5\n",
" df['avg_close_5'] = grouped['close'].apply(lambda x: x.rolling(window=5).mean() / x)\n",
"\n",
" # 计算标准差指标\n",
" df['std_return_5'] = grouped['close'].apply(lambda x: x.pct_change().rolling(window=5).std())\n",
" df['std_return_15'] = grouped['close'].apply(lambda x: x.pct_change().rolling(window=15).std())\n",
" df['std_return_25'] = grouped['close'].apply(lambda x: x.pct_change().rolling(window=25).std())\n",
" df['std_return_90'] = grouped['close'].apply(lambda x: x.pct_change().rolling(window=90).std())\n",
" df['std_return_90_2'] = grouped['close'].apply(lambda x: x.shift(10).pct_change().rolling(window=90).std())\n",
"\n",
" # 计算比值指标\n",
" df['std_return_5 / std_return_90'] = df['std_return_5'] / df['std_return_90']\n",
" df['std_return_5 / std_return_25'] = df['std_return_5'] / df['std_return_25']\n",
"\n",
" # 计算标准差差值\n",
" df['std_return_90 - std_return_90_2'] = df['std_return_90'] - df['std_return_90_2']\n",
"\n",
" return df\n",
"\n",
"\n",
"def get_act_factor(df, cat=True):\n",
" # 按股票和日期排序\n",
" df = df.sort_values(by=['ts_code', 'trade_date'])\n",
" grouped = df.groupby('ts_code', group_keys=False)\n",
" # 计算 EMA 指标\n",
" df['ema_5'] = grouped['close'].apply(\n",
" lambda x: pd.Series(talib.EMA(x.values, timeperiod=5), index=x.index)\n",
" )\n",
" df['ema_13'] = grouped['close'].apply(\n",
" lambda x: pd.Series(talib.EMA(x.values, timeperiod=13), index=x.index)\n",
" )\n",
" df['ema_20'] = grouped['close'].apply(\n",
" lambda x: pd.Series(talib.EMA(x.values, timeperiod=20), index=x.index)\n",
" )\n",
" df['ema_60'] = grouped['close'].apply(\n",
" lambda x: pd.Series(talib.EMA(x.values, timeperiod=60), index=x.index)\n",
" )\n",
"\n",
" # 计算 act_factor1, act_factor2, act_factor3, act_factor4\n",
" df['act_factor1'] = grouped['ema_5'].apply(\n",
" lambda x: np.arctan((x / x.shift(1) - 1) * 100) * 57.3 / 50\n",
" )\n",
" df['act_factor2'] = grouped['ema_13'].apply(\n",
" lambda x: np.arctan((x / x.shift(1) - 1) * 100) * 57.3 / 40\n",
" )\n",
" df['act_factor3'] = grouped['ema_20'].apply(\n",
" lambda x: np.arctan((x / x.shift(1) - 1) * 100) * 57.3 / 21\n",
" )\n",
" df['act_factor4'] = grouped['ema_60'].apply(\n",
" lambda x: np.arctan((x / x.shift(1) - 1) * 100) * 57.3 / 10\n",
" )\n",
"\n",
" if cat:\n",
" df['cat_af1'] = df['act_factor1'] > 0\n",
" df['cat_af2'] = df['act_factor2'] > df['act_factor1']\n",
" df['cat_af3'] = df['act_factor3'] > df['act_factor2']\n",
" df['cat_af4'] = df['act_factor4'] > df['act_factor3']\n",
"\n",
" # 计算 act_factor5 和 act_factor6\n",
" df['act_factor5'] = df['act_factor1'] + df['act_factor2'] + df['act_factor3'] + df['act_factor4']\n",
" df['act_factor6'] = (df['act_factor1'] - df['act_factor2']) / np.sqrt(df['act_factor1']**2 + df['act_factor2']**2)\n",
"\n",
" # 根据 trade_date 截面计算排名\n",
" df['rank_act_factor1'] = df.groupby('trade_date', group_keys=False)['act_factor1'].rank(ascending=False, pct=True)\n",
" df['rank_act_factor2'] = df.groupby('trade_date', group_keys=False)['act_factor2'].rank(ascending=False, pct=True)\n",
" df['rank_act_factor3'] = df.groupby('trade_date', group_keys=False)['act_factor3'].rank(ascending=False, pct=True)\n",
"\n",
" return df\n",
"\n",
"\n",
"def get_money_flow_factor(df):\n",
" # 计算资金流相关因子（字段名称见 tushare 数据说明）\n",
" df['active_buy_volume_large'] = df['buy_lg_vol'] / df['net_mf_vol']\n",
" df['active_buy_volume_big'] = df['buy_elg_vol'] / df['net_mf_vol']\n",
" df['active_buy_volume_small'] = df['buy_sm_vol'] / df['net_mf_vol']\n",
"\n",
" df['buy_lg_vol_minus_sell_lg_vol'] = (df['buy_lg_vol'] - df['sell_lg_vol']) / df['net_mf_vol']\n",
" df['buy_elg_vol_minus_sell_elg_vol'] = (df['buy_elg_vol'] - df['sell_elg_vol']) / df['net_mf_vol']\n",
"\n",
" df['log(circ_mv)'] = np.log(df['circ_mv'])\n",
" return df\n",
"\n",
"\n",
"def get_alpha_factor(df):\n",
" df = df.sort_values(by=['ts_code', 'trade_date'])\n",
" grouped = df.groupby('ts_code')\n",
"\n",
" # alpha_022: 当前 close 与 5 日前 close 差值\n",
" df['alpha_022'] = grouped['close'].transform(lambda x: x - x.shift(5))\n",
"\n",
" # alpha_003: (close - open) / (high - low)\n",
" df['alpha_003'] = np.where(df['high'] != df['low'],\n",
