I want to do this by $70:
(1) edit the error in stacked autoencoder code. This code used as a feature selection algorithm.
(2) Add LSTM code as a classification method the first time, then CNN as a classification method another time.
Please I want it by proper settings that appropriate for the type of the dataset and the purpose of classification.
(3) Count the mean and median for each case in the code for all the 5 folds. (i used Kfold function , 5 folds).
Example: the mean of the results of the feature selection with classifier LSTM in (fold1 + fold2 + fold3 +fold4 +fold5).
then the mean of the results of the feature selection with classifier CNN in (fold1 + fold2 + fold3 +fold4 +fold5).
The type of dataset contains software metrics.
The purpose of this methodology to predict defects in software.
X refers to all attributes.
Y refers to the class.
The class is the last attribute in the dataset that determines defect or not defect, buggy or clean,...etc.
the work on python 3.6 on pycharm.