Evaluation method based on markov chain model

Abstract

The evaluation outcome often has time lags. As a result, the decision makers often make decision with time lags. In order to deal with the problem, a novel evaluation method based on Markov chain was put forward. Make the mark sequence given by experts for each index a Markov chain, the initial distribution and transition probabilities matrix can be computed based upon the statistical analysis of the marks. Further more, the stationary distribution can be obtained. The mathematical expectation of the marks was made the weight of each index. The weight computed based upon the initial distribution, transition probability matrix and stationary distribution were treated as the initial weight, evolutional weight and stationary weight respectively. When the expertspsila survey is available, the weight can be updated in a timely basis and the amount of information is enlarged. Then the decision makers can make decision more scientifically in time. At last, a descriptive example, the effect of which is favorable, was given to demonstrate the method.