基于PSR验前信息可信度的Bayes数据融合方法

Bayes Data Fusion Method Based on Prior Information Credibility of PSR

针对传统Bayes数据融合方法在解决小子样试验鉴定问题时存在的局限性,提出了一种基于相似反映比例(Proportion of Similar Response,PSR)验前信息可信度的Bayes数据融合方法。该方法从Bayes理论出发,在考虑验前信息可信度的基础上,用PSR检验值作为信息可信度的度量,并考虑了射击试验不同组数据之间的差异,解决了"数据淹没"问题以及验前概率与使用方风险计算的难题。应用案例表明:该方法可以较好地将仿真数据和试验数据进行融合,适用于武器系统射击精度及弹药密集度试验的相关计算。

Aiming at limitation of traditional Bayes data fusion method in solving small sample test appraisal,Bayes data fusion method based on prior information credibility of Proportion of Similar Response(PSR) is presented.The method is based on Bayes theory,on the basis of considering prior information credibility,PSR test value is made as measure of information credibility,and differences among different groups of firing test data are considered,so as to solve difficult problem of data obliteration and calculation between prior probability and consumer's risk.Application example shows that the method can fuse simulation data and test data more effectively,and is suitable for use in calculation of firing accuracy and ammunition intensity of the weapon system.