Hi Suraj

I have done over sampling to reduce the biasedness of result.

I haven't change the values of minority class just increased the number datapoints with same values so that number of data points in class A will be

equals to class B That will build more generalized model with balanced dataset.

let's say;

a lady has two children then i ,i repeatedly taken the same values again and again so that the number of datapoints of class A will be equal to class B .As values has not been changed it won't affect the final prediction.