Biclustering in data mining using a memetic multi-objective evolutionary algorithm

Abstract

In this paper, a new memetic strategy that integrates a multi-objective evolutionary algorithm (the SPEA2) with a local search technique for data mining is presented. The algorithm explores a Term Frequency-Inverse Document Frequency (TF-IDF) data matrix in order to find biclusters that fulfill several objectives. The case of study was a dataset corresponding to the Reuters-21578 corpus. Our algorithm performed satisfactorily, finding biclusters that have large size and coherent values, yielding to undeniably promising outcomes. Nonetheless, more experiments with data from other corpus are necessary, thus leading to more concluding results