" (df['close'] - df['open']) / (df['high'] - df['low']),\n",
" 0)\n",
"\n",
" # alpha_007: 计算过去5日 close 与 vol 的相关性，并按 trade_date 排名\n",
" df['alpha_007'] = grouped.apply(lambda x: x['close'].rolling(5).corr(x['vol'])).reset_index(level=0, drop=True)\n",
" df['alpha_007'] = df.groupby('trade_date', group_keys=False)['alpha_007'].rank(ascending=True, pct=True)\n",
"\n",
" # alpha_013: 计算过去5日 close 之和 - 20日 close 之和，并按 trade_date 排名\n",
" df['alpha_013'] = grouped['close'].transform(lambda x: x.rolling(5).sum() - x.rolling(20).sum())\n",
" df['alpha_013'] = df.groupby('trade_date', group_keys=False)['alpha_013'].rank(ascending=True, pct=True)\n",
"\n",
" return df\n",
"\n"
],
"outputs": [],
"execution_count": 26
},
{
"cell_type": "code",
"id": "a735bc02ceb4d872",
"metadata": {
"jupyter": {
"source_hidden": true,
"is_executing": true
},
"ExecuteTime": {
"start_time": "2025-02-23T15:47:15.450197Z"
}
},
"source": [
"def read_industry_data(h5_filename):\n",
" # 读取 H5 文件中所有的行业数据\n",
" industry_data = pd.read_hdf(h5_filename, key='sw_daily', columns=[\n",
" 'ts_code', 'trade_date', 'open', 'close', 'high', 'low', 'pe', 'pb', 'vol'\n",
" ]) # 假设 H5 文件的键是 'industry_data'\n",
" industry_data = industry_data.reindex()\n",
" industry_data['trade_date'] = pd.to_datetime(df['trade_date'], format='%Y%m%d')\n",
"\n",
" grouped = industry_data.groupby('ts_code', group_keys=False)\n",
" industry_data['obv'] = grouped.apply(\n",
" lambda x: pd.Series(talib.OBV(x['close'].values, x['vol'].values), index=x.index)\n",
" )\n",
" industry_data['return_5'] = grouped['close'].apply(lambda x: x / x.shift(5) - 1)\n",
" # industry_data['return_20'] = grouped['close'].apply(lambda x: x / x.shift(20) - 1)\n",
"\n",
" # industry_data = get_act_factor(industry_data, cat=False)\n",
" # industry_data = industry_data.sort_values(by=['trade_date', 'ts_code'])\n",
"\n",
" # 计算每天每个 ts_code 的因子和当天所有 ts_code 的中位数的偏差\n",
" factor_columns = ['obv', 'return_5', 'return_20', 'act_factor1', 'act_factor2', 'act_factor3', 'act_factor4'] # 因子列\n",
"\n",
" for factor in factor_columns:\n",
" if factor in industry_data.columns:\n",
" # 计算每天每个 ts_code 的因子值与当天所有 ts_code 的中位数的偏差\n",
" industry_data[f'{factor}_deviation'] = industry_data.groupby('trade_date')[factor].transform(lambda x: x - x.median())\n",
"\n",
" industry_data['return_5_percentile'] = industry_data.groupby('trade_date')['return_5'].transform(lambda x: x.rank(pct=True))\n",
" industry_data = industry_data.drop(columns=['open', 'close', 'high', 'low', 'pe', 'pb', 'vol'])\n",
"\n",
"\n",
" industry_data = industry_data.rename(\n",
" columns={col: f'industry_{col}' for col in industry_data.columns if col not in ['ts_code', 'trade_date']})\n",
"\n",
" industry_data = industry_data.rename(columns={'ts_code': 'cat_l2_code'})\n",
" return industry_data\n",
"\n",
"industry_df = read_industry_data('../../data/sw_daily.h5')\n"
],
"outputs": [],
"execution_count": null
},
{
"cell_type": "code",
"id": "53f86ddc0677a6d7",
"metadata": {
"scrolled": true,
"ExecuteTime": {
"end_time": "2025-02-23T15:38:37.800520Z",
"start_time": "2025-02-23T15:38:37.753477Z"
}
},
"source": [
"origin_columns = df.columns.tolist()\n",
"origin_columns = [col for col in origin_columns if col not in ['turnover_rate', 'pe_ttm', 'volume_ratio', 'l2_code']]\n",
"origin_columns = [col for col in origin_columns if col not in index_data.columns]\n"
],
"outputs": [],
"execution_count": 9
},
{
"cell_type": "code",
"id": "dbe2fd8021b9417f",
"metadata": {
"scrolled": true,
"ExecuteTime": {
"end_time": "2025-02-23T15:39:35.360052Z",
"start_time": "2025-02-23T15:38:37.818014Z"
}
},
"source": [
"def filter_data(df):\n",
" # df = df.groupby('trade_date').apply(lambda x: x.nlargest(1000, 'act_factor1'))\n",
" df = df[~df['is_st']]\n",
" df = df[~df['ts_code'].str.endswith('BJ')]\n",
" df = df[~df['ts_code'].str.startswith('30')]\n",
" df = df[~df['ts_code'].str.startswith('68')]\n",
" df = df[~df['ts_code'].str.startswith('8')]\n",
" df = df.reset_index(drop=True)\n",
" return df\n",
"\n",
"\n",
"df = filter_data(df)\n",
"df = get_technical_factor(df)\n",
"df = get_act_factor(df)\n",
