An agent-based simulated annealing algorithm for data reduction

Abstract

Data reduction in the supervised machine learning aims at deciding which features and instances from the training set should be retained for further use during the learning process. Data reduction can result in increased capabilities and generalization properties of the learning model and shorter learning process time. It can also help in scaling up to a large data sources. This paper proposes an approach based on a combination of the simulated annealing technique and the multi-agent architecture designed for solving the data reduction problem. The paper includes the overview of the proposed approach and shows the computational experiment results. Experiment has shown that the proposed agent-based simulated annealing outperforms the traditional simulated annealing approach when solving the data reduction problem. © 2010 Springer-Verlag.