Chemo-inspired Genetic Algorithm for Optimizing the Piecewise Aggregate Approximation

Abstract

In a previous work we presented DEWPAA: an improved version of the piecewise aggregate approximation representation method of time series. DEWPAA uses differential evolution to set weights to different segments of the time series according to their information content. In this paper we use a hybrid of bacterial foraging and genetic algorithm (CGA) to set the weights of the different segments in our improved piecewise aggregate approximation. Our experiments show that the new hybrid gives better results in time series classification.