EL之GB(GBR):利用GBR对回归问题(实数值评分预测)建模
输出结果
![](http://image109.360doc.com/DownloadImg/2021/09/2818/231252069_5_20210929015654755)
![](http://image109.360doc.com/DownloadImg/2021/09/2818/231252069_6_2021092901565583)
![](http://image109.360doc.com/DownloadImg/2021/09/2818/231252069_7_20210929015655349.png)
设计思路
![](http://image109.360doc.com/DownloadImg/2021/09/2818/231252069_8_20210929015655521)
核心代码
wineGBMModel = ensemble.GradientBoostingRegressor(n_estimators=nEst,
max_depth=depth,
learning_rate=learnRate,
subsample = subSamp,
loss='ls')
wineGBMModel.fit(xTrain, yTrain)
msError = []
predictions = wineGBMModel.staged_predict(xTest)