{"id":64461,"date":"2026-01-23T15:56:22","date_gmt":"2026-01-23T07:56:22","guid":{"rendered":"https:\/\/www.wsisp.com\/helps\/64461.html"},"modified":"2026-01-23T15:56:22","modified_gmt":"2026-01-23T07:56:22","slug":"catboost%ef%bc%9a%e6%9c%80%e5%bc%ba%e6%97%b6%e9%97%b4%e5%ba%8f%e5%88%97%e9%a2%84%e6%b5%8b%e7%ae%97%e6%b3%95%ef%bc%88%e9%99%84%e6%95%b0%e6%8d%ae%e6%ba%90%e7%a0%81%ef%bc%89","status":"publish","type":"post","link":"https:\/\/www.wsisp.com\/helps\/64461.html","title":{"rendered":"CatBoost\uff1a\u6700\u5f3a\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u7b97\u6cd5\uff08\u9644\u6570\u636e\u6e90\u7801\uff09"},"content":{"rendered":"

\u4e00\u3001CatBoost\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u539f\u7406：<\/h2>\n
  • \n

    \u7279\u5f81\u5de5\u7a0b\u8f6c\u6362<\/h3>\n
      \n
    • \n

      \u5c06\u65f6\u95f4\u5e8f\u5217\u95ee\u9898\u8f6c\u5316\u4e3a\u76d1\u7763\u5b66\u4e60\u95ee\u9898<\/p>\n<\/li>\n

    • \n

      \u4f7f\u7528\u6ed1\u52a8\u7a97\u53e3\u521b\u5efa\u6ede\u540e\u7279\u5f81<\/p>\n<\/li>\n

    • \n

      \u6dfb\u52a0\u7edf\u8ba1\u7279\u5f81（\u5747\u503c\u3001\u6807\u51c6\u5dee\u3001\u6700\u5927\u503c\u3001\u6700\u5c0f\u503c）<\/p>\n<\/li>\n<\/ul>\n<\/li>\n

    • \n

      \u5904\u7406\u65f6\u5e8f\u4f9d\u8d56<\/h3>\n
        \n
      • \n

        \u901a\u8fc7\u6ede\u540e\u7279\u5f81\u6355\u6349\u65f6\u95f4\u4f9d\u8d56\u5173\u7cfb<\/p>\n<\/li>\n

      • \n

        \u6a21\u578b\u80fd\u591f\u5b66\u4e60\u5386\u53f2\u6a21\u5f0f\u5bf9\u672a\u6765\u503c\u7684\u5f71\u54cd<\/p>\n<\/li>\n<\/ul>\n<\/li>\n

      • \n

        \u6709\u5e8f\u63d0\u5347\u673a\u5236<\/h3>\n
          \n
        • \n

          CatBoost\u4f7f\u7528\u6709\u5e8f\u63d0\u5347（Ordered Boosting）\u5904\u7406\u65f6\u5e8f\u6570\u636e<\/p>\n<\/li>\n

        • \n

          \u907f\u514d\u76ee\u6807\u6cc4\u6f0f，\u786e\u4fdd\u65f6\u95f4\u4e0a\u7684\u56e0\u679c\u5173\u7cfb<\/p>\n<\/li>\n<\/ul>\n<\/li>\n

        • \n

          \u7c7b\u522b\u7279\u5f81\u5904\u7406<\/h3>\n
            \n
          • \n

            \u5185\u7f6e\u7c7b\u522b\u7279\u5f81\u5904\u7406\u80fd\u529b，\u9002\u5408\u5305\u542b\u5206\u7c7b\u53d8\u91cf\u7684\u65f6\u5e8f\u6570\u636e<\/p>\n<\/li>\n

          • \n

            \u4f7f\u7528\u76ee\u6807\u7edf\u8ba1\u548c\u6392\u5e8f\u63d0\u5347\u51cf\u5c11\u8fc7\u62df\u5408<\/p>\n<\/li>\n<\/ul>\n<\/li>\n

          • \n

            \u68af\u5ea6\u63d0\u5347\u6846\u67b6<\/h3>\n
              \n
            • \n

              \u57fa\u4e8e\u51b3\u7b56\u6811\u7684\u68af\u5ea6\u63d0\u5347\u7b97\u6cd5<\/p>\n<\/li>\n

            • \n

              \u9010\u6b65\u4f18\u5316\u635f\u5931\u51fd\u6570，\u51cf\u5c11\u9884\u6d4b\u8bef\u5dee<\/p>\n<\/li>\n<\/ul>\n<\/li>\n

