{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "## 1. 导入依赖" ] }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "import os\n", "from datetime import datetime\n", "from typing import List\n", "\n", "import polars as pl\n", "\n", "from src.factors import FactorEngine\n", "from src.training import (\n", " DateSplitter,\n", " LightGBMModel,\n", " STFilter,\n", " StandardScaler,\n", " # StockFilterConfig, # 已删除，使用 StockPoolManager + filter_func 替代\n", " StockPoolManager,\n", " Trainer,\n", " Winsorizer,\n", " NullFiller,\n", " check_data_quality,\n", ")\n", "from src.training.config import TrainingConfig\n", "\n" ] }, { "metadata": {}, "cell_type": "markdown", "source": "## 2. 定义辅助函数" }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "def register_factors(\n", " engine: FactorEngine,\n", " selected_factors: List[str],\n", " factor_definitions: dict,\n", " label_factor: dict,\n", ") -> List[str]:\n", " \"\"\"注册因子（selected_factors 从 metadata 查询，factor_definitions 用 DSL 表达式注册）\"\"\"\n", " print(\"=\" * 80)\n", " print(\"注册因子\")\n", " print(\"=\" * 80)\n", "\n", " # 注册 SELECTED_FACTORS 中的因子（已在 metadata 中）\n", " print(\"\\n注册特征因子（从 metadata）:\")\n", " for name in selected_factors:\n", " engine.add_factor(name)\n", " print(f\" - {name}\")\n", "\n", " # 注册 FACTOR_DEFINITIONS 中的因子（通过表达式，尚未在 metadata 中）\n", " print(\"\\n注册特征因子（表达式）:\")\n", " for name, expr in factor_definitions.items():\n", " engine.add_factor(name, expr)\n", " print(f\" - {name}: {expr}\")\n", "\n", " # 注册 label 因子（通过表达式）\n", " print(\"\\n注册 Label 因子（表达式）:\")\n", " for name, expr in label_factor.items():\n", " engine.add_factor(name, expr)\n", " print(f\" - {name}: {expr}\")\n", "\n", " # 特征列 = SELECTED_FACTORS + FACTOR_DEFINITIONS 的 keys\n", " feature_cols = selected_factors + list(factor_definitions.keys())\n", "\n", " print(f\"\\n特征因子数: {len(feature_cols)}\")\n", " print(f\" - 来自 metadata: {len(selected_factors)}\")\n", " print(f\" - 来自表达式: {len(factor_definitions)}\")\n", " print(f\"Label: {list(label_factor.keys())[0]}\")\n", " print(f\"已注册因子总数: {len(engine.list_registered())}\")\n", "\n", " return feature_cols\n", "\n", "\n", "def prepare_data(\n", " engine: FactorEngine,\n", " feature_cols: List[str],\n", " start_date: str,\n", " end_date: str,\n", ") -> pl.DataFrame:\n", " print(\"\\n\" + \"=\" * 80)\n", " print(\"准备数据\")\n", " print(\"=\" * 80)\n", "\n", " # 计算因子（全市场数据）\n", " print(f\"\\n计算因子: {start_date} - {end_date}\")\n", " factor_names = feature_cols + [LABEL_NAME] # 包含 label\n", "\n", " data = engine.compute(\n", " factor_names=factor_names,\n", " start_date=start_date,\n", " end_date=end_date,\n", " )\n", "\n", " print(f\"数据形状: {data.shape}\")\n", " print(f\"数据列: {data.columns}\")\n", " print(f\"\\n前5行预览:\")\n", " print(data.head())\n", "\n", " return data\n", "\n" ] }, { "metadata": {}, "cell_type": "markdown", "source": [ "## 3. 配置参数\n", "#\n", "### 3.1 因子定义" ] }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "# 特征因子定义字典：新增因子只需在此处添加一行\n", "LABEL_NAME = \"future_return_5\"\n", "\n", "# 当前选择的因子列表（从 FACTOR_DEFINITIONS 中选择要使用的因子）\n", "SELECTED_FACTORS = [\n", " # ================= 1. 价格、趋势与路径依赖 =================\n", " \"ma_5\",\n", " \"ma_20\",\n", " \"ma_ratio_5_20\",\n", " \"bias_10\",\n", " \"high_low_ratio\",\n", " \"bbi_ratio\",\n", " \"return_5\",\n", " \"return_20\",\n", " \"kaufman_ER_20\",\n", " \"mom_acceleration_10_20\",\n", " \"drawdown_from_high_60\",\n", " \"up_days_ratio_20\",\n", " # ================= 2. 波动率、风险调整与高阶矩 =================\n", " \"volatility_5\",\n", " \"volatility_20\",\n", " \"volatility_ratio\",\n", " \"std_return_20\",\n", " \"sharpe_ratio_20\",\n", " \"min_ret_20\",\n", " \"volatility_squeeze_5_60\",\n", " # ================= 3. 