Comparison of the classification rules generated by See 5.0 and SSCO Systems

Archive - Central European Conference on Information and Intelligent Systems, CECIIS - 2012

Visnja Ognjenovic, Vladimir Brtka, Ivana Berkovic, Eleonora Brtka

Last modified: 2012-07-12

Abstract

This paper presents the comparison of the classification rules generated by See5.0/C5.0 and SSCO systems. See5.0/C5.0 system is based on C4.5 algorithm, while SSCO system is based on an algorithm, theoretically correlated to Rough Set Theory. Both systems generate classification rules in the IF THEN form.

The goal of comparison of the classification rules, generated by those two systems is detection and extraction of important rules in the terms of classification power. Some experimental comparison of two systems has been done using the Wisconsin Breast Cancer Database (January 8, 1991), obtained from UCI Machine Learning Repository.