              \u4e8c\u3001\u4e0e\u5176\u4ed6\u7b97\u6cd5\u7684\u4f18\u52bf：<\/h2>\n
            • \n

              vs \u4f20\u7edf\u7edf\u8ba1\u65b9\u6cd5（ARIMA, SARIMA）<\/h3>\n
                \n
              • \n

                \u65e0\u9700\u624b\u52a8\u8bc6\u522b\u53c2\u6570（p,d,q）<\/p>\n<\/li>\n

              • \n

                \u80fd\u81ea\u52a8\u6355\u6349\u975e\u7ebf\u6027\u5173\u7cfb<\/p>\n<\/li>\n

              • \n

                \u5904\u7406\u7f3a\u5931\u503c\u548c\u5f02\u5e38\u503c\u66f4\u9c81\u68d2<\/p>\n<\/li>\n<\/ul>\n<\/li>\n

              • \n

                vs LSTM\/RNN<\/h3>\n
                  \n
                • \n

                  \u8bad\u7ec3\u901f\u5ea6\u66f4\u5feb，\u8ba1\u7b97\u8d44\u6e90\u9700\u6c42\u66f4\u4f4e<\/p>\n<\/li>\n

                • \n

                  \u8d85\u53c2\u6570\u8c03\u6574\u66f4\u7b80\u5355<\/p>\n<\/li>\n

                • \n

                  \u5bf9\u5c0f\u6570\u636e\u96c6\u8868\u73b0\u66f4\u597d，\u4e0d\u6613\u8fc7\u62df\u5408<\/p>\n<\/li>\n<\/ul>\n<\/li>\n

                • \n

                  vs XGBoost\/LightGBM<\/h3>\n
                    \n
                  • \n

                    \u5185\u7f6e\u7c7b\u522b\u7279\u5f81\u5904\u7406，\u65e0\u9700\u72ec\u70ed\u7f16\u7801<\/p>\n<\/li>\n

                  • \n

                    \u6709\u5e8f\u63d0\u5347\u51cf\u5c11\u8fc7\u62df\u5408，\u63d0\u5347\u6cdb\u5316\u80fd\u529b<\/p>\n<\/li>\n

                  • \n

                    \u81ea\u52a8\u5904\u7406\u7f3a\u5931\u503c，\u9884\u5904\u7406\u66f4\u7b80\u5355<\/p>\n<\/li>\n

                  • \n

                    \u5982\u9700\u8981\u6570\u636e\u53ca\u6e90\u7801，\u5173\u6ce8\u5fae\u4fe1\u516c\u4f17\u53f7\u3010\u7231\u6572\u4ee3\u7801\u7684\u7834\u6653\u3011，\u540e\u53f0\u79c1\u4fe1\u3010Catboost0123\u3011<\/p>\n<\/li>\n<\/ul>\n<\/li>\n

                  • \n

                    vs Prophet<\/h3>\n
                      \n
                    • \n

                      \u66f4\u7075\u6d3b\u7684\u7279\u5f81\u5de5\u7a0b\u80fd\u529b<\/p>\n<\/li>\n

                    • \n

                      \u80fd\u7ed3\u5408\u5916\u90e8\u7279\u5f81\u4e00\u8d77\u5efa\u6a21<\/p>\n<\/li>\n

                    • \n

                      \u5bf9\u7a81\u53d8\u70b9（breakpoints）\u7684\u9002\u5e94\u6027\u66f4\u5f3a<\/p>\n<\/li>\n<\/ul>\n<\/li>\n

                    • \n

                      \u7efc\u5408\u4f18\u52bf<\/h3>\n
                        \n
                      • \n

                        \u8bad\u7ec3\u901f\u5ea6\u5feb，\u652f\u6301GPU\u52a0\u901f<\/p>\n<\/li>\n

                      • \n

                        \u5185\u7f6e\u6b63\u5219\u5316，\u9632\u6b62\u8fc7\u62df\u5408<\/p>\n<\/li>\n

                      • \n

                        \u63d0\u4f9b\u7279\u5f81\u91cd\u8981\u6027\u5206\u6790<\/p>\n<\/li>\n

                      • \n

                        \u5bf9\u5f02\u5e38\u503c\u4e0d\u654f\u611f，\u7a33\u5b9a\u6027\u597d<\/p>\n<\/li>\n

                      • \n

                        \u5982\u9700\u8981\u6570\u636e\u53ca\u6e90\u7801，\u5173\u6ce8\u5fae\u4fe1\u516c\u4f17\u53f7\u3010\u7231\u6572\u4ee3\u7801\u7684\u7834\u6653\u3011，\u540e\u53f0\u79c1\u4fe1\u3010Catboost0123\u3011<\/p>\n<\/li>\n<\/ul>\n<\/li>\n

