No Detail Found!

Change-point analysis of coal mining disasters data in various time resolutions.

Author(s)
Muhammad Jaudet Department of Electrical Engineering, Pakistan Institute of Engineering and Applied Sciences, Islamabad, Pakistan
Naeem Iqbal Department of Electrical Engineering, Pakistan Institute of Engineering and Applied Sciences, Islamabad, Pakistan
Sikander M. Mirza Department of Physics and Applied Mathematics, Pakistan Institute of Engineering and Applied Sciences, P.O. Nilore, Islamabad, Pakistan
Abstract
British coal mining disasters data at various time resolutions (yearly, monthly, weekly and daily) have been analysed as stochastic Poisson process with a variety of segmentation techniques available in literature. Segmentation techniques not only include hierarchical and dynamic programming but also a novel MCMC (Markov Chain Monte Carlo) technique for finding multiple change-points has been applied. It is observed that for higher counts per bin, classical Poisson likelihood performs better than Bayesian integrated likelihoods. Reduction of bin size not only brings likelihood curves closer but Bayesian integrated likelihoods starts to segment data better than classical Poisson likelihood in terms of root mean squared error: These results of likelihood maximization and fixed dimension MCMC are in excellent agreement with the findings of other researchers.
Publication Details
Page(s) 31-40
DOI DOI not available
Published Journal: Mehran University Research Journal of Engineering and Technology, Volume: 28, Issue: 1, Year: 2009

Request the full text from PASTIC

0

Citations

0

Downloads

36

Views