Ant colony optimization for mining gradual patterns

Gradual pattern extraction is a field in Knowledge Discovery in Databases that maps correlations between attributes of a data set as gradual dependencies. A gradual dependency may take the form: the more Attribute(K), the less Attribute(L). Classical approa-ches for extracting gradual patterns extend either a breath-first search or a depth-first search strategy. However, these strategies can be computationally expensive and inefficient especially when dealing with large data sets. In this study, we investigate 3 population-based optimization techniques (i.e. ant colony optimization, genetic algorithm and particle swarm optimization) that may be employed improve the efficiency of mining gradual patterns. We show that ant colony optimization technique is better suited for gradual pattern mining task than the other 2 techniques. Through computational experiments on real-world data sets, we compared the computational performance of the proposed algorithms that implement the 3 population-based optimization techniques to classical algorithms for the task of gradual pattern mining and we show that the proposed algorithms outperform their classical counterparts.

Automated summary

Gradual pattern extraction is a field in Knowledge Discovery in Databases that aims to map correlations between attributes of a data set as gradual dependencies. In this study, three population-based optimization techniques are investigated to improve the efficiency of mining gradual patterns. The results show that ant colony optimization technique outperforms genetic algorithm and particle swarm optimization in the task of gradual pattern mining.