                        \u4e09\u3001\u5e94\u7528\u793a\u4f8b<\/h2>\n

                        1.\u6570\u636e\u51c6\u5907<\/h3>\n<\/p>\n

                        \"\u56fe\u7247\"<\/p>\n

                        2.\u5bfc\u5165\u7b2c\u4e09\u65b9\u5e93<\/h3>\n

                        import\u00a0pandas\u00a0as\u00a0pd
                        \nimport\u00a0numpy\u00a0as\u00a0np
                        \nimport\u00a0joblib
                        \nimport\u00a0matplotlib.pyplot\u00a0as\u00a0plt
                        \nfrom\u00a0sklearn.metrics\u00a0import\u00a0mean_absolute_error, mean_squared_error, r2_score
                        \nimport\u00a0warnings
                        \nwarnings.filterwarnings('ignore') <\/p>\n

                        \u5982\u9700\u8981\u6570\u636e\u53ca\u6e90\u7801，\u5173\u6ce8\u5fae\u4fe1\u516c\u4f17\u53f7\u3010\u7231\u6572\u4ee3\u7801\u7684\u7834\u6653\u3011，\u540e\u53f0\u79c1\u4fe1\u3010Catboost0123\u3011<\/p>\n

                        3.\u6a21\u578b\u914d\u7f6e \u00a0<\/h3>\n

                        # \u6570\u636e\u914d\u7f6e
                        \nDATA_PATH =\u00a0'data.csv'
                        \nTARGET_COLUMN =\u00a0'Open'
                        \nTEST_SIZE = 0.2
                        \nRANDOM_STATE = 42
                        \n# \u7279\u5f81\u5de5\u7a0b\u914d\u7f6e
                        \nLOOKBACK_WINDOW = 30 \u00a0# \u4f7f\u7528\u8fc7\u53bb30\u5929\u7684\u6570\u636e\u9884\u6d4b\u4e0b\u4e00\u5929
                        \nFORECAST_HORIZON = 1 \u00a0# \u9884\u6d4b\u672a\u67651\u5929
                        \n# CatBoost\u6a21\u578b\u914d\u7f6e
                        \nCATBOOST_PARAMS = {
                        \n\u00a0 \u00a0\u00a0'iterations': 1000,
                        \n\u00a0 \u00a0\u00a0'learning_rate': 0.03,
                        \n\u00a0 \u00a0\u00a0'depth': 6,
                        \n\u00a0 \u00a0\u00a0'loss_function':\u00a0'RMSE',
                        \n\u00a0 \u00a0\u00a0'verbose': 100,
                        \n\u00a0 \u00a0\u00a0'random_seed': RANDOM_STATE,
                        \n\u00a0 \u00a0\u00a0'task_type':\u00a0'CPU', \u00a0# \u53ef\u6539\u4e3a'GPU'\u5982\u679c\u6709GPU
                        \n\u00a0 \u00a0\u00a0'early_stopping_rounds': 50
                        \n}
                        \n# \u7ed8\u56fe\u914d\u7f6e
                        \nPLOT_CONFIG = {
                        \n\u00a0 \u00a0\u00a0'figsize': (12, 8),
                        \n\u00a0 \u00a0\u00a0'train_color':\u00a0'blue',
                        \n\u00a0 \u00a0\u00a0'test_color':\u00a0'red',
                        \n\u00a0 \u00a0\u00a0'pred_color':\u00a0'green',
                        \n\u00a0 \u00a0\u00a0'dpi': 100
                        \n} <\/p>\n

                        4.\u52a0\u8f7d\u6570\u636e<\/h3>\n

                        def\u00a0load_and_prepare_data():
                        \n\u00a0 \u00a0\u00a0"""\u52a0\u8f7d\u5e76\u51c6\u5907\u6570\u636e"""
                        \n\u00a0 \u00a0\u00a0print("\u52a0\u8f7d\u6570\u636e…")
                        \n\u00a0 \u00a0 df = pd.read_csv(DATA_PATH, parse_dates=['Date'])
                        \n\u00a0 \u00a0 df = df.sort_values('Date')
                        \n\u00a0 \u00a0 df.set_index('Date', inplace=True)
                        \n\u00a0 \u00a0\u00a0# \u786e\u4fdd\u6ca1\u6709\u7f3a\u5931\u503c
                        \n\u00a0 \u00a0 df = df.fillna(method='ffill') <\/p>\n