日内微观结构与异象 =================\n", " \"overnight_intraday_diff\",\n", " \"upper_shadow_ratio\",\n", " \"capital_retention_20\",\n", " \"max_ret_20\",\n", " # ================= 4. 量能、流动性与量价背离 =================\n", " \"volume_ratio_5_20\",\n", " \"turnover_rate_mean_5\",\n", " \"turnover_deviation\",\n", " \"amihud_illiq_20\",\n", " \"turnover_cv_20\",\n", " \"pv_corr_20\",\n", " \"close_vwap_deviation\",\n", " # ================= 5. 基本面财务特征 =================\n", " \"roe\",\n", " \"roa\",\n", " \"profit_margin\",\n", " \"debt_to_equity\",\n", " \"current_ratio\",\n", " \"net_profit_yoy\",\n", " \"revenue_yoy\",\n", " \"healthy_expansion_velocity\",\n", " # ================= 6. 基本面估值与截面动量共振 =================\n", " \"EP\",\n", " \"BP\",\n", " \"CP\",\n", " \"market_cap_rank\",\n", " \"turnover_rank\",\n", " \"return_5_rank\",\n", " \"EP_rank\",\n", " \"pe_expansion_trend\",\n", " \"value_price_divergence\",\n", " \"active_market_cap\",\n", " \"ebit_rank\",\n", "]\n", "\n", "# 因子定义字典（完整因子库）\n", "FACTOR_DEFINITIONS = {\n", " # ================= 1. 价格、趋势与路径依赖 (Trend, Momentum & Path Dependency) =================\n", " \"ma_5\": \"ts_mean(close, 5)\",\n", " \"ma_20\": \"ts_mean(close, 20)\",\n", " \"ma_ratio_5_20\": \"ts_mean(close, 5) / (ts_mean(close, 20) + 1e-8) - 1\", # 均线发散度\n", " \"bias_10\": \"close / (ts_mean(close, 10) + 1e-8) - 1\", # 10日乖离率\n", " \"high_low_ratio\": \"(close - ts_min(low, 20)) / (ts_max(high, 20) - ts_min(low, 20) + 1e-8)\", # 威廉指标变形\n", " \"bbi_ratio\": \"(ts_mean(close, 3) + ts_mean(close, 6) + ts_mean(close, 12) + ts_mean(close, 24)) / (4 * close + 1e-8)\", # 多空指标比率\n", " \"return_5\": \"(close / (ts_delay(close, 5) + 1e-8)) - 1\", # 5日动量\n", " \"return_20\": \"(close / (ts_delay(close, 20) + 1e-8)) - 1\", # 20日动量\n", " # [高阶] Kaufman 趋势效率 (极高价值) - 衡量趋势流畅度，剔除无序震荡\n", " \"kaufman_ER_20\": \"abs(close - ts_delay(close, 20)) / (ts_sum(abs(close - ts_delay(close, 1)), 20) + 1e-8)\",\n", " # [高阶] 动量加速度 - 寻找二阶导数大于0，正在加速爆发的股票\n", " \"mom_acceleration_10_20\": \"(close / (ts_delay(close, 10) + 1e-8) - 1) - (ts_delay(close, 10) / (ts_delay(close, 20) + 1e-8) - 1)\",\n", " # [高阶] 高点距离衰减 - 衡量套牢盘压力\n", " \"drawdown_from_high_60\": \"close / (ts_max(high, 60) + 1e-8) - 1\",\n", " # [高阶] 趋势一致性 - 过去20天内收红的天数比例\n", " \"up_days_ratio_20\": \"ts_sum(close > ts_delay(close, 1), 20) / 20\",\n", " # ================= 2. 波动率、风险调整与高阶矩 (Volatility & Risk-Adjusted Returns) =================\n", " \"volatility_5\": \"ts_std(close, 5)\",\n", " \"volatility_20\": \"ts_std(close, 20)\",\n", " \"volatility_ratio\": \"ts_std(close, 5) / (ts_std(close, 20) + 1e-8)\", # 波动率期限结构\n", " \"std_return_20\": \"ts_std((close / (ts_delay(close, 1) + 1e-8)) - 1, 20)\", # 真实收益率波动率\n", " # [高阶] 夏普趋势比率 - 惩罚暴涨暴跌，奖励稳健爬坡\n", " \"sharpe_ratio_20\": \"ts_mean(close / (ts_delay(close, 1) + 1e-8) - 1, 20) / (ts_std(close / (ts_delay(close, 1) + 1e-8) - 1, 20) + 1e-8)\",\n", " # [高阶] 尾部崩盘风险 - 过去一个月最大单日跌幅\n", " \"min_ret_20\": \"ts_min(close / (ts_delay(close, 1) + 1e-8) - 1, 20)\",\n", " # [高阶] 波动率挤压比 - 寻找盘整到极致面临变盘的股票 (布林带收口)\n", " \"volatility_squeeze_5_60\": \"ts_std(close, 5) / (ts_std(close, 60) + 1e-8)\",\n", " # ================= 3. 日内微观结构与异象 (Intraday Microstructure & Anomalies) =================\n", " # [高阶] 隔夜与日内背离 - 差值越小说明主力越喜欢在盘中吸筹\n", " \"overnight_intraday_diff\": \"(open / (ts_delay(close, 1) + 1e-8) - 1) - (close / (open + 1e-8) - 1)\",\n", " # [高阶] 上影线抛压极值 - 冲高回落被套牢的概率\n", " \"upper_shadow_ratio\": \"(high - ((open + close + abs(open - close)) / 2)) / (high - low + 1e-8)\",\n", " # [高阶] 资金沉淀率 - 衡量主力日内高抛低吸洗盘的剧烈程度\n", " \"capital_retention_20\": \"ts_sum(abs(close - open), 20) / (ts_sum(high - low, 20) + 1e-8)\",\n", " # [高阶] MAX 彩票效应 - 反转因子，剔除近期有过妖股连板特征的标的\n", " \"max_ret_20\": \"ts_max(close / (ts_delay(close, 1) + 1e-8) - 1, 20)\",\n", " # ================= 4. 