Improved Nature-Inspired Algorithms for Numeric Association Rule Mining

Nowadays, only a few papers exist dealing with Association Rule Mining with numerical attributes. When we are confronted with solving this problem using nature-inspired algorithms, two issues emerge: How to shrink the values of the upper and lower bounds of attributes properly, and How to define the evaluation function properly? This paper proposes shrinking the interval of attributes using the so-called shrinking coefficient, while the evaluation function is defined as a weighted sum of support, confidence, inclusion and shrink coefficient. The four nature-inspired algorithms were applied on sport datasets generated by a random generator from the web. The results of the experiments revealed that, although there are differences between selecting a specific algorithm, they could be applied to the problem in practice.