Performance Evaluation of Rule Based Classification Algorithms

The growth of the internet has created a vast new arena for information generation. There is huge amount of data available in Information Industry. Databases today can range in size of the terabytes or more bytes of data. To address these issues, researchers turned to a new research area called Data Mining. Data mining is a technology that blends traditional data analysis methods with sophisticated algorithms for processing large volume of data. Experimental evaluation of rule based classification algorithm is performed using WEKA open source tool. Five rule based classification algorithm considered are OneR, PART, Decision Table, DTNB and Ridor algorithms. Chess End Game data set is used in experimental evaluation of algorithms .Cross validation testing technique is considered for experiment. For comparing the five algorithm three performance parameters number of classified instances, accuracy and error rate are considered. The results of experiment are presented in tabular and graphical form. From this study it is found that PART is best algorithm for classification.