量能、流动性与量价背离 (Volume, Liquidity & Divergence) =================\n", " \"volume_ratio_5_20\": \"ts_mean(vol, 5) / (ts_mean(vol, 20) + 1e-8)\", # 相对放量比\n", " \"turnover_rate_mean_5\": \"ts_mean(turnover_rate, 5)\", # 活跃度\n", " \"turnover_deviation\": \"(turnover_rate - ts_mean(turnover_rate, 10)) / (ts_std(turnover_rate, 10) + 1e-8)\", # 换手率偏离度\n", " # [高阶] Amihud 非流动性异象 (绝对核心) - 衡量砸盘/拉升的摩擦成本\n", " \"amihud_illiq_20\": \"ts_mean(abs(close / (ts_delay(close, 1) + 1e-8) - 1) / (amount + 1e-8), 20)\",\n", " # [高阶] 换手率惩罚因子 - 换手率忽高忽低说明游资接力，行情极不稳定\n", " \"turnover_cv_20\": \"ts_std(turnover_rate, 20) / (ts_mean(turnover_rate, 20) + 1e-8)\",\n", " # [高阶] 纯粹量价相关性 - 检验是否是\"放量上涨，缩量下跌\"的良性多头\n", " \"pv_corr_20\": \"ts_corr(close / (ts_delay(close, 1) + 1e-8) - 1, vol, 20)\",\n", " # [高阶] 收盘价与均价背离 - 专门抓尾盘突袭拉升骗线的股票\n", " \"close_vwap_deviation\": \"close / (amount / (vol * 100 + 1e-8) + 1e-8) - 1\",\n", " # ================= 5. 基本面财务特征 (Fundamental Quality & Structure) =================\n", " \"roe\": \"n_income / (total_hldr_eqy_exc_min_int + 1e-8)\", # 净资产收益率\n", " \"roa\": \"n_income / (total_assets + 1e-8)\", # 总资产收益率\n", " \"profit_margin\": \"n_income / (revenue + 1e-8)\", # 销售净利率\n", " \"debt_to_equity\": \"total_liab / (total_hldr_eqy_exc_min_int + 1e-8)\", # 杠杆率\n", " \"current_ratio\": \"total_cur_assets / (total_cur_liab + 1e-8)\", # 短期偿债安全垫\n", " # [高阶] 利润同比增速 (日频延后252天等于去年同期)\n", " \"net_profit_yoy\": \"(n_income / (ts_delay(n_income, 252) + 1e-8)) - 1\",\n", " # [高阶] 营收同比增速\n", " \"revenue_yoy\": \"(revenue / (ts_delay(revenue, 252) + 1e-8)) - 1\",\n", " # [高阶] 资产负债表扩张斜率 - 剔除单纯靠举债扩张的公司\n", " \"healthy_expansion_velocity\": \"(total_assets / (ts_delay(total_assets, 252) + 1e-8) - 1) - (total_liab / (ts_delay(total_liab, 252) + 1e-8) - 1)\",\n", " # ================= 6. 基本面估值与截面动量共振 (Valuation & Cross-Sectional Ranking) =================\n", " # 估值水平绝对值 (Tushare 市值单位需要 * 10000 转换为元)\n", " \"EP\": \"n_income / (total_mv * 10000 + 1e-8)\", # 盈利收益率 (1/PE)\n", " \"BP\": \"total_hldr_eqy_exc_min_int / (total_mv * 10000 + 1e-8)\", # 账面市值比 (1/PB)\n", " \"CP\": \"n_cashflow_act / (total_mv * 10000 + 1e-8)\", # 经营现金流收益率 (1/PCF)\n", " # 全市场截面排名因子\n", " \"market_cap_rank\": \"cs_rank(total_mv)\", # 规模因子 (Size)\n", " \"turnover_rank\": \"cs_rank(turnover_rate)\",\n", " \"return_5_rank\": \"cs_rank((close / (ts_delay(close, 5) + 1e-8)) - 1)\",\n", " \"EP_rank\": \"cs_rank(n_income / (total_mv + 1e-8))\", # 谁最便宜\n", " # [高阶] 戴维斯双击动量 - 估值相对上一年是否在扩张\n", " \"pe_expansion_trend\": \"(total_mv / (n_income + 1e-8)) / (ts_delay(total_mv, 60) / (ts_delay(n_income, 60) + 1e-8) + 1e-8) - 1\",\n", " # [高阶] 业绩与价格背离度 - 截面做差：利润排名全市场第一，但近20日价格排名倒数第一，捕捉被错杀的潜伏股\n", " \"value_price_divergence\": \"cs_rank((n_income - ts_delay(n_income, 252)) / (abs(ts_delay(n_income, 252)) + 1e-8)) - cs_rank(close / (ts_delay(close, 20) + 1e-8))\",\n", " # [高阶] 流动性溢价调整后市值 - 识别僵尸大盘股和极度活跃的小微盘\n", " \"active_market_cap\": \"total_mv * ts_mean(turnover_rate, 20)\",\n", " \"ebit_rank\": \"cs_rank(ebit)\",\n", "}\n", "\n", "# Label 因子定义（不参与训练，用于计算目标）\n", "LABEL_FACTOR = {\n", " LABEL_NAME: \"(ts_delay(close, -5) / ts_delay(open, -1)) - 1\", # 未来5日收益率\n", "}" ] }, { "metadata": {}, "cell_type": "markdown", "source": "### 3.2 训练参数配置" }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "# 日期范围配置（正确的 train/val/test 三分法）\n", "# Train: 用于训练模型参数\n", "# Val: 用于验证/早停/调参（位于 train 之后，test 之前）\n", "# Test: 仅用于最终评估，完全独立于训练过程\n", "TRAIN_START = \"20200101\"\n", "TRAIN_END = \"20231231\"\n", "VAL_START = \"20240101\"\n", "VAL_END = \"20241231\"\n", "TEST_START = \"20250101\"\n", "TEST_END = \"20261231\"\n", "\n", "# 模型参数配置\n", "MODEL_PARAMS = {\n", " \"objective\": \"regression\",\n", " \"metric\": \"mae\", # 改为 MAE，对异常值更稳健\n", " # 树结构控制（防过拟合核心）\n", " # \"num_leaves\": 20, # 从31降为20，降低模型复杂度\n", " # \"max_depth\": 