PROJECT TITLE :

Unified Representation of Sets of Heterogeneous Markov Transition Matrices

ABSTRACT:

Markov chains are terribly efficient models and are extensively applied in a wide selection of fields covering queuing theory, signal processing, performance analysis, time series, and finance. For discrete finite initial-order Markov chains, which are among the foremost used models of this family, the transition matrix will be seen because the model parameter, since it encompasses the set of chances governing the system state. Estimating such a matrix is, however, not an straightforward task because of potential opposing skilled reports or variability of conditions beneath that the estimation method is allotted. In this paper, we have a tendency to propose an explicit approach to infer a consensus transition matrix, defined in accordance with the theory of proof, from a family of information samples or transition matrices. To validate our technique, experiments are conducted on nonstationary label images and daily rainfall data. The obtained results confirm the interest of the proposed evidential modeling with respect to the standard Bayesian one.