Using decision trees to summarize associative classification rules

Association rule mining is one of the most popular issues within data mining. It discovers items that co-occur frequently within a set of transactions and determines rules based on these co-occurrence relations. Classification problems have adopted association rules for years (associative classification). Once the rules have been generated, however, their lack of organization causes a readability problem, meaning it is difficult for users to analyze them and obtain a good understanding of the domain. Therefore, our work presents two algorithms that use decision trees to summarize associative classification rules. The obtained classification model combines the advantages of associative classification and decision trees. It organizes knowledge in a more readable, compact, well-organized form and is easier to use than associative classification. It also provides better classification accuracy than the traditional decision tree algorithm.