16, # 显式限制深度，防止过度拟合噪声\n", " # \"min_child_samples\": 50, # 叶子最小样本数，防止学习极端样本\n", " # \"min_child_weight\": 0.001,\n", " # 学习参数\n", " \"learning_rate\": 0.01, # 降低学习率，配合更多树\n", " \"n_estimators\": 1000, # 增加树数量，配合早停\n", " # 采样策略（关键防过拟合）\n", " \"subsample\": 0.8, # 每棵树随机采样80%数据（行采样）\n", " \"subsample_freq\": 5, # 每5轮迭代进行一次 subsample\n", " \"colsample_bytree\": 0.8, # 每棵树随机选择80%特征（列采样）\n", " # 正则化\n", " \"reg_alpha\": 0.1, # L1正则，增加稀疏性\n", " \"reg_lambda\": 1.0, # L2正则，平滑权重\n", " # 数值稳定性\n", " \"verbose\": -1,\n", " \"random_state\": 42,\n", "}\n", "\n", "\n", "# 股票池筛选函数\n", "# 使用新的 StockPoolManager API：传入自定义筛选函数和所需列/因子\n", "# 筛选函数接收单日 DataFrame，返回布尔 Series\n", "#\n", "# 筛选逻辑（针对单日数据）：\n", "# 1. 先排除创业板、科创板、北交所（ST过滤由STFilter组件处理）\n", "# 2. 然后选取市值最小的500只股票\n", "def stock_pool_filter(df: pl.DataFrame) -> pl.Series:\n", " \"\"\"股票池筛选函数（单日数据）\n", "\n", " 筛选条件：\n", " 1. 排除创业板（代码以 300 开头）\n", " 2. 排除科创板（代码以 688 开头）\n", " 3. 排除北交所（代码以 8、9 或 4 开头）\n", " 4. 选取当日市值最小的500只股票\n", " \"\"\"\n", " # 代码筛选（排除创业板、科创板、北交所）\n", " code_filter = (\n", " ~df[\"ts_code\"].str.starts_with(\"30\") # 排除创业板\n", " & ~df[\"ts_code\"].str.starts_with(\"68\") # 排除科创板\n", " & ~df[\"ts_code\"].str.starts_with(\"8\") # 排除北交所\n", " & ~df[\"ts_code\"].str.starts_with(\"9\") # 排除北交所\n", " & ~df[\"ts_code\"].str.starts_with(\"4\") # 排除北交所\n", " )\n", "\n", " # 在已筛选的股票中，选取市值最小的500只\n", " # 按市值升序排序，取前500\n", " valid_df = df.filter(code_filter)\n", " n = min(1000, len(valid_df))\n", " small_cap_codes = valid_df.sort(\"total_mv\").head(n)[\"ts_code\"]\n", "\n", " # 返回布尔 Series：是否在被选中的股票中\n", " return df[\"ts_code\"].is_in(small_cap_codes)\n", "\n", "\n", "# 定义筛选所需的基础列\n", "STOCK_FILTER_REQUIRED_COLUMNS = [\"total_mv\"] # ST过滤由STFilter组件处理\n", "\n", "# 可选：定义筛选所需的因子（如果需要用因子进行筛选）\n", "# STOCK_FILTER_REQUIRED_FACTORS = {\n", "# \"market_cap_rank\": \"cs_rank(total_mv)\",\n", "# }\n", "\n", "\n", "# 输出配置（相对于本文件所在目录）\n", "OUTPUT_DIR = \"output\"\n", "SAVE_PREDICTIONS = True\n", "PERSIST_MODEL = False\n", "\n", "# Top N 配置：每日推荐股票数量\n", "TOP_N = 5 # 可调整为 10, 20 等" ] }, { "metadata": {}, "cell_type": "markdown", "source": [ "## 4. 训练流程\n", "#\n", "### 4.1 初始化组件" ] }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "print(\"\\n\" + \"=\" * 80)\n", "print(\"LightGBM 回归模型训练\")\n", "print(\"=\" * 80)\n", "\n", "# 1. 创建 FactorEngine（启用 metadata 功能）\n", "print(\"\\n[1] 创建 FactorEngine\")\n", "engine = FactorEngine(metadata_path=\"data/factors.jsonl\")\n", "\n", "# 2. 使用 metadata 定义因子\n", "print(\"\\n[2] 定义因子（从 metadata 注册）\")\n", "feature_cols = register_factors(\n", " engine, SELECTED_FACTORS, FACTOR_DEFINITIONS, LABEL_FACTOR\n", ")\n", "target_col = LABEL_NAME\n", "\n", "# 3. 准备数据（使用模块级别的日期配置）\n", "print(\"\\n[3] 准备数据\")\n", "\n", "data = prepare_data(\n", " engine=engine,\n", " feature_cols=feature_cols,\n", " start_date=TRAIN_START,\n", " end_date=TEST_END,\n", ")\n", "\n", "# 4. 打印配置信息\n", "print(f\"\\n[配置] 训练期: {TRAIN_START} - {TRAIN_END}\")\n", "print(f\"[配置] 验证期: {VAL_START} - {VAL_END}\")\n", "print(f\"[配置] 测试期: {TEST_START} - {TEST_END}\")\n", "print(f\"[配置] 特征数: {len(feature_cols)}\")\n", "print(f\"[配置] 目标变量: {target_col}\")\n", "\n", "# 5. 创建模型\n", "model = LightGBMModel(params=MODEL_PARAMS)\n", "\n", "# 6. 创建数据处理器（使用函数返回的完整特征列表）\n", "processors = [\n", " NullFiller(feature_cols=feature_cols, strategy=\"mean\"),\n", " Winsorizer(feature_cols=feature_cols, lower=0.01, upper=0.99),\n", " StandardScaler(feature_cols=feature_cols),\n", "]\n", "\n", "# 7. 创建数据划分器（正确的 train/val/test 三分法）\n", "# Train: 训练模型参数 | Val: 验证/早停 | Test: 最终评估\n", "splitter = DateSplitter(\n", " train_start=TRAIN_START,\n", " train_end=TRAIN_END,\n", " val_start=VAL_START,\n", " val_end=VAL_END,\n", " test_start=TEST_START,\n", " test_end=TEST_END,\n", ")\n", "\n", "# 8. 