Self-adjusting Associative Rules Generator for Classification : An Evolutionary Computation Approach

The problem of generating efficient association rules can seen as search problem since many different sets of rules are possible from a given set of instances. As the application of evolutionary computation in searching is well studied, it is possible to utilize evolutionary computation in mining for efficient association rules. In this paper, a program known as self-adjusting associative rules generator (SARG) is described. SARG is a data mining program which can generate associative rules for classification. It is an improvement of the data mining program called genetic programming for inductive learning (GPIL). Both utilize evolutionary computation in inductive learning. The shortcoming of GPIL lies in the operations crossover and selection. These two operations were inflexible and not able to adjust themselves in order to select suitable methods for the task at hand. SARG introduces new method of crossover known as MaxToMin crossover together with a self-adjusting reproduction. It has been tested on several benchmark data sets available in the public domain. Comparison between GPIL and SARG revealed that SARG achieved better performance and was able to classify these data sets with higher accuracy. The paper also discusses relevant aspects of SARG and suggests directions for future work