How can i calculate number of correctly classified test instance for each base classifier and finally for bagging classifier?

Question

in my task i almost did everything, but i could not find a way to do that.

Firstly I created an instance of multi-layer perceptron network with these attributes, (fou