创建股票池管理器\n", "# 使用新的 API：传入自定义筛选函数和所需列\n", "pool_manager = StockPoolManager(\n", " filter_func=stock_pool_filter,\n", " required_columns=STOCK_FILTER_REQUIRED_COLUMNS, # 筛选所需的额外列\n", " # required_factors=STOCK_FILTER_REQUIRED_FACTORS, # 可选：筛选所需的因子\n", " data_router=engine.router,\n", ")\n", "print(\"[股票池筛选] 使用自定义函数进行股票池筛选\")\n", "print(f\"[股票池筛选] 所需基础列: {STOCK_FILTER_REQUIRED_COLUMNS}\")\n", "print(\"[股票池筛选] 筛选逻辑: 排除创业板/科创板/北交所后，每日选市值最小的500只\")\n", "# print(f\"[股票池筛选] 所需因子: {list(STOCK_FILTER_REQUIRED_FACTORS.keys())}\")\n", "\n", "# 9. 创建 ST 股票过滤器\n", "st_filter = STFilter(\n", " data_router=engine.router,\n", ")\n", "\n", "# 10. 创建训练器\n", "trainer = Trainer(\n", " model=model,\n", " pool_manager=pool_manager,\n", " processors=processors,\n", " filters=[st_filter], # 使用STFilter过滤ST股票\n", " splitter=splitter,\n", " target_col=target_col,\n", " feature_cols=feature_cols,\n", " persist_model=PERSIST_MODEL,\n", ")" ] }, { "metadata": {}, "cell_type": "markdown", "source": "### 4.2 执行训练" }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "print(\"\\n\" + \"=\" * 80)\n", "print(\"开始训练\")\n", "print(\"=\" * 80)\n", "\n", "# 步骤 1: 股票池筛选\n", "print(\"\\n[步骤 1/6] 股票池筛选\")\n", "print(\"-\" * 60)\n", "if pool_manager:\n", " print(\" 执行每日独立筛选股票池...\")\n", " filtered_data = pool_manager.filter_and_select_daily(data)\n", " print(f\" 筛选前数据规模: {data.shape}\")\n", " print(f\" 筛选后数据规模: {filtered_data.shape}\")\n", " print(f\" 筛选前股票数: {data['ts_code'].n_unique()}\")\n", " print(f\" 筛选后股票数: {filtered_data['ts_code'].n_unique()}\")\n", " print(f\" 删除记录数: {len(data) - len(filtered_data)}\")\n", "else:\n", " filtered_data = data\n", " print(\" 未配置股票池管理器，跳过筛选\")" ] }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "# 步骤 2: 划分训练/验证/测试集（正确的三分法）\n", "print(\"\\n[步骤 2/6] 划分训练集、验证集和测试集\")\n", "print(\"-\" * 60)\n", "if splitter:\n", " # 正确的三分法：train用于训练，val用于验证/早停，test仅用于最终评估\n", " train_data, val_data, test_data = splitter.split(filtered_data)\n", " print(f\" 训练集数据规模: {train_data.shape}\")\n", " print(f\" 验证集数据规模: {val_data.shape}\")\n", " print(f\" 测试集数据规模: {test_data.shape}\")\n", " print(f\" 训练集股票数: {train_data['ts_code'].n_unique()}\")\n", " print(f\" 验证集股票数: {val_data['ts_code'].n_unique()}\")\n", " print(f\" 测试集股票数: {test_data['ts_code'].n_unique()}\")\n", " print(\n", " f\" 训练集日期范围: {train_data['trade_date'].min()} - {train_data['trade_date'].max()}\"\n", " )\n", " print(\n", " f\" 验证集日期范围: {val_data['trade_date'].min()} - {val_data['trade_date'].max()}\"\n", " )\n", " print(\n", " f\" 测试集日期范围: {test_data['trade_date'].min()} - {test_data['trade_date'].max()}\"\n", " )\n", "\n", " print(\"\\n 训练集前5行预览:\")\n", " print(train_data.head())\n", " print(\"\\n 验证集前5行预览:\")\n", " print(val_data.head())\n", " print(\"\\n 测试集前5行预览:\")\n", " print(test_data.head())\n", "else:\n", " train_data = filtered_data\n", " test_data = filtered_data\n", " print(\" 未配置划分器，全部作为训练集\")" ] }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "# 步骤 3: 数据质量检查（必须在预处理之前）\n", "print(\"\\n[步骤 3/7] 数据质量检查\")\n", "print(\"-\" * 60)\n", "print(\" [说明] 此检查在 fillna 等处理之前执行，用于发现数据问题\")\n", "\n", "print(\"\\n 检查训练集...\")\n", "check_data_quality(train_data, feature_cols, raise_on_error=True)\n", "\n", "if \"val_data\" in locals() and val_data is not None:\n", " print(\"\\n 检查验证集...\")\n", " check_data_quality(val_data, feature_cols, raise_on_error=True)\n", "\n", "print(\"\\n 检查测试集...\")\n", "check_data_quality(test_data, feature_cols, raise_on_error=True)\n", "\n", "print(\" [成功] 数据质量检查通过，未发现异常\")\n" ] }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "# 步骤 4: 训练集数据处理\n", "print(\"\\n[步骤 4/7] 训练集数据处理\")\n", "print(\"-\" * 60)\n", "fitted_processors = []\n", "if processors:\n", " for i, processor in enumerate(processors, 1):\n", " print(f\" [{i}/{len(processors)}] 应用处理器: {processor.