International Journal of Computer Science Issues

More than a traditional journal...

Data mining technology is becoming increasingly important and popular due the huge amounts of digital data that is stored globally. It provides methods and techniques to analyze these huge data repositories to extract useful information, which then is used to feed the decision making process. Classification is one of the data mining approaches to analyzing data. Other popular approaches are association rule mining and clustering. Various classification techniques have been identified in the literature including decision tree classification, rule-based classification, nave Bayesian classification, Bayesian belief networks, and rule-based nave Bayesian classification. One of the main differences between these classification techniques is the representation scheme used by each classification technique. A representation scheme captures the classification criteria and knowledge that a system learns from a pre-classified training set. In this paper we provide a comparative assessment of some these representation schemes and describes the advantages and disadvantages of each classification technique and its underlying representation scheme.