Abstract—Utilising association rule discovery to learn classifiers in data mining is known as Associative Classification (AC). In the last decade, AC algorithms proved to be effective in devising high accurate classification systems from various types of supervised data sets. Yet, there are new emerging trends and that can further enhance the performance of current AC methods or necessitate the development of new methods. This paper sheds the light on four possible new research trends within AC that could enhance the predictive performance of the classifier or their quality in terms of rules. These possible research directions are considered starting research points for other scholars in rule based classification in data mining.

Index Terms—artificial intelligence, associative classification, classification, research trends