__class__.__name__}\")\n", " train_data_before = len(train_data)\n", " train_data = processor.fit_transform(train_data)\n", " train_data_after = len(train_data)\n", " fitted_processors.append(processor)\n", " print(f\" 处理前记录数: {train_data_before}\")\n", " print(f\" 处理后记录数: {train_data_after}\")\n", " if train_data_before != train_data_after:\n", " print(f\" 删除记录数: {train_data_before - train_data_after}\")\n", "\n", "print(\"\\n 训练集处理后前5行预览:\")\n", "print(train_data.head())\n", "print(f\"\\n 训练集特征统计:\")\n", "print(f\" 特征数: {len(feature_cols)}\")\n", "print(f\" 样本数: {len(train_data)}\")\n", "print(f\" 缺失值统计:\")\n", "for col in feature_cols[:5]: # 只显示前5个特征的缺失值\n", " null_count = train_data[col].null_count()\n", " if null_count > 0:\n", " print(f\" {col}: {null_count} ({null_count / len(train_data) * 100:.2f}%)\")" ] }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "# 步骤 4: 训练模型\n", "print(\"\\n[步骤 5/7] 训练模型\")\n", "print(\"-\" * 60)\n", "print(f\" 模型类型: LightGBM\")\n", "print(f\" 训练样本数: {len(train_data)}\")\n", "print(f\" 特征数: {len(feature_cols)}\")\n", "print(f\" 目标变量: {target_col}\")\n", "\n", "X_train = train_data.select(feature_cols)\n", "y_train = train_data.select(target_col).to_series()\n", "\n", "print(f\"\\n 目标变量统计:\")\n", "print(f\" 均值: {y_train.mean():.6f}\")\n", "print(f\" 标准差: {y_train.std():.6f}\")\n", "print(f\" 最小值: {y_train.min():.6f}\")\n", "print(f\" 最大值: {y_train.max():.6f}\")\n", "print(f\" 缺失值: {y_train.null_count()}\")\n", "\n", "print(\"\\n 开始训练...\")\n", "model.fit(X_train, y_train)\n", "print(\" 训练完成！\")" ] }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "# 步骤 5: 测试集数据处理\n", "print(\"\\n[步骤 6/7] 测试集数据处理\")\n", "print(\"-\" * 60)\n", "if processors and test_data is not train_data:\n", " for i, processor in enumerate(fitted_processors, 1):\n", " print(\n", " f\" [{i}/{len(fitted_processors)}] 应用处理器: {processor.__class__.__name__}\"\n", " )\n", " test_data_before = len(test_data)\n", " test_data = processor.transform(test_data)\n", " test_data_after = len(test_data)\n", " print(f\" 处理前记录数: {test_data_before}\")\n", " print(f\" 处理后记录数: {test_data_after}\")\n", "else:\n", " print(\" 跳过测试集处理\")" ] }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "# 步骤 6: 生成预测\n", "print(\"\\n[步骤 7/7] 生成预测\")\n", "print(\"-\" * 60)\n", "X_test = test_data.select(feature_cols)\n", "print(f\" 测试样本数: {len(X_test)}\")\n", "print(\" 预测中...\")\n", "predictions = model.predict(X_test)\n", "print(f\" 预测完成！\")\n", "\n", "print(f\"\\n 预测结果统计:\")\n", "print(f\" 均值: {predictions.mean():.6f}\")\n", "print(f\" 标准差: {predictions.std():.6f}\")\n", "print(f\" 最小值: {predictions.min():.6f}\")\n", "print(f\" 最大值: {predictions.max():.6f}\")\n", "\n", "# 保存结果到 trainer\n", "trainer.results = test_data.with_columns([pl.Series(\"prediction\", predictions)])" ] }, { "metadata": {}, "cell_type": "markdown", "source": "### 4.3 训练指标曲线" }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "print(\"\\n\" + \"=\" * 80)\n", "print(\"训练指标曲线\")\n", "print(\"=\" * 80)\n", "\n", "# 重新训练以收集指标（因为之前的训练没有保存评估结果）\n", "print(\"\\n重新训练模型以收集训练指标...\")\n", "\n", "import lightgbm as lgb\n", "\n", "# 准备数据（使用 val 做验证，test 不参与训练过程）\n", "X_train_np = X_train.to_numpy()\n", "y_train_np = y_train.to_numpy()\n", "X_val_np = val_data.select(feature_cols).to_numpy()\n", "y_val_np = val_data.select(target_col).to_series().to_numpy()\n", "\n", "# 创建数据集\n", "train_dataset = lgb.Dataset(X_train_np, label=y_train_np)\n", "val_dataset = lgb.Dataset(X_val_np, label=y_val_np, reference=train_dataset)\n", "\n", "# 用于存储评估结果\n", "evals_result = {}\n", "\n", "# 使用与原模型相同的参数重新训练\n", "# 正确的三分法：train用于训练，val用于验证，test不参与训练过程\n", "# 添加早停：如果验证指标连续100轮没有改善则停止训练\n", "booster_with_eval = lgb.train(\n", " MODEL_PARAMS,\n", " train_dataset,\n", " num_boost_round=MODEL_PARAMS.get(\"n_estimators\", 