"df = get_money_flow_factor(df)\n",
"df = get_alpha_factor(df)\n",
"# df = df.merge(industry_df, on=['l2_code', 'trade_date'], how='left')\n",
"df = df.rename(columns={'l2_code': 'cat_l2_code'})\n",
"# df = df.merge(index_data, on='trade_date', how='left')\n",
"\n",
"print(df.info())"
],
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"<class 'pandas.core.frame.DataFrame'>\n",
"Index: 5538535 entries, 1962 to 5538534\n",
"Data columns (total 72 columns):\n",
" # Column Dtype \n",
"--- ------ ----- \n",
" 0 ts_code object \n",
" 1 trade_date datetime64[ns]\n",
" 2 open float64 \n",
" 3 close float64 \n",
" 4 high float64 \n",
" 5 low float64 \n",
" 6 vol float64 \n",
" 7 turnover_rate float64 \n",
" 8 pe_ttm float64 \n",
" 9 circ_mv float64 \n",
" 10 volume_ratio float64 \n",
" 11 is_st bool \n",
" 12 up_limit float64 \n",
" 13 down_limit float64 \n",
" 14 buy_sm_vol float64 \n",
" 15 sell_sm_vol float64 \n",
" 16 buy_lg_vol float64 \n",
" 17 sell_lg_vol float64 \n",
" 18 buy_elg_vol float64 \n",
" 19 sell_elg_vol float64 \n",
" 20 net_mf_vol float64 \n",
" 21 cat_l2_code object \n",
" 22 log_close float64 \n",
" 23 up float64 \n",
" 24 down float64 \n",
" 25 atr_14 float64 \n",
" 26 atr_6 float64 \n",
" 27 obv float64 \n",
" 28 maobv_6 float64 \n",
" 29 obv-maobv_6 float64 \n",
" 30 rsi_3 float64 \n",
" 31 rsi_6 float64 \n",
" 32 rsi_9 float64 \n",
" 33 return_5 float64 \n",
" 34 return_10 float64 \n",
" 35 return_20 float64 \n",
" 36 avg_close_5 float64 \n",
" 37 std_return_5 float64 \n",
" 38 std_return_15 float64 \n",
" 39 std_return_25 float64 \n",
" 40 std_return_90 float64 \n",
" 41 std_return_90_2 float64 \n",
" 42 std_return_5 / std_return_90 float64 \n",
" 43 std_return_5 / std_return_25 float64 \n",
" 44 std_return_90 - std_return_90_2 float64 \n",
" 45 ema_5 float64 \n",
" 46 ema_13 float64 \n",
" 47 ema_20 float64 \n",
" 48 ema_60 float64 \n",
" 49 act_factor1 float64 \n",
" 50 act_factor2 float64 \n",
" 51 act_factor3 float64 \n",
" 52 act_factor4 float64 \n",
" 53 cat_af1 bool \n",
" 54 cat_af2 bool \n",
" 55 cat_af3 bool \n",
" 56 cat_af4 bool \n",
" 57 act_factor5 float64 \n",
" 58 act_factor6 float64 \n",
" 59 rank_act_factor1 float64 \n",
" 60 rank_act_factor2 float64 \n",
" 61 rank_act_factor3 float64 \n",
" 62 active_buy_volume_large float64 \n",
" 63 active_buy_volume_big float64 \n",
" 64 active_buy_volume_small float64 \n",
" 65 buy_lg_vol_minus_sell_lg_vol float64 \n",
" 66 buy_elg_vol_minus_sell_elg_vol float64 \n",
" 67 log(circ_mv) float64 \n",
" 68 alpha_022 float64 \n",
" 69 alpha_003 float64 \n",
" 70 alpha_007 float64 \n",
" 71 alpha_013 float64 \n",
"dtypes: bool(5), datetime64[ns](1), float64(64), object(2)\n",
"memory usage: 2.8+ GB\n",
"None\n"
]
}
],
"execution_count": 10
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:39:35.623327Z",
"start_time": "2025-02-23T15:39:35.485449Z"
}
},
"cell_type": "code",
"source": [
"def create_deviation_within_dates(df, feature_columns):\n",
" groupby_col = 'cat_l2_code' # 使用 trade_date 进行分组\n",
" new_columns = {}\n",
" ret_feature_columns = feature_columns[:]\n",
"\n",
" # 自动选择所有数值型特征\n",
" num_features = [col for col in feature_columns if 'cat' not in col and 'index' not in col]\n",
" num_features = [col for col in feature_columns if 'cat' not in col and 'industry' not in col]\n",
"\n",
" # 遍历所有数值型特征\n",
" for feature in num_features:\n",
" if feature == 'trade_date': # 不需要对 'trade_date' 计算偏差\n",
" continue\n",
"\n",
" grouped_median = df.groupby(['trade_date', groupby_col])[feature].transform('median')\n",
" deviation_col_name = f'deviation_median_{feature}'\n",
" new_columns[deviation_col_name] = df[feature] - grouped_median\n",
" ret_feature_columns.append(deviation_col_name)\n",
"\n",
" grouped_mean = df.groupby(groupby_col)[feature].transform('mean')\n",
" deviation_col_name = f'deviation_mean_{feature}'\n",
" new_columns[deviation_col_name] = df[feature] - grouped_mean\n",
" ret_feature_columns.append(deviation_col_name)\n",
"\n",
" # 将新计算的偏差特征与原始 DataFrame 合并\n",
" df = pd.concat([df, pd.DataFrame(new_columns)], axis=1)\n",
"\n",
