Swarm optimisation as a new tool for data mining

Abstract

This paper proposes the use of Particle Swarm Optimisers as a tool for data mining. To evaluate its usefulness, we empirically compare the performance of three variants of the Particle Optimiser with another evolutionary algorithm, namely a Genetic Algorithm, in rule discovery for clussijkution tusks. Such tusks are considered core tools for Decision Support Systems in a widespread urea, ranging from the industry, commerce, military and scientljic fields. The data sources used here for experimental testing are commonly used and considered us a de facto standard for rule discovery algorithms reliability ranking. The results o htuined in these domains seem to indicate that Particle Swarm Optimisers are competitive with other evolutionary techniques, and can he successfully applied to more demanding problem domains.0-7695-1926-1/03/$17.00 (C) 2003 IEEE

Exploring comprehensible classification rules from trained neural networks integrated with a time-varying binary particle swarm optimizer

“…In order to improve the efficiency of proposed binary PSO algorithm, parameters are designed in an adaptive manner. Although there are a few studies in the data mining literature, the results prove the competitiveness of different PSO algorithms in the classification area (Wang et al, 2007;Holden and Freitas, 2008;Sousa et al, 2003). The main contribution of this study is to present the capability of binary-PSO algorithm to uncover the hidden information in a trained ANN so as to produce accurate and comprehensible classification rules.…”

Hybrid particle swarm optimization for rule discovery in the diagnosis of coronary artery disease

“…The PSO has been also applied for classification task (Sousa et al, 2003;Sousa et al, 2004). In these two works, three PSO variants have been compared.…”

“…The authors have also extended the algorithm to use k‐means clustering algorithm as the initial step to seed the swarm. The PSO has been also applied for classification task (Sousa et al, 2003; Sousa et al, 2004). In these two works, three PSO variants have been compared.…”

Swarm Intelligence in Data Mining

“…Sousa et al [61], [62] have proposed the use of PSO as a tool for data mining. In order to evaluate the usefulness of PSO for data mining, an empirical comparison of the performance of three variants of PSO with another evolutionary algorithm (Genetic Algorithm), in rule discovery for classification tasks is used.…”