ML之回歸預測：利用十(xgboost,10-1)種機器學習算法對無人駕駛汽車系統參數(2017年的data,18+2)進行回歸預測值VS真實值——bug調試記錄

?

?

?

目錄

輸出結果

1、增加XGBR算法

---

?

?

?

?

輸出結果

1、增加XGBR算法

1、增加XGBR算法時候，采用網格搜索的方法

XGBR_grid_model Training time： 135.60037931849538 輸出XGBR_grid_model模型的最優參數： {'learning_rate': 0.03, 'max_depth': 4, 'n_estimators': 100} XGBR_grid_model_best_score: 0.7993051604810518 XGBR_grid_model_score： -0.339632045000555

2、增加XGBR算法時候，調用得到的最佳參數，卻輸不出最佳參數對應的準確度！

?

3、增加XGBR算法時候，對新數據進行預測，遇到bug：因為XGBR算法，要求傳入的數據都為數值型

成功解決ValueError: DataFrame.dtypes for label must be int, float or bool

?

4、默認配置參數，輸出的準確度較低：54%
(1)、通過特征選擇可得到最高準確度為67.69%

ML之回歸預測：利用十種機器學習(LiR+xgboost(特征重要性+特征選擇))算法對無人駕駛汽車系統參數(2017年的data,18+2)進行回歸預測值VS真實值

?

XGBR：The value of default parameter of XGBR is 0.539021005879261 XGBR：R-squared value of DecisionTreeRegressor: 0.539021005879261 XGBR：測試[12.8, 13.0]行數據， [53.609283 53.47569 53.920986 53.920986 53.61446 53.61446 53.82615753.773552 53.44194 53.09419 54.089035 54.353115 53.321796 53.29324353.970306 53.752117 53.460567 53.44237 53.840027 53.920986 52.3508557.61133 57.598843 57.911274 58.06042 57.945023 57.665913] ------------------------------------------------------------ XGBR_model.feature_importances_： [0.08985916 0.01444405 0.08940411 0. 0.03163605 0.018708770.00869713 0.12159647 0.03933521 0.12161936 0.02289704 0.042608730.02663714 0.04179822 0.03441375 0.14182347 0.01409451 0.14042689] Thresh=0.000, n=18, Accuracy: 52.56% Thresh=0.009, n=17, Accuracy: 52.56% Thresh=0.014, n=16, Accuracy: 51.94% Thresh=0.014, n=15, Accuracy: 52.89% Thresh=0.019, n=14, Accuracy: 54.23% Thresh=0.023, n=13, Accuracy: 53.38% Thresh=0.027, n=12, Accuracy: 52.86% Thresh=0.032, n=11, Accuracy: 53.08% Thresh=0.034, n=10, Accuracy: 64.67% Thresh=0.039, n=9, Accuracy: 67.69% Thresh=0.042, n=8, Accuracy: 66.25% Thresh=0.043, n=7, Accuracy: 67.52% Thresh=0.089, n=6, Accuracy: 64.64% Thresh=0.090, n=5, Accuracy: 40.05% Thresh=0.122, n=4, Accuracy: 37.93% Thresh=0.122, n=3, Accuracy: 19.23% Thresh=0.140, n=2, Accuracy: 22.15% Thresh=0.142, n=1, Accuracy: 11.77%

?

?

?

?

?

?

?

?

?

?

?

?

?

?

總結

以上是生活随笔為你收集整理的ML之回归预测：利用十(xgboost,10-1)种机器学习算法对无人驾驶汽车系统参数(2017年的data,18+2)进行回归预测值VS真实值——bug调试记录的全部內容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網站內容還不錯，歡迎將生活随笔推薦給好友。