I am working on microarray datasets. This type of dataset is continuous especially Colon. If i want to apply feature selection such as chisquare , IG , etc..when i upload colon dataset to perform feature selection, both chi square and other algorithms seems inactive,Therefore, is it necessary to normalize all dataset between [0,1] before feature selection. From ChiSquaredAttributeEval code it seems that its dicretize/ binarize the dataset prior FS.
What the data format appropriate to these feature selection algorithms ?
Thanks