Comparative Study of Heuristic Hybrid of Markov Chain Monte Carlo and Dynamic Programming Methodologies for Network Fault Analysis

Citation

Jaudet M, Iqbal N, Mirza NM, Mirza SM & Hussain A (2007) Comparative Study of Heuristic Hybrid of Markov Chain Monte Carlo and Dynamic Programming Methodologies for Network Fault Analysis. International Journal of Computer Science and Network Security, 7 (4), pp. 32-41. http://paper.ijcsns.org/07_book/200704/20070405.pdf

Abstract
Modeling of network-faults based time-sequence data by piecewise constant intensity function has been carried out using a heuristic approach that employs both Markov Chain Monte Carlo approach (MCMC) and Dynamic Programming algorithm (DPA) methodologies. The results for synthetic as well as for real data show that both MCMC and DPA have close agreement between predicted and actual values. Remarkable speedup (4 to 5 times) has been observed by augmentation of the heuristic method. Due to higher efficiency the proposed approach is well suited for cases with larger data sets requiring near-optimal solution.

Keywords
Data mining; Dynamic programming; Event sequence; Change-points; Maximum likelihood; MCMC

Journal
International Journal of Computer Science and Network Security: Volume 7, Issue 4