100),\n", " valid_sets=[train_dataset, val_dataset],\n", " valid_names=[\"train\", \"val\"],\n", " callbacks=[\n", " lgb.record_evaluation(evals_result),\n", " lgb.early_stopping(stopping_rounds=100, verbose=True),\n", " ],\n", ")\n", "\n", "print(\"训练完成，指标已收集\")\n", "\n", "# 获取指标名称\n", "metric_name = list(evals_result[\"train\"].keys())[0]\n", "print(f\"\\n评估指标: {metric_name}\")\n", "\n", "# 提取训练和验证指标\n", "train_metric = evals_result[\"train\"][metric_name]\n", "val_metric = evals_result[\"val\"][metric_name]\n", "\n", "# 显示早停信息\n", "actual_rounds = len(train_metric)\n", "expected_rounds = MODEL_PARAMS.get(\"n_estimators\", 100)\n", "print(f\"\\n[早停信息]\")\n", "print(f\" 配置的最大轮数: {expected_rounds}\")\n", "print(f\" 实际训练轮数: {actual_rounds}\")\n", "if actual_rounds < expected_rounds:\n", " print(f\" 早停状态: 已触发（连续100轮验证指标未改善）\")\n", "else:\n", " print(f\" 早停状态: 未触发（达到最大轮数）\")\n", "\n", "print(f\"\\n最终指标:\")\n", "print(f\" 训练 {metric_name}: {train_metric[-1]:.6f}\")\n", "print(f\" 验证 {metric_name}: {val_metric[-1]:.6f}\")" ] }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "# 绘制训练指标曲线\n", "import matplotlib.pyplot as plt\n", "\n", "fig, ax = plt.subplots(figsize=(12, 6))\n", "\n", "# 绘制训练集和验证集的指标曲线（注意：val用于验证，test不参与训练）\n", "iterations = range(1, len(train_metric) + 1)\n", "ax.plot(\n", " iterations, train_metric, label=f\"Train {metric_name}\", linewidth=2, color=\"blue\"\n", ")\n", "ax.plot(\n", " iterations, val_metric, label=f\"Validation {metric_name}\", linewidth=2, color=\"red\"\n", ")\n", "\n", "ax.set_xlabel(\"Iteration\", fontsize=12)\n", "ax.set_ylabel(metric_name.upper(), fontsize=12)\n", "ax.set_title(\n", " f\"Training and Validation {metric_name.upper()} Curve\",\n", " fontsize=14,\n", " fontweight=\"bold\",\n", ")\n", "ax.legend(fontsize=10)\n", "ax.grid(True, alpha=0.3)\n", "\n", "# 标记最佳验证指标点（用于早停决策）\n", "best_iter = val_metric.index(min(val_metric))\n", "best_metric = min(val_metric)\n", "ax.axvline(\n", " x=best_iter + 1,\n", " color=\"green\",\n", " linestyle=\"--\",\n", " alpha=0.7,\n", " label=f\"Best Iteration ({best_iter + 1})\",\n", ")\n", "ax.scatter([best_iter + 1], [best_metric], color=\"green\", s=100, zorder=5)\n", "ax.annotate(\n", " f\"Best: {best_metric:.6f}\\nIter: {best_iter + 1}\",\n", " xy=(best_iter + 1, best_metric),\n", " xytext=(best_iter + 1 + len(iterations) * 0.1, best_metric),\n", " fontsize=9,\n", " arrowprops=dict(arrowstyle=\"->\", color=\"green\", alpha=0.7),\n", ")\n", "\n", "plt.tight_layout()\n", "plt.show()\n", "\n", "print(f\"\\n[指标分析]\")\n", "print(f\" 最佳验证 {metric_name}: {best_metric:.6f}\")\n", "print(f\" 最佳迭代轮数: {best_iter + 1}\")\n", "print(f\" 早停建议: 如果验证指标连续10轮不下降，建议在第 {best_iter + 1} 轮停止训练\")\n", "print(f\"\\n[重要提醒] 验证集仅用于早停/调参，测试集完全独立于训练过程！\")" ] }, { "metadata": {}, "cell_type": "markdown", "source": "### 4.4 查看结果" }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "print(\"\\n\" + \"=\" * 80)\n", "print(\"训练结果\")\n", "print(\"=\" * 80)\n", "\n", "results = trainer.results\n", "\n", "print(f\"\\n结果数据形状: {results.shape}\")\n", "print(f\"结果列: {results.columns}\")\n", "print(f\"\\n结果前10行预览:\")\n", "print(results.head(10))\n", "print(f\"\\n结果后5行预览:\")\n", "print(results.tail())\n", "\n", "print(f\"\\n每日预测样本数统计:\")\n", "daily_counts = results.group_by(\"trade_date\").agg(pl.len()).sort(\"trade_date\")\n", "print(f\" 最小: {daily_counts['len'].min()}\")\n", "print(f\" 最大: {daily_counts['len'].max()}\")\n", "print(f\" 平均: {daily_counts['len'].mean():.2f}\")\n", "\n", "# 展示某一天的前10个预测结果\n", "sample_date = results[\"trade_date\"][0]\n", "sample_data = results.filter(results[\"trade_date\"] == sample_date).head(10)\n", "print(f\"\\n示例日期 {sample_date} 