1.The main code is used to train different model based on the given feature and lable datasets. So just change the name of the datasets to your own ones and input the needed parameters to indicate the model you want use, oversampling method and inner cross-validation method.

2.For each algorithm, you will get a probability prediction. You first combine them together into one file, then just use the mojority_vote code to apply ensemble model.

3.For data preprocessing , you have to determine which ones are catogorical data and which ones are nummerical.Then just change the feature names to do the preprocessing .