                        5.\u7279\u5f81\u5de5\u7a0b<\/h3>\n

                        def\u00a0create_features(data, target_col, window_size):
                        \n\u00a0 \u00a0\u00a0"""\u521b\u5efa\u65f6\u95f4\u5e8f\u5217\u7279\u5f81"""
                        \n\u00a0 \u00a0 features = []
                        \n\u00a0 \u00a0 targets = []
                        \n\u00a0 \u00a0\u00a0for\u00a0i\u00a0in\u00a0range(window_size,\u00a0len(data)):
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0# \u4f7f\u7528\u8fc7\u53bbwindow_size\u5929\u7684\u6570\u636e
                        \n\u00a0 \u00a0 \u00a0 \u00a0 feature_window = data[target_col].iloc[i-window_size:i].values
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0# \u6dfb\u52a0\u7edf\u8ba1\u7279\u5f81
                        \n\u00a0 \u00a0 \u00a0 \u00a0 feature_mean = np.mean(feature_window)
                        \n\u00a0 \u00a0 \u00a0 \u00a0 feature_std = np.std(feature_window)
                        \n\u00a0 \u00a0 \u00a0 \u00a0 feature_max = np.max(feature_window)
                        \n\u00a0 \u00a0 \u00a0 \u00a0 feature_min = np.min(feature_window)
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0# \u6dfb\u52a0\u6ede\u540e\u7279\u5f81
                        \n\u00a0 \u00a0 \u00a0 \u00a0 lag_features =\u00a0list(feature_window)
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0# \u7ec4\u5408\u6240\u6709\u7279\u5f81
                        \n\u00a0 \u00a0 \u00a0 \u00a0 all_features = lag_features + [feature_mean, feature_std, feature_max, feature_min]
                        \n\u00a0 \u00a0 \u00a0 \u00a0 features.append(all_features)
                        \n\u00a0 \u00a0 \u00a0 \u00a0 targets.append(data[target_col].iloc[i])
                        \n\u00a0 \u00a0\u00a0return\u00a0np.array(features), np.array(targets) <\/p>\n

                        \u5982\u9700\u8981\u6570\u636e\u53ca\u6e90\u7801，\u5173\u6ce8\u5fae\u4fe1\u516c\u4f17\u53f7\u3010\u7231\u6572\u4ee3\u7801\u7684\u7834\u6653\u3011，\u540e\u53f0\u79c1\u4fe1\u3010Catboost0123\u3011<\/p>\n

                        6.\u8bad\u7ec3\u6a21\u578b<\/h3>\n

                        def\u00a0train_model():
                        \n\u00a0 \u00a0\u00a0"""\u8bad\u7ec3\u6a21\u578b"""
                        \n\u00a0 \u00a0\u00a0print("\u5f00\u59cb\u8bad\u7ec3CatBoost\u6a21\u578b…")
                        \n\u00a0 \u00a0\u00a0# \u52a0\u8f7d\u6570\u636e
                        \n\u00a0 \u00a0 df = load_and_prepare_data()
                        \n\u00a0 \u00a0\u00a0# \u521b\u5efa\u7279\u5f81
                        \n\u00a0 \u00a0 X, y = create_features(df, TARGET_COLUMN, LOOKBACK_WINDOW)
                        \n\u00a0 \u00a0\u00a0# \u5212\u5206\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6
                        \n\u00a0 \u00a0 train_size =\u00a0int(len(X) * (1\u00a0– TEST_SIZE))
                        \n\u00a0 \u00a0 X_train, X_test = X[:train_size], X[train_size:]
                        \n\u00a0 \u00a0 y_train, y_test = y[:train_size], y[train_size:]
                        \n\u00a0 \u00a0\u00a0print(f"\u8bad\u7ec3\u96c6\u5927\u5c0f:\u00a0{X_train.shape[0]}")
                        \n\u00a0 \u00a0\u00a0print(f"\u6d4b\u8bd5\u96c6\u5927\u5c0f:\u00a0{X_test.shape[0]}")
                        \n\u00a0 \u00a0\u00a0# \u521b\u5efa\u5e76\u8bad\u7ec3\u6a21\u578b
                        \n\u00a0 \u00a0 model = CatBoostRegressor(**CATBOOST_PARAMS)
                        \n\u00a0 \u00a0 model.fit(
                        \n\u00a0 \u00a0 \u00a0 \u00a0 X_train, y_train,
                        \n\u00a0 \u00a0 \u00a0 \u00a0 eval_set=(X_test, y_test),
                        \n\u00a0 \u00a0 \u00a0 \u00a0 use_best_model=True,
                        \n\u00a0 \u00a0 \u00a0 \u00a0 plot=False
                        \n\u00a0 \u00a0 )
                        \n\u00a0 \u00a0\u00a0# \u4fdd\u5b58\u6a21\u578b
                        \n\u00a0 \u00a0 joblib.dump(model,\u00a0'catboost_timeseries_model.pkl')
                        \n\u00a0 \u00a0\u00a0print("\u6a21\u578b\u5df2\u4fdd\u5b58\u4e3a 'catboost_timeseries_model.pkl'")
                        \n\u00a0 \u00a0\u00a0# \u9884\u6d4b
                        \n\u00a0 \u00a0 y_pred_train = model.predict(X_train)
                        \n\u00a0 \u00a0 y_pred_test = model.predict(X_test)
                        \n\u00a0 \u00a0\u00a0return\u00a0{
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'model': model,
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'X_train': X_train,
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'X_test': X_test,
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'y_train': y_train,
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'y_test': y_test,
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'y_pred_train': y_pred_train,
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'y_pred_test': y_pred_test,
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'dates': df.index[LOOKBACK_WINDOW:]
                        \n\u00a0 \u00a0 } <\/p>\n

