EL之Bagging(DTR)：利用Bagging對回歸問題(實數值評分預測)模組化(調2參)

輸出結果

設計思路

核心代碼

bagFract = 1.0        #----------------------☆☆☆☆☆

nBagSamples = int(len(xTrain) * bagFract)

for iTrees in range(numTreesMax):

   idxBag = []

   for i in range(nBagSamples):

       idxBag.append(random.choice(range(len(xTrain))))

   xTrainBag = [xTrain[i] for i in idxBag]

   yTrainBag = [yTrain[i] for i in idxBag]

   modelList.append(DecisionTreeRegressor(max_depth=treeDepth))

   modelList[-1].fit(xTrainBag, yTrainBag)

   latestPrediction = modelList[-1].predict(xTest)

   predList.append(list(latestPrediction))