Vol 17 no.2 2017

Ameen A. O., Olagunju M., Awotunde J. B., Adebakin T.O., Alabi I.O.

Department of Computer Science, University of Ilorin, Ilorin, Nigeria

Abstract

This paper conducted a performance evaluation on the most commonly data mining algorithms: Support Vector Machines (Radial basis function), C4.5 decision tree algorithm and Adaboost, using the two previous algorithms as base classifiers (ensemble approach), on breast cancer diagnostic removing redundant or irrelevant features using Chi-square. Result shows that while C4.5 builds its classification model in a short time, The Adaboost with SVM as its base classifier when three features are removed proved to be the best algorithm in classifying breast cancer.

Full Text:

PDF