这个目前是最优解+1

config_examples/live.json

@@ -19,9 +19,6 @@
       "timeout": 20000
     },
     "pair_whitelist": [
-      "AAVE/USDT",
-      "ADA/USDT",
-      "ASTER/USDT",
       "AVAX/USDT",
       "BCH/USDT",
       "BNB/USDT",
@@ -29,25 +26,14 @@
       "DOGE/USDT",
       "ETH/USDT",
       "FIL/USDT",
-      "HYPE/USDT",
       "IP/USDT",
-      "KITE/USDT",
-      "LINK/USDT",
       "LTC/USDT",
-      "PENGU/USDT",
       "PEPE/USDT",
-      "PI/USDT",
-      "PUMP/USDT",
-      "SNX/USDT",
       "SOL/USDT",
       "SUI/USDT",
-      "TON/USDT",
-      "TRX/USDT",
-      "UNI/USDT",
-      "USDG/USDT",
-      "WLFI/USDT",
+      "TRB/USDT",
+      "TRUMP/USDT",
       "XAUT/USDT",
-      "XPL/USDT",
       "XRP/USDT"
     ],
     "pair_blacklist": [

config_examples/live_XGBoostRegressorMultiTarget.json

@@ -1,5 +1,5 @@
 {
-  "dryrun": false;
+  "dryrun": false,
   "exchange": {
     "name": "okx",
     "key": "cbda9fde-b9e3-4a2d-94f9-e5c3705dfb5c",
@@ -83,11 +83,11 @@
     "live_retrain_hours": 2,
     "outlier_detection": {
       "method": "IsolationForest",
-      "contamination": 0.15  // 适度提高异常值检测阈值，过滤更多噪声
+      "contamination": 0.15
     },
     "feature_selection": {
       "method": "recursive_elimination",
-      "feature_number": 20  // 限制特征数量，避免维度爆炸导致过拟合
+      "feature_number": 20
     },
     "feature_parameters": {
       "include_timeframes": [
@@ -102,32 +102,32 @@
       ],
       "outlier_protection_percentage": 0.1,
       "label_period_candles": 100,
-      "include_shifted_candles": 2,  // 减少移位蜡烛数，降低特征冗余
+      "include_shifted_candles": 2,
       "DI_threshold": 0.5,
       "weight_factor": 0.9,
-      "principal_component_analysis": true,  // 开启PCA降维，减少特征维度
+      "principal_component_analysis": true,
       "use_SVM_to_remove_outliers": true,
       "indicator_periods_candles": [
         10,
         20,
-        50  // 减少指标周期数，避免特征过多
+        50
       ],
       "plot_feature_importances": 10
     },
     "data_split_parameters": {
-      "test_size": 0.2,  // 增大测试集比例，更严格验证模型泛化能力
+      "test_size": 0.2,
       "shuffle": false
     },
     "model_training_parameters": {
-      "n_estimators": 200,  // 减少树的数量，降低模型复杂度
-      "learning_rate": 0.05,  // 适度提高学习率，加快收敛且避免过拟合
-      "max_depth": 5,  // 降低树深度，核心缓解过拟合
-      "min_child_weight": 3,  // 提高子节点最小权重，避免拟合噪声
-      "subsample": 0.7,  // 降低样本采样率，增强正则化
-      "colsample_bytree": 0.7,  // 降低特征采样率，避免过拟合
+      "n_estimators": 200,
+      "learning_rate": 0.05,
+      "max_depth": 5,
+      "min_child_weight": 3,
+      "subsample": 0.7,
+      "colsample_bytree": 0.7,
       "verbosity": 0,
-      "reg_alpha": 0.1,  // 新增L1正则化，限制权重
-      "reg_lambda": 1.0  // 新增L2正则化，核心缓解过拟合
+      "reg_alpha": 0.1,
+      "reg_lambda": 1.0
     }
   },
   "fee": 0.0008,