" # for feature in ['obv', 'return_20', 'act_factor1', 'act_factor2', 'act_factor3', 'act_factor4']:\n",
" # df[f'deviation_industry_{feature}'] = df[feature] - df[f'industry_{feature}']\n",
"\n",
" return df, ret_feature_columns\n"
],
"id": "d345bcc43b15579e",
"outputs": [],
"execution_count": 11
},
{
"cell_type": "code",
"id": "5f3d9aece75318cd",
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:40:13.266857Z",
"start_time": "2025-02-23T15:39:35.672924Z"
}
},
"source": [
"def get_qcuts(series, quantiles):\n",
" q = pd.qcut(series, q=quantiles, labels=False, duplicates='drop')\n",
" return q[-1] # 返回窗口最后一个元素的分位数标签\n",
"\n",
"\n",
"import pandas as pd\n",
"\n",
"\n",
"def remove_outliers_label_percentile(label: pd.Series, lower_percentile: float = 0.01, upper_percentile: float = 0.99):\n",
" if not (0 <= lower_percentile < upper_percentile <= 1):\n",
" raise ValueError(\"Percentile values must satisfy 0 <= lower_percentile < upper_percentile <= 1.\")\n",
"\n",
" # Calculate lower and upper bounds based on percentiles\n",
" lower_bound = label.quantile(lower_percentile)\n",
" upper_bound = label.quantile(upper_percentile)\n",
"\n",
" # Filter out values outside the bounds\n",
" filtered_label = label[(label >= lower_bound) & (label <= upper_bound)]\n",
"\n",
" # Print the number of removed outliers\n",
" print(f\"Removed {len(label) - len(filtered_label)} outliers.\")\n",
" return filtered_label\n",
"\n",
"\n",
"def calculate_risk_adjusted_target(df, days=5):\n",
" df = df.sort_values(by=['ts_code', 'trade_date'])\n",
"\n",
" df['future_close'] = df.groupby('ts_code')['close'].shift(-days)\n",
" df['future_return'] = (df['future_close'] - df['close']) / df['close']\n",
"\n",
" df['future_volatility'] = df.groupby('ts_code')['future_return'].rolling(days, min_periods=1).std().reset_index(\n",
" level=0, drop=True)\n",
" df['sharpe_ratio'] = df['future_return'] * df['future_volatility']\n",
" df['sharpe_ratio'].replace([np.inf, -np.inf], np.nan, inplace=True)\n",
"\n",
" return df['sharpe_ratio']\n",
"\n",
"\n",
"future_close = df.groupby('ts_code')['close'].shift(-4)\n",
"future_return = (future_close - df['close']) / df['close']\n",
"df['label'] = future_return\n",
"df['label'] = remove_outliers_label_percentile(df['label'])\n",
"\n",
"# df = df.apply(lambda x: x.astype('float32') if x.dtype in ['float64', 'float32'] else x)\n",
"df = df.sort_values(by=['trade_date', 'ts_code'])\n",
"train_data = df[(df['trade_date'] <= '2023-01-01') & (df['trade_date'] >= '2016-01-01')]\n",
"test_data = df[df['trade_date'] >= '2023-01-01']\n",
"\n",
"train_data = train_data.merge(industry_df, on=['cat_l2_code', 'trade_date'], how='left')\n",
"# train_data = train_data.merge(index_data, on='trade_date', how='left')\n",
"test_data = test_data.merge(industry_df, on=['cat_l2_code', 'trade_date'], how='left')\n",
"# test_data = test_data.merge(index_data, on='trade_date', how='left')\n",
"\n",
"train_data = train_data.groupby('trade_date', group_keys=False).apply(lambda x: x.nlargest(1000, 'return_20'))\n",
"test_data = test_data.groupby('trade_date', group_keys=False).apply(lambda x: x.nlargest(1000, 'return_20'))\n",
"\n",
"train_data, test_data = train_data.replace([np.inf, -np.inf], np.nan), test_data.replace([np.inf, -np.inf], np.nan)\n"
],
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Removed 123852 outliers.\n"
]
}
],
"execution_count": 12
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:40:13.500203Z",
"start_time": "2025-02-23T15:40:13.345289Z"
}
},
"cell_type": "code",
"source": [
"feature_columns = [col for col in train_data.columns if col not in ['trade_date',\n",
" 'ts_code',\n",
" 'label']]\n",
"feature_columns = [col for col in feature_columns if 'future' not in col]\n",
"feature_columns = [col for col in feature_columns if 'score' not in col]\n",
"feature_columns = [col for col in feature_columns if col not in origin_columns]\n",
"feature_columns = [col for col in feature_columns if not col.startswith('_')]\n",
"print(feature_columns)"
],
"id": "93d47ef451968346",