                        7.\u6a21\u578b\u9884\u6d4b<\/h3>\n

                        def plot_predictions(results):
                        \n\u00a0 \u00a0\u00a0"""\u7ed8\u5236\u9884\u6d4b\u66f2\u7ebf"""
                        \n\u00a0 \u00a0 dates = results['dates']
                        \n\u00a0 \u00a0 train_size =\u00a0len(results['y_train'])
                        \n\u00a0 \u00a0 plt.figure(figsize=PLOT_CONFIG['figsize'])
                        \n\u00a0 \u00a0 # \u8bad\u7ec3\u96c6\u90e8\u5206
                        \n\u00a0 \u00a0 train_dates = dates[:train_size]
                        \n\u00a0 \u00a0 plt.plot(train_dates, results['y_train'],\u00a0
                        \n\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0label='\u8bad\u7ec3\u96c6\u771f\u5b9e\u503c', color=PLOT_CONFIG['train_color'], alpha=0.7)
                        \n\u00a0 \u00a0 plt.plot(train_dates, results['y_pred_train'],\u00a0
                        \n\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0label='\u8bad\u7ec3\u96c6\u9884\u6d4b\u503c', color=PLOT_CONFIG['pred_color'], alpha=0.7, linestyle='–')
                        \n\u00a0 \u00a0 # \u6d4b\u8bd5\u96c6\u90e8\u5206
                        \n\u00a0 \u00a0 test_dates = dates[train_size:]
                        \n\u00a0 \u00a0 plt.plot(test_dates, results['y_test'],\u00a0
                        \n\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0label='\u6d4b\u8bd5\u96c6\u771f\u5b9e\u503c', color=PLOT_CONFIG['test_color'], alpha=0.7)
                        \n\u00a0 \u00a0 plt.plot(test_dates, results['y_pred_test'],\u00a0
                        \n\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0label='\u6d4b\u8bd5\u96c6\u9884\u6d4b\u503c', color='orange', alpha=0.7, linestyle='–')
                        \n\u00a0 \u00a0 plt.title('CatBoost\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b – \u8bad\u7ec3\u96c6 vs \u6d4b\u8bd5\u96c6')
                        \n\u00a0 \u00a0 plt.xlabel('\u65e5\u671f')
                        \n\u00a0 \u00a0 plt.ylabel('\u5f00\u76d8\u4ef7')
                        \n\u00a0 \u00a0 plt.legend()
                        \n\u00a0 \u00a0 plt.grid(True, alpha=0.3)
                        \n\u00a0 \u00a0 plt.xticks(rotation=45)
                        \n\u00a0 \u00a0 plt.tight_layout()
                        \n\u00a0 \u00a0 plt.savefig('prediction_plot.png', dpi=PLOT_CONFIG['dpi'])
                        \n\u00a0 \u00a0 plt.show() <\/p>\n

