Abstract: Associative classiſtation is a supervised classiſtation approach, integrating association mining and classiſtation. Several studies in data mining have shown that associative classiſtation achieves higher
classiſtation accuracy than do traditional classiſtation techniques. However, the associative classiſtation suï¬ers from a major drawback: The huge number of the generated classiſtation rules which takes eï¬orts to
select the best ones in order to construct the classi﬑r. To overcome such
drawback, we have proposed an associative classiſtation method that reduces associative classiſtation rules without jeopardizing the classiſtation accuracy. Moreover, we will introduce in this paper two diï¬erent strategies to classify new instances based on some interestingness measures that arise from data mining literature in order to select the best
rules during classiſtation. A detailed description of this method is presented in this paper, as well as the experimentation study on 12 benchmark data sets proving that our approach is highly competitive in terms
of accuracy in comparison with popular classiſtation approaches.