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"['turnover_rate', 'pe_ttm', 'volume_ratio', 'cat_l2_code', 'log_close', 'up', 'down', 'atr_14', 'atr_6', 'obv', 'maobv_6', 'obv-maobv_6', 'rsi_3', 'rsi_6', 'rsi_9', 'return_5', 'return_10', 'return_20', 'avg_close_5', 'std_return_5', 'std_return_15', 'std_return_25', 'std_return_90', 'std_return_90_2', 'std_return_5 / std_return_90', 'std_return_5 / std_return_25', 'std_return_90 - std_return_90_2', 'ema_5', 'ema_13', 'ema_20', 'ema_60', 'act_factor1', 'act_factor2', 'act_factor3', 'act_factor4', 'cat_af1', 'cat_af2', 'cat_af3', 'cat_af4', 'act_factor5', 'act_factor6', 'rank_act_factor1', 'rank_act_factor2', 'rank_act_factor3', 'active_buy_volume_large', 'active_buy_volume_big', 'active_buy_volume_small', 'buy_lg_vol_minus_sell_lg_vol', 'buy_elg_vol_minus_sell_elg_vol', 'log(circ_mv)', 'alpha_022', 'alpha_003', 'alpha_007', 'alpha_013', 'industry_obv', 'industry_return_5', 'industry_obv_deviation', 'industry_return_5_deviation', 'industry_return_5_percentile', '000852.SH_MACD', '000905.SH_MACD', '399006.SZ_MACD', '000852.SH_MACD_hist', '000905.SH_MACD_hist', '399006.SZ_MACD_hist', '000852.SH_RSI', '000905.SH_RSI', '399006.SZ_RSI', '000852.SH_Signal_line', '000905.SH_Signal_line', '399006.SZ_Signal_line', '000852.SH_daily_return', '000905.SH_daily_return', '399006.SZ_daily_return']\n"
]
}
],
"execution_count": 13
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:40:34.827593Z",
"start_time": "2025-02-23T15:40:13.547615Z"
}
},
"cell_type": "code",
"source": [
"# feature_columns_new = feature_columns[:]\n",
"train_data, feature_columns_new = create_deviation_within_dates(train_data, feature_columns)\n",
"print(f'feature_columns size: {len(feature_columns_new)}')\n",
"test_data, feature_columns_new = create_deviation_within_dates(test_data, feature_columns)\n",
"print(f'feature_columns size: {len(feature_columns_new)}')\n",
"\n",
"train_data = train_data.dropna(subset=feature_columns_new)\n",
"train_data = train_data.dropna(subset=['label'])\n",
"train_data = train_data.reset_index(drop=True)\n",
"\n",
"test_data = test_data.dropna(subset=feature_columns_new)\n",
"test_data = test_data.dropna(subset=['label'])\n",
"test_data = test_data.reset_index(drop=True)\n",
"\n",
"print(len(train_data))\n",
"print(f\"最小日期: {train_data['trade_date'].min().strftime('%Y-%m-%d')}\")\n",
"print(f\"最大日期: {train_data['trade_date'].max().strftime('%Y-%m-%d')}\")\n",
"print(len(test_data))\n",
"print(f\"最小日期: {test_data['trade_date'].min().strftime('%Y-%m-%d')}\")\n",
"print(f\"最大日期: {test_data['trade_date'].max().strftime('%Y-%m-%d')}\")"
],
"id": "572576eea818c865",
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"feature_columns size: 202\n",
"feature_columns size: 202\n",
"807303\n",
"最小日期: 2017-03-21\n",
"最大日期: 2022-12-30\n",
"297608\n",
"最小日期: 2023-01-03\n",
"最大日期: 2025-02-10\n"
]
}
],
"execution_count": 14
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:40:35.116436Z",
"start_time": "2025-02-23T15:40:34.974047Z"
}
},
"cell_type": "code",
"source": [
"cat_columns = [col for col in df.columns if col.startswith('cat')]\n",
"for col in cat_columns:\n",
" train_data[col] = train_data[col].astype('category')\n",
" test_data[col] = test_data[col].astype('category')"
],
"id": "2d7e37432f551aea",
"outputs": [],
"execution_count": 15
},
{
"cell_type": "code",
"id": "8f134d435f71e9e2",
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:40:36.646379Z",
"start_time": "2025-02-23T15:40:35.116436Z"
}
},
"source": [
"import lightgbm as lgb\n",
"import numpy as np\n",
"import matplotlib.pyplot as plt\n",
"import optuna\n",
"from sklearn.model_selection import KFold\n",
"from sklearn.metrics import mean_absolute_error\n",
"import os\n",
"import json\n",
"import pickle\n",
"import hashlib\n",
"from catboost import Pool\n",
"\n",
"def train_light_model(train_data_df, test_data_df, params, feature_columns, callbacks, evals,\n",
" print_feature_importance=True, num_boost_round=100,\n",
" use_optuna=False):\n",
" train_data_df, test_data_df = train_data_df.dropna(subset=['label']), test_data_df.dropna(subset=['label'])\n",
" X_train = train_data_df[feature_columns]\n",
" y_train = train_data_df['label']\n",
"\n",