                        8.\u6a21\u578b\u8bc4\u4ef7\u5e76\u7ed8\u56fe<\/h3>\n

                        def\u00a0evaluate_predictions(y_true, y_pred, set_name="\u6570\u636e\u96c6"):
                        \n\u00a0 \u00a0\u00a0"""\u8bc4\u4f30\u9884\u6d4b\u7ed3\u679c"""
                        \n\u00a0 \u00a0 mae = mean_absolute_error(y_true, y_pred)
                        \n\u00a0 \u00a0 mse = mean_squared_error(y_true, y_pred)
                        \n\u00a0 \u00a0 rmse = np.sqrt(mse)
                        \n\u00a0 \u00a0 r2 = r2_score(y_true, y_pred)
                        \n\u00a0 \u00a0\u00a0# \u8ba1\u7b97MAPE（\u907f\u514d\u9664\u96f6）
                        \n\u00a0 \u00a0 mape = np.mean(np.abs((y_true – y_pred) \/ np.maximum(np.abs(y_true),\u00a01e-10))) *\u00a0100
                        \n\u00a0 \u00a0\u00a0print(f"\\\\n{set_name}\u8bc4\u4f30\u7ed3\u679c:")
                        \n\u00a0 \u00a0\u00a0print(f"MAE:\u00a0{mae:.4f}")
                        \n\u00a0 \u00a0\u00a0print(f"MSE:\u00a0{mse:.4f}")
                        \n\u00a0 \u00a0\u00a0print(f"RMSE:\u00a0{rmse:.4f}")
                        \n\u00a0 \u00a0\u00a0print(f"R\u00b2:\u00a0{r2:.4f}")
                        \n\u00a0 \u00a0\u00a0print(f"MAPE:\u00a0{mape:.2f}%")
                        \n\u00a0 \u00a0\u00a0return\u00a0{
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'MAE': mae,
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'MSE': mse,
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'RMSE': rmse,
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'R2': r2,
                        \n\u00a0 \u00a0 \u00a0 \u00a0\u00a0'MAPE': mape
                        \n\u00a0 \u00a0 }
                        \ndef\u00a0plot_density_scatter(ax, y_true, y_pred, title):
                        \n\u00a0 \u00a0\u00a0"""\u7ed8\u5236\u6e10\u53d8\u5bc6\u5ea6\u6563\u70b9\u56fe"""
                        \n\u00a0 \u00a0\u00a0from\u00a0scipy.stats\u00a0import\u00a0gaussian_kde
                        \n\u00a0 \u00a0\u00a0# \u8ba1\u7b97\u70b9\u5bc6\u5ea6
                        \n\u00a0 \u00a0 xy = np.vstack([y_true, y_pred])
                        \n\u00a0 \u00a0 z = gaussian_kde(xy)(xy)
                        \n\u00a0 \u00a0\u00a0# \u6309\u5bc6\u5ea6\u6392\u5e8f，\u4f7f\u5f97\u5bc6\u5ea6\u9ad8\u7684\u70b9\u663e\u793a\u5728\u4e0a\u5c42
                        \n\u00a0 \u00a0 idx = z.argsort()
                        \n\u00a0 \u00a0 x, y, z = y_true[idx], y_pred[idx], z[idx]
                        \n\u00a0 \u00a0 scatter = ax.scatter(x, y, c=z, s=20, alpha=0.6, cmap='viridis')
                        \n\u00a0 \u00a0 plt.colorbar(scatter, ax = ax, label =\u00a0"\u70b9\u5bc6\u5ea6")
                        \n\u00a0 \u00a0\u00a0# \u6dfb\u52a0\u5bf9\u89d2\u7ebf（\u5b8c\u7f8e\u9884\u6d4b\u7ebf）
                        \n\u00a0 \u00a0 lims = [np.min([ax.get_xlim(), ax.get_ylim()]),\u00a0
                        \n\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 np.max([ax.get_xlim(), ax.get_ylim()])]
                        \n\u00a0 \u00a0 ax.plot(lims, lims,\u00a0'r–', alpha=0.8, label='\u5b8c\u7f8e\u9884\u6d4b\u7ebf')
                        \n\u00a0 \u00a0 ax.set_xlabel('\u771f\u5b9e\u503c')
                        \n\u00a0 \u00a0 ax.set_ylabel('\u9884\u6d4b\u503c')
                        \n\u00a0 \u00a0 ax.set_title(title)
                        \n\u00a0 \u00a0 ax.legend()
                        \n\u00a0 \u00a0 ax.grid(True, alpha=0.3)
                        \n\u00a0 \u00a0\u00a0return\u00a0scatter <\/p>\n

                        \u4ee5\u4e0b\u662f\u8f93\u51fa\u7684\u56fe\u4ef6：<\/p>\n<\/p>\n

                        \"\u56fe\u7247\"<\/p>\n<\/p>\n

                        \"\u56fe\u7247\"<\/p>\n<\/p>\n

                        \"\u56fe\u7247\"<\/p>\n<\/p>\n

                        \"\u56fe\u7247\"<\/p>\n","protected":false},"excerpt":{"rendered":"

