Personnel BDAR Ability Assessment Model Based on Bayesian Stochastic Assessment Method

Abstract

It is important for us to evaluate each trainee’s ability with the aim of improving BDAR training efficiency. The typical assessment methods include fuzzy evaluation, gray correlation evaluation, neural network and so on. All of these methods are not able to make full use of the historical information. Determining membership function in first two methods is not easy. And ANN needs a lot of data sample which is difficult to obtain in BDAR training. So we can’t use these methods to model the assessment of personnel BDAR ability. Then we introduce Bayesian Stochastic Assessment Method which can deal well with the nonlinear and random problem. Each indexes’ standard is given according to the characteristics of BDAR training. A modified normal distribution which can make full use of historical information was put forward to determine the prior probability. And the poster probability is determined by distance method.