" X_val = test_data_df[feature_columns]\n",
" y_val = test_data_df['label']\n",
"\n",
" categorical_feature = [i for i, col in enumerate(feature_columns) if col.startswith('cat')]\n",
" print(f'categorical_feature: {categorical_feature}')\n",
" train_data = lgb.Dataset(X_train, label=y_train, categorical_feature=categorical_feature)\n",
" val_data = lgb.Dataset(X_val, label=y_val, categorical_feature=categorical_feature)\n",
" model = lgb.train(\n",
" params, train_data, num_boost_round=num_boost_round,\n",
" valid_sets=[train_data, val_data], valid_names=['train', 'valid'],\n",
" callbacks=callbacks\n",
" )\n",
"\n",
" if print_feature_importance:\n",
" lgb.plot_metric(evals)\n",
" # lgb.plot_tree(model, figsize=(20, 8))\n",
" lgb.plot_importance(model, importance_type='split', max_num_features=20)\n",
" plt.show()\n",
" return model\n",
"\n",
"\n",
"from catboost import CatBoostRegressor\n",
"import pandas as pd\n",
"\n",
"\n",
"def train_catboost(train_data_df, test_data_df, feature_columns, params=None, plot=False):\n",
"\n",
" train_data_df, test_data_df = train_data_df.dropna(subset=['label']), test_data_df.dropna(subset=['label'])\n",
" X_train = train_data_df[feature_columns]\n",
" y_train = train_data_df['label']\n",
"\n",
" X_val = test_data_df[feature_columns]\n",
" y_val = test_data_df['label']\n",
"\n",
" cat_features = [i for i, col in enumerate(feature_columns) if col.startswith('cat')]\n",
" print(f'cat_features: {cat_features}')\n",
" train_pool = Pool(data=X_train, label=y_train, cat_features=cat_features)\n",
" val_pool = Pool(data=X_val, label=y_val, cat_features=cat_features)\n",
"\n",
" model = CatBoostRegressor(**params)\n",
" model.fit(train_pool,\n",
" eval_set=[train_pool, val_pool])\n",
"\n",
" return model"
],
"outputs": [],
"execution_count": 16
},
{
"cell_type": "code",
"id": "4a4542e1ed6afe7d",
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:40:36.799266Z",
"start_time": "2025-02-23T15:40:36.714195Z"
}
},
"source": [
"light_params = {\n",
" 'objective': 'regression',\n",
" 'metric': 'l2',\n",
" # 'objective': 'quantile', # 分位回归\n",
" # 'metric': 'quantile', # 使用 quantile 作为评估指标\n",
" 'alpha': 0.75, # 90% 分位数\n",
" 'learning_rate': 0.05,\n",
" 'is_unbalance': True,\n",
" 'num_leaves': 1024,\n",
" 'min_data_in_leaf': 128,\n",
" 'max_depth': 32,\n",
" 'max_bin': 1024,\n",
" 'feature_fraction': 0.7,\n",
" 'bagging_fraction': 0.7,\n",
" 'bagging_freq': 5,\n",
" 'lambda_l1': 1,\n",
" 'lambda_l2': 1,\n",
" # 'boosting_type': 'dart',\n",
" 'verbosity': -1\n",
"}"
],
"outputs": [],
"execution_count": 17
},
{
"cell_type": "code",
"id": "beeb098799ecfa6a",
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:41:02.591899Z",
"start_time": "2025-02-23T15:40:36.877547Z"
}
},
"source": [
"print('train data size: ', len(train_data))\n",
"\n",
"evals = {}\n",
"model = train_light_model(train_data, test_data, light_params, feature_columns_new,\n",
" [lgb.log_evaluation(period=500),\n",
" lgb.callback.record_evaluation(evals),\n",
" lgb.early_stopping(50, first_metric_only=True)\n",
" ], evals,\n",
" num_boost_round=500, use_optuna=False,\n",
" print_feature_importance=True)"
],
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"train data size: 807303\n",
"categorical_feature: [3, 35, 36, 37, 38]\n",
"Training until validation scores don't improve for 50 rounds\n",
"Early stopping, best iteration is:\n",
"[24]\ttrain's l2: 0.00229628\tvalid's l2: 0.00252369\n",
"Evaluated only: l2\n"
]
},
{
"data": {
"text/plain": [
"<Figure size 640x480 with 1 Axes>"
],
"image/png": "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
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/plain": [
"<Figure size 640x480 with 1 Axes>"
],
