Question

What should be done when there is high similarity for both classes in data for classification?

Answer 1

(1) Find similarity of each attribute with another(2) Find similarity of each attribute with class label.Now the attributes which have maximum similarity or dependency with other attributes and at the same time minimum similarity with class label should be removed before classification as they contribute minimum in classification process or sometimes degrades the performance also.Now please suggest me some methods through which I can find similarity among attributes.One of similarity measure that I have find is (1) Correlation(2) Mutual Information