Title :
The application of evaluation method based on HMM for results validity of complex simulation system
Author :
Song, Heng-Jie ; Yang, Ming ; Wang, Zi-cai
Author_Institution :
Simulation Res. Center, Harbin Inst. of Technol., China
Abstract :
According to the characteristic of randomization and temporal logic in complex simulation systems, a new evaluation method based on hidden Markov model (HMM) is presented, which applies multivariate statistical theory to quantificationally evaluate the results validity of complex simulation system. By importing matrix of observed state vector, the method enhances the clarity of description on scenario and running states of simulation systems, which ultimately implements an exploring approach to quantitative analysis for the results validity. Furthermore, quantificational evaluation criterion of results validity is given and the critical algorithm adopted in the process of quantificational evaluation is discussed in detail.
Keywords :
hidden Markov models; matrix algebra; simulation; statistical analysis; temporal logic; vectors; HMM; complex simulation system; hidden Markov model; multivariate statistical theory; quantificational evaluation criterion; quantitative analysis; randomization characteristic; state vector; temporal logic; Accreditation; Analytical models; Cybernetics; Electronic mail; Hidden Markov models; Logic; Machine learning; Usability; Weapons; Complex simulation system; HMM; evaluation; results validity;
Conference_Titel :
Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
Conference_Location :
Guangzhou, China
Print_ISBN :
0-7803-9091-1
DOI :
10.1109/ICMLC.2005.1527752