                        \u4e00\u3001CatBoost\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u539f\u7406： \u7279\u5f81\u5de5\u7a0b\u8f6c\u6362 \u5c06\u65f6\u95f4\u5e8f\u5217\u95ee\u9898\u8f6c\u5316\u4e3a\u76d1\u7763\u5b66\u4e60\u95ee\u9898 \u4f7f\u7528\u6ed1\u52a8\u7a97\u53e3\u521b\u5efa\u6ede\u540e\u7279\u5f81 \u6dfb\u52a0\u7edf\u8ba1\u7279\u5f81（\u5747\u503c\u3001\u6807\u51c6\u5dee\u3001\u6700\u5927\u503c\u3001\u6700\u5c0f\u503c） \u5904\u7406\u65f6\u5e8f\u4f9d\u8d56 \u901a\u8fc7\u6ede\u540e\u7279\u5f81\u6355\u6349\u65f6\u95f4\u4f9d\u8d56\u5173\u7cfb \u6a21\u578b\u80fd\u591f\u5b66\u4e60\u5386\u53f2\u6a21\u5f0f\u5bf9\u672a\u6765\u503c\u7684\u5f71\u54cd \u6709\u5e8f\u63d0\u5347\u673a\u5236 CatBoost\u4f7f\u7528\u6709\u5e8f\u63d0\u5347（Ordered Boosting）\u5904\u7406\u65f6\u5e8f\u6570\u636e \u907f\u514d\u76ee\u6807\u6cc4\u6f0f，\u786e\u4fdd\u65f6\u95f4\u4e0a\u7684\u56e0\u679c\u5173\u7cfb \u7c7b\u522b\u7279\u5f81<\/p>\n","protected":false},"author":2,"featured_media":64456,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[50,86],"topic":[],"class_list":["post-64461","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-server","tag-50","tag-86"],"yoast_head":"\nCatBoost\uff1a\u6700\u5f3a\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u7b97\u6cd5\uff08\u9644\u6570\u636e\u6e90\u7801\uff09 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.wsisp.com\/helps\/64461.html\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"CatBoost\uff1a\u6700\u5f3a\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u7b97\u6cd5\uff08\u9644\u6570\u636e\u6e90\u7801\uff09 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"og:description\" content=\"\u4e00\u3001CatBoost\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u539f\u7406： \u7279\u5f81\u5de5\u7a0b\u8f6c\u6362 \u5c06\u65f6\u95f4\u5e8f\u5217\u95ee\u9898\u8f6c\u5316\u4e3a\u76d1\u7763\u5b66\u4e60\u95ee\u9898 \u4f7f\u7528\u6ed1\u52a8\u7a97\u53e3\u521b\u5efa\u6ede\u540e\u7279\u5f81 \u6dfb\u52a0\u7edf\u8ba1\u7279\u5f81（\u5747\u503c\u3001\u6807\u51c6\u5dee\u3001\u6700\u5927\u503c\u3001\u6700\u5c0f\u503c） \u5904\u7406\u65f6\u5e8f\u4f9d\u8d56 \u901a\u8fc7\u6ede\u540e\u7279\u5f81\u6355\u6349\u65f6\u95f4\u4f9d\u8d56\u5173\u7cfb \u6a21\u578b\u80fd\u591f\u5b66\u4e60\u5386\u53f2\u6a21\u5f0f\u5bf9\u672a\u6765\u503c\u7684\u5f71\u54cd \u6709\u5e8f\u63d0\u5347\u673a\u5236 CatBoost\u4f7f\u7528\u6709\u5e8f\u63d0\u5347（Ordered Boosting）\u5904\u7406\u65f6\u5e8f\u6570\u636e \u907f\u514d\u76ee\u6807\u6cc4\u6f0f，\u786e\u4fdd\u65f6\u95f4\u4e0a\u7684\u56e0\u679c\u5173\u7cfb \u7c7b\u522b\u7279\u5f81\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.wsisp.com\/helps\/64461.html\" \/>\n<meta property=\"og:site_name\" content=\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"article:published_time\" content=\"2026-01-23T07:56:22+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/01\/20260123075619-697329a3d3c98.png\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"4 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/64461.html\",\"url\":\"https:\/\/www.wsisp.com\/helps\/64461.html\",\"name\":\"CatBoost\uff1a\u6700\u5f3a\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u7b97\u6cd5\uff08\u9644\u6570\u636e\u6e90\u7801\uff09 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"isPartOf\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\"},\"datePublished\":\"2026-01-23T07:56:22+00:00\",\"dateModified\":\"2026-01-23T07:56:22+00:00\",\"author\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/64461.html#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.wsisp.com\/helps\/64461.html\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/64461.html#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/www.wsisp.com\/helps\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"CatBoost\uff1a\u6700\u5f3a\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u7b97\u6cd5\uff08\u9644\u6570\u636e\u6e90\u7801\uff09\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\",\"url\":\"https:\/\/www.wsisp.com\/helps\/\",\"name\":\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"description\":\"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"contentUrl\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"caption\":\"admin\"},\"sameAs\":[\"http:\/\/wp.wsisp.com\"],\"url\":\"https:\/\/www.wsisp.com\/helps\/author\/admin\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"CatBoost\uff1a\u6700\u5f3a\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u7b97\u6cd5\uff08\u9644\u6570\u636e\u6e90\u7801\uff09 