的前10条预测:\")\n", "print(sample_data.select([\"ts_code\", \"trade_date\", target_col, \"prediction\"]))" ] }, { "metadata": {}, "cell_type": "markdown", "source": "### 4.4 保存结果" }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "print(\"\\n\" + \"=\" * 80)\n", "print(\"保存预测结果\")\n", "print(\"=\" * 80)\n", "\n", "# 确保输出目录存在\n", "os.makedirs(OUTPUT_DIR, exist_ok=True)\n", "\n", "# 生成时间戳\n", "start_dt = datetime.strptime(TEST_START, \"%Y%m%d\")\n", "end_dt = datetime.strptime(TEST_END, \"%Y%m%d\")\n", "date_str = f\"{start_dt.strftime('%Y%m%d')}_{end_dt.strftime('%Y%m%d')}\"\n", "\n", "# 保存每日 Top N\n", "print(f\"\\n[1/1] 保存每日 Top {TOP_N} 股票...\")\n", "topn_output_path = os.path.join(OUTPUT_DIR, f\"regression_output.csv\")\n", "\n", "# 按日期分组，取每日 top N\n", "topn_by_date = []\n", "unique_dates = results[\"trade_date\"].unique().sort()\n", "for date in unique_dates:\n", " day_data = results.filter(results[\"trade_date\"] == date)\n", " # 按 prediction 降序排序，取前 N\n", " topn = day_data.sort(\"prediction\", descending=True).head(TOP_N)\n", " topn_by_date.append(topn)\n", "\n", "# 合并所有日期的 top N\n", "topn_results = pl.concat(topn_by_date)\n", "\n", "# 格式化日期并调整列顺序：日期、分数、股票\n", "topn_to_save = topn_results.select(\n", " [\n", " pl.col(\"trade_date\").str.slice(0, 4)\n", " + \"-\"\n", " + pl.col(\"trade_date\").str.slice(4, 2)\n", " + \"-\"\n", " + pl.col(\"trade_date\").str.slice(6, 2).alias(\"date\"),\n", " pl.col(\"prediction\").alias(\"score\"),\n", " pl.col(\"ts_code\"),\n", " ]\n", ")\n", "topn_to_save.write_csv(topn_output_path, include_header=True)\n", "print(f\" 保存路径: {topn_output_path}\")\n", "print(\n", " f\" 保存行数: {len(topn_to_save)}（{len(unique_dates)}个交易日 × 每日top{TOP_N}）\"\n", ")\n", "print(f\"\\n 预览（前15行）:\")\n", "print(topn_to_save.head(15))" ] }, { "metadata": {}, "cell_type": "markdown", "source": "### 4.5 特征重要性" }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "importance = model.feature_importance()\n", "if importance is not None:\n", " print(\"\\n特征重要性:\")\n", " print(importance.sort_values(ascending=False))\n", "\n", "print(\"\\n\" + \"=\" * 80)\n", "print(\"训练完成！\")\n", "print(\"=\" * 80)" ] }, { "metadata": {}, "cell_type": "markdown", "source": [ "## 5. 可视化分析\n", "#\n", "使用训练好的模型直接绘图。\n", "- **特征重要性图**：辅助特征选择\n", "- **决策树图**：理解决策逻辑" ] }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "# 导入可视化库\n", "import matplotlib.pyplot as plt\n", "import lightgbm as lgb\n", "import pandas as pd\n", "\n", "# 从封装的model中取出底层Booster\n", "booster = model.model\n", "print(f\"模型类型: {type(booster)}\")\n", "print(f\"特征数量: {len(feature_cols)}\")" ] }, { "metadata": {}, "cell_type": "markdown", "source": [ "### 5.1 绘制特征重要性（辅助特征选择）\n", "#\n", "**解读**：\n", "- 重要性高的特征对模型贡献大\n", "- 重要性为0的特征可以考虑删除\n", "- 可以帮助理解哪些因子最有效" ] }, { "metadata": {}, "cell_type": "code", "outputs": [], "execution_count": null, "source": [ "print(\"绘制特征重要性...\")\n", "\n", "fig, ax = plt.subplots(figsize=(10, 8))\n", "lgb.plot_importance(\n", " booster,\n", " max_num_features=20,\n", " importance_type=\"gain\",\n", " title=\"Feature Importance (Gain)\",\n", " ax=ax,\n", ")\n", "ax.set_xlabel(\"Importance (Gain)\")\n", "plt.tight_layout()\n", "plt.show()\n", "\n", "# 打印重要性排名\n", "importance_gain = pd.Series(\n", " booster.feature_importance(importance_type=\"gain\"), index=feature_cols\n", ").sort_values(ascending=False)\n", "\n", "print(\"\\n[特征重要性排名 - Gain]\")\n", "print(importance_gain)\n", "\n", "# 识别低重要性特征\n", "zero_importance = importance_gain[importance_gain == 0].index.tolist()\n", "if zero_importance:\n", " print(f\"\\n[低重要性特征] 以下{len(zero_importance)}个特征重要性为0，可考虑删除:\")\n", " for feat in zero_importance:\n", " print(f\" - {feat}\")\n", "else:\n", " print(\"\\n所有特征都有一定重要性\")\n" ] } ], "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": "3.10.0" } }, "nbformat": 4, "nbformat_minor": 4 }