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAz4AAAHFCAYAAAAzASwrAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8pXeV/AAAACXBIWXMAAA9hAAAPYQGoP6dpAAEAAElEQVR4nOzdeXxNx//48dfNvgrZEyIJElsoYgutWCJBElvtBB97W7XXvlXt1FItYgu6xUctLQ0aEksaxJL0E7FUrUWCEAlRWc/vD7+crytBaDQp7+fjcR/tnXmfOTNncO/cmTNHoyiKghBCCCGEEEK8wXSKuwJCCCGEEEII8brJwEcIIYQQQgjxxpOBjxBCCCGEEOKNJwMfIYQQQgghxBtPBj5CCCGEEEKIN54MfIQQQgghhBBvPBn4CCGEEEIIId54MvARQgghhBBCvPFk4COEEEIIIYR448nARwghhCjh1q9fj0ajKfA1ZsyY13LO06dPM336dC5fvvxayv87Ll++jEajYf369cVdlVcWFhbG9OnTi7saQrxV9Iq7AkIIIYQonJCQEKpUqaKV5ujo+FrOdfr0aT799FOaNm2Ki4vLaznHq3JwcODw4cNUrFixuKvyysLCwvjqq69k8CPEP0gGPkIIIcS/hIeHB3Xr1i3uavwtWVlZaDQa9PRe/SuIoaEhDRs2LMJa/XMePnyIiYlJcVdDiLeSLHUTQggh3hCbNm3Cy8sLU1NTzMzM8PPzIzY2Vivm+PHjdOvWDRcXF4yNjXFxcaF79+5cuXJFjVm/fj2dO3cGoFmzZuqyurylZS4uLvTt2zff+Zs2bUrTpk3V9/v370ej0fD1118zevRoypYti6GhIX/88QcAe/fupUWLFpQqVQoTExMaN27Mvn37XtjOgpa6TZ8+HY1Gw//+9z86d+6MhYUFlpaWjBo1iuzsbM6dO0erVq0wNzfHxcWF+fPna5WZV9dvvvmGUaNGYW9vj7GxMd7e3vmuIcBPP/2El5cXJiYmmJub07JlSw4fPqwVk1enkydP0qlTJ8qUKUPFihXp27cvX331FYDWssW8ZYVfffUVTZo0wdbWFlNTU2rUqMH8+fPJysrKd709PDw4duwY7733HiYmJlSoUIG5c+eSm5urFXvv3j1Gjx5NhQoVMDQ0xNbWljZt2nD27Fk1JjMzk5kzZ1KlShUMDQ2xsbHhP//5D7dv335hnwjxbyADHyGEEOJfIicnh+zsbK1XntmzZ9O9e3eqVavGf//7X77++mvu37/Pe++9x+nTp9W4y5cvU7lyZZYsWcKePXuYN28eiYmJ1KtXj+TkZAD8/f2ZPXs28PhL+OHDhzl8+DD+/v6vVO8JEyZw9epVVq5cyY4dO7C1teWbb77B19eXUqVKsWHDBv773/9iaWmJn59foQY/z9KlSxfeeecdtmzZwsCBA1m8eDEjR46kffv2+Pv7s23bNpo3b864cePYunVrvuMnTpzIxYsXWbNmDWvWrOHGjRs0bdqUixcvqjHfffcd7dq1o1SpUnz//fesXbuWlJQUmjZtSlRUVL4yO3bsSKVKldi8eTMrV65kypQpdOrUCUC9tocPH8bBwQGACxcu0KNHD77++mt27txJ//79WbBgAYMHD85XdlJSEj179qRXr1789NNPtG7dmgkTJvDNN9+oMffv3+fdd98lODiY//znP+zYsYOVK1fi7u5OYmIiALm5ubRr1465c+fSo0cPfv75Z+bOnUt4eDhNmzblr7/+euU+EaLEUIQQQghRooWEhChAga+srCzl6tWrip6envLxxx9rHXf//n3F3t5e6dKlyzPLzs7OVh48eKCYmpoqS5cuVdM3b96sAEpkZGS+Y5ydnZU+ffrkS/f29la8vb3V95GRkQqgNGnSRCsuPT1dsbS0VAIDA7XSc3JylHfeeUepX7/+c66Goly6dEkBlJCQEDVt2rRpCqB8/vnnWrG1atVSAGXr1q1qWlZWlmJjY6N07NgxX13r1Kmj5ObmqumXL19W9PX1lQEDBqh1dHR0VGrUqKHk5OSocffv31dsbW2VRo0a5avT1KlT87Xho48+UgrzNSwnJ0fJyspSNm7cqOjq6ip3795V87y9vRVAOXr0qNYx1apVU/z8/NT3M2bMUAAlPDz8mef5/vvvFUDZsmWLVvqxY8cUQFm+fPkL6ypESSczPkIIIcS/xMaNGzl27JjWS09Pjz179pCdnU3v3r21ZoOMjIzw9vZm//79ahkPHjxg3LhxVKpUCT09PfT09DAzMyM9PZ0zZ868lnq///77Wu+jo6O5e/cuffr00apvbm4urVq14tixY6Snp7/SuQICArTeV61aFY1GQ+vWrdU0PT09KlWqpLW8L0+PHj3QaDTqe2dnZxo1akRkZCQA586d48aNGwQFBaGj839fo8zMzHj//fc5cuQIDx8+fG77XyQ2Npa2bdtiZWWFrq4u+vr69O7dm5ycHH7//XetWHt7e+rXr6+VVrNmTa227dq1C3d3d3x8fJ55zp07d1K6dGkCAwO1+qRWrVrY29tr/RkS4t9KNjcQQggh/iWqVq1a4OYGN2/eBKBevXoFHvfkF/QePXqwb98+pkyZQr169ShVqhQajYY2bdq8tuVMeUu4nq5v3nKvgty9exdTU9OXPpelpaXWewMDA0xMTDAyMsqXnpaWlu94e3v7AtN+++03AO7cuQPkbxM83mEvNzeXlJQUrQ0MCop9lqtXr/Lee+9RuXJlli5diouLC0ZGRsTExPDRRx/l6yMrK6t8ZRgaGmrF3b59m/Llyz/3vDdv3uTevXsYGBgUmJ+3DFKIfzMZ+AghhBD/ctbW1gD88MMPODs7PzMuNTWVnTt3Mm3aNMaPH6+mZ2RkcPfu3UKfz8jIiIyMjHzpycnJal2e9OQMypP1XbZs2TN3Z7Ozsyt0fYpSUlJSgWl5A4y8/+bdG/OkGzduoKOjQ5kyZbTSn27/82zfvp309HS2bt2q1ZdxcXGFLuNpNjY2XLt27bkx1tbWWFlZsXv37gLzzc3NX/n8QpQUMvARQggh/uX8/PzQ09PjwoULz11WpdFoUBQFQ0NDrfQ1a9aQk5OjlZYXU9AskIuLC//73/+00n7//XfOnTtX4MDnaY0bN6Z06dKcPn2aoUOHvjD+n/T9998zatQodbBy5coVoqOj6d27NwCVK1embNmyfPfdd4wZM0aNS09PZ8uWLepOby/y5PU1NjZW0/PKe7KPFEVh9erVr9ym1q1bM3XqVCIiImjevHmBMQEBAYSGhpKTk0ODBg1e+VxClGQy8BFCCCH+5VxcXJgxYwaTJk3i4sWLtGrVijJlynDz5k1iYmIwNTXl008/pVSpUjRp0oQFCxZgbW2Ni4sLBw4cYO3atZQuXVqrTA8PDwBWrVqFubk5RkZGuLq6YmVlRVBQEL169eLDDz/k/fff58qVK8yfPx8bG5tC1dfMzIxly5bRp08f7t69S6dOnbC1teX27dv89ttv3L59mxUrVhT1ZSqUW7du0aFDBwYOHEhqairTpk3DyMiICRMmAI+XDc6fP5+ePXsSEBDA4MGDycjIYMGCBdy7d4+5c+cW6jw1atQAYN68ebRu3RpdXV1q1qxJy5YtMTAwoHv37owdO5ZHjx6xYsUKUlJSXrlNI0aMYNOmTbRr147x48dTv359/vrrLw4cOEBAQADNmjWjW7dufPvtt7Rp04bhw4dTv3599PX1uXbtGpGRkbRr144OHTq8ch2EKAlkcwMhhBDiDTBhwgR++OEHfv/9d/r06YOfnx9jx47lypUrNGnSRI377rvvaNasGWPHjqVjx44cP36c8PBwLCwstMpzdXVlyZIl/PbbbzRt2pR69eqxY8cO4PF9QvPnz2fPnj0EBASwYsUKVqxYgbu7e6Hr26tXLyIjI3nw4AGDBw/Gx8eH4cOHc/LkSVq0aFE0F+UVzJ49G2dnZ/7zn//Qr18/HBwciIyMpGLFimpMjx492L59O3fu3KFr16785z//oVSpUkRGRvLuu+8W6jw9evRgwIABLF++HC8vL+rVq8eNGzeoUqUKW7ZsISUlhY4dO/L
},
"metadata": {},
"output_type": "display_data"
}
],
"execution_count": 18
},
{
"cell_type": "code",
"id": "63235069-dc59-48fb-961a-e80373e41a61",
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:41:02.758847Z",
"start_time": "2025-02-23T15:41:02.670568Z"
}
},
"source": [
"print('train data size: ', len(train_data))\n",
"\n",
"catboost_params = {\n",
" 'loss_function': 'MAE', # 90% 分位回归\n",
" 'iterations': 5000, # 训练轮数\n",
" 'learning_rate': 0.05, # 学习率，较低以防止过拟合\n",
