A new approach based on evaluation algorithm for classification problems in data mining

Data mining is a method for finding unknown relation and trends that analyses great amount of data and transforms insignificant information to significant knowledge for users. Classification in data mining is an approach finding out related attributes of each class and providing display with rule statements from given data sets.In this thesis a new algorithm ?Rough-Mep algorithm? is proposed for discovering of classification rules based on rough set theory and evaluation algorithms. The effectiveness of our approach is tested on eight publicly available binary and n-ary classification data sets.