Boosted incremental nelder-mead simplex algorithm: Distributed regression in wireless sensor networks

Abstract

Wireless sensor networks (WSNs) have been of great interest among academia and industry due to their diverse applications in recent years. The main goal of a WSN is data collection. As the amount of the collected data increases, it would be essential to develop some techniques to analyze them. In this paper, we propose an in-network optimization algorithm based on Nelder-Mead simplex to incrementally do regression analysis over distributed data. Then, we improve the resulted regressor by the application of boosting concept from machine learning theory. Simulation results show that the proposed algorithm not only increases accuracy but is also more efficient in terms of communication compared to its gradient based counterparts. © 2008 International Federation for Information Processing.