" 'depth': 10, # 树的深度，防止过拟合\n",
" # 'l1_leaf_reg': 10.0, # l1 正则化，提高泛化能力\n",
" # 'bagging_temperature': 1, # 降低过拟合\n",
" # 'subsample': 0.8, # 每轮随机 80% 的样本，减少过拟合\n",
" 'colsample_bylevel': 0.8, # 每层 80% 特征子集，防止过拟合\n",
" 'random_seed': 42, # 固定随机种子，保证可复现\n",
" 'verbose': 500, # 每 100 轮打印一次信息\n",
" 'early_stopping_rounds': 100, # 早停，防止过拟合\n",
" # 'task_type': 'GPU'\n",
"}\n",
"\n",
"# model = train_catboost(train_data, test_data, feature_columns_new, catboost_params)"
],
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"train data size: 807303\n"
]
}
],
"execution_count": 19
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:41:02.928981Z",
"start_time": "2025-02-23T15:41:02.836475Z"
}
},
"cell_type": "code",
"source": [
"from tqdm import tqdm\n",
"\n",
"def incremental_training(test_data: pd.DataFrame,\n",
" model,\n",
" days: int,\n",
" back_days: int,\n",
" feature_columns: list,\n",
" params: dict,\n",
" model_type: str = 'lightgbm'):\n",
" if model_type not in ['lightgbm', 'catboost']:\n",
" raise ValueError(\"model_type must be either 'lightgbm' or 'catboost'\")\n",
"\n",
" test_data = test_data.sort_values(by='trade_date')\n",
" scores = []\n",
" unique_trade_dates = sorted(test_data['trade_date'].unique())\n",
"\n",
" new_model = None\n",
" for i in tqdm(range(0, len(unique_trade_dates), days)):\n",
" # Get the current window of trade dates\n",
" current_dates = unique_trade_dates[i:i + days]\n",
" window_data = test_data[test_data['trade_date'].isin(current_dates)]\n",
" X = window_data[feature_columns]\n",
"\n",
" if new_model is not None:\n",
" window_scores = new_model.predict(X, prediction_type='RawFormulaVal')\n",
" else:\n",
" window_scores = model.predict(X, prediction_type='RawFormulaVal')\n",
" scores.extend(window_scores)\n",
"\n",
" # Prepare data for incremental training\n",
" current_dates = unique_trade_dates[max(0, i - back_days):i + days]\n",
" window_data = test_data[test_data['trade_date'].isin(current_dates)]\n",
" X_train = window_data[feature_columns]\n",
" y_train = window_data['label'] # Assuming 'label' is what you're predicting\n",
" # Incrementally train the model\n",
" if len(y_train.unique()) > 1:\n",
" if model_type == 'lightgbm':\n",
" categorical_feature = [i for i, col in enumerate(feature_columns) if col.startswith('cat')]\n",
" train_data = lgb.Dataset(X_train, label=y_train, categorical_feature=categorical_feature)\n",
" new_model = lgb.train(params,\n",
" train_set=train_data,\n",
" num_boost_round=100,\n",
" init_model=model,\n",
" keep_training_booster=True)\n",
" elif model_type == 'catboost':\n",
" from catboost import Pool\n",
" train_data = Pool(data=X_train, label=y_train, cat_features=[col for col in feature_columns if col.startswith('cat')])\n",
" # model.set_params(**params)\n",
" model.fit(train_data, init_model=model)\n",
" else:\n",
" print(current_dates)\n",
"\n",
" # Add the scores as a new 'score' column to the test_data\n",
" test_data['score'] = scores\n",
" return test_data"
],
"id": "465944b1d463e4b1",
"outputs": [],
"execution_count": 20
},
{
"cell_type": "code",
"id": "7359f89064a124d2",
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:41:43.347794Z",
"start_time": "2025-02-23T15:41:03.010181Z"
}
},
"source": [
"\n",
"predictions_test = incremental_training(test_data, model, 5, 0, feature_columns_new, light_params, model_type='lightgbm')\n",
"predictions_test = predictions_test.loc[predictions_test.groupby('trade_date')['score'].idxmax()]\n",
"predictions_test[['trade_date', 'score', 'ts_code']].to_csv('predictions_test.tsv', index=False)\n"
],
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"100%|██████████| 101/101 [00:39<00:00, 2.53it/s]\n"
]
}
],
"execution_count": 21
},
{
"cell_type": "code",
"id": "b427ce41-9739-4e9e-bea8-5f2551fec5d7",
"metadata": {
"ExecuteTime": {
"end_time": "2025-02-23T15:41:43.468762Z",
"start_time": "2025-02-23T15:41:43.453009Z"
}
},
"source": [],
"outputs": [],
"execution_count": null
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.19"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
Reference in New Issue Copy Permalink