A System Based on Swarm Particle Optimization to Extract Knowledge from Times Series Data

In this paper, a Particle Swarm Optimizer is proposed to extract the behavior patterns from time dependents phenomena. A time series can be created by the temporal phenomenon observation and the local tendency dynamic of the phenomenon can be described by a binary codification of the time series, where "1" is assigned for positive trends and "0" is assigned for otherwise. The proposed algorithm searches for the behavior patterns embedded in the time series, or rules that describe the laws that govern the dynamics of the studied phenomenon. Therefore, each particle of the swarm consists of a trial rule with a recognition window and a forecast value. A set of eight artificial time series (with and without noise) are used to evaluate the proposed method. The experimental results show that the proposed method is a promising approach for tendency forecasting and extraction of knowledge from the time series data.