eduzhai > Applied Sciences > Engineering >

Ensemble of Clustering Approaches for Feature Selection of High Dimensional Data

  • Save

... pages left unread,continue reading

Document pages: 7 pages

Abstract: As a prolific research area in data mining and machine learning, feature selection and associated problems bring an enormous number of proposed solutions. However, many researches are composed in new way, but fail to classify at some time for a particular data. In general, a feature selection methods search out local minima but it s does not obtain optimal minima. So, in this paper, we have proposed a method that comprised of ensemble techniques for feature selection which enable to achieve the minimum classification error. We proposed a hybrid approach, using bits from k-mean techniques and called it Feature Co-association Ensemble (CFE). For experimental analysis, we have used seven benchmark high dimensional datasets from UCI repository and validation proves its worthiness among the exiting approaches of feature selection methods.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×