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.wsisp.com\/helps\/64461.html","og_locale":"zh_CN","og_type":"article","og_title":"CatBoost\uff1a\u6700\u5f3a\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u7b97\u6cd5\uff08\u9644\u6570\u636e\u6e90\u7801\uff09 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","og_description":"\u4e00\u3001CatBoost\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u539f\u7406： \u7279\u5f81\u5de5\u7a0b\u8f6c\u6362 \u5c06\u65f6\u95f4\u5e8f\u5217\u95ee\u9898\u8f6c\u5316\u4e3a\u76d1\u7763\u5b66\u4e60\u95ee\u9898 \u4f7f\u7528\u6ed1\u52a8\u7a97\u53e3\u521b\u5efa\u6ede\u540e\u7279\u5f81 \u6dfb\u52a0\u7edf\u8ba1\u7279\u5f81（\u5747\u503c\u3001\u6807\u51c6\u5dee\u3001\u6700\u5927\u503c\u3001\u6700\u5c0f\u503c） \u5904\u7406\u65f6\u5e8f\u4f9d\u8d56 \u901a\u8fc7\u6ede\u540e\u7279\u5f81\u6355\u6349\u65f6\u95f4\u4f9d\u8d56\u5173\u7cfb \u6a21\u578b\u80fd\u591f\u5b66\u4e60\u5386\u53f2\u6a21\u5f0f\u5bf9\u672a\u6765\u503c\u7684\u5f71\u54cd \u6709\u5e8f\u63d0\u5347\u673a\u5236 CatBoost\u4f7f\u7528\u6709\u5e8f\u63d0\u5347（Ordered Boosting）\u5904\u7406\u65f6\u5e8f\u6570\u636e \u907f\u514d\u76ee\u6807\u6cc4\u6f0f，\u786e\u4fdd\u65f6\u95f4\u4e0a\u7684\u56e0\u679c\u5173\u7cfb \u7c7b\u522b\u7279\u5f81","og_url":"https:\/\/www.wsisp.com\/helps\/64461.html","og_site_name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","article_published_time":"2026-01-23T07:56:22+00:00","og_image":[{"url":"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/01\/20260123075619-697329a3d3c98.png"}],"author":"admin","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"4 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.wsisp.com\/helps\/64461.html","url":"https:\/\/www.wsisp.com\/helps\/64461.html","name":"CatBoost\uff1a\u6700\u5f3a\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u7b97\u6cd5\uff08\u9644\u6570\u636e\u6e90\u7801\uff09 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","isPartOf":{"@id":"https:\/\/www.wsisp.com\/helps\/#website"},"datePublished":"2026-01-23T07:56:22+00:00","dateModified":"2026-01-23T07:56:22+00:00","author":{"@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41"},"breadcrumb":{"@id":"https:\/\/www.wsisp.com\/helps\/64461.html#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.wsisp.com\/helps\/64461.html"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.wsisp.com\/helps\/64461.html#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/www.wsisp.com\/helps"},{"@type":"ListItem","position":2,"name":"CatBoost\uff1a\u6700\u5f3a\u65f6\u95f4\u5e8f\u5217\u9884\u6d4b\u7b97\u6cd5\uff08\u9644\u6570\u636e\u6e90\u7801\uff09"}]},{"@type":"WebSite","@id":"https:\/\/www.wsisp.com\/helps\/#website","url":"https:\/\/www.wsisp.com\/helps\/","name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","description":"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41","name":"admin","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/","url":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","contentUrl":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","caption":"admin"},"sameAs":["http:\/\/wp.wsisp.com"],"url":"https:\/\/www.wsisp.com\/helps\/author\/admin"}]}},"_links":{"self":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/64461","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/comments?post=64461"}],"version-history":[{"count":0,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/64461\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media\/64456"}],"wp:attachment":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media?parent=64461"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/categories?post=64461"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/tags?post=64461"},{"taxonomy":"topic","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/topic?post=64461"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}