A study of interestingness measures for knowledge discovery in databases—a genetic approach

Abstract

One of the vital areas of attention in the field of knowledge discovery is to analyze the interestingness measures in rule discovery and to select the best one according to the situation. There is a wide variety of interestingness measures available in data mining literature and it is difficult for user to select appropriate measure in a particular application domain. The main contribution of the paper is to compare these interestingness measures on diverse datasets by using genetic algorithm and select the best one according to the situation.