About this Journal Submit a Manuscript Table of Contents
International Journal of Distributed Sensor Networks
Volume 3 (2007), Issue 3, Pages 311-330
Original Article

Estimation of the Hyperexponential Density with Applications in Sensor Networks

1School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA
2Department of Computer Science, University of Texas at Dallas, Richardson, Texas, USA

Copyright © 2007 Hindawi Publishing Corporation. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


This paper solves the problem of estimation of the parameters of a hyperexponential density and presents a practical application of the solution in sensor networks. Two novel algorithms for estimating the parameters of the density are formulated. In the first algorithm, an objective function is constructed as a function of the unknown component means and an estimate of the cumulative distribution function (cdf) of the hyperexponential density. The component means are obtained by minimizing this objective function, using quasi-Newtonian techniques. The mixing probabilities are then computed using these known means and linear least squares analysis. In the second algorithm, an objective function of the unknown component means, mixing probabilities, and an estimate of the cdf is constructed. All the 2M parameters are computed by minimizing this objective function, using quasi-Newtonian techniques. The developed algorithms are also compared to the basic EM algorithm, and their relative advantages over the EM algorithm are discussed. The algorithms developed are computationally efficient and easily implemented, and hence, are suitable for low-power and sensor nodes with limited storage and computational capacity. In particular, we demonstrate how the structure of these algorithms may be exploited to be effectively utilized in practical situations, and are hence ideal for sensor networks.