可靠性信息融合中继承因子的确定方法
Determining Similarity Factor in Reliability Information Fusion

为了解决存在相似产品信息的可靠性信息融合问题，提出了基于证据理论确定不同相似产品信息融合时继承因子的方法，充分利用了不同相似产品之间的信息。以成败型产品为例，首先，将多种相似产品信息表示为各自的共轭先验分布并建立混合先验分布，接着给出对应的混合后验分布；然后，从先验分布的重叠面积出发，定义了不同产品之间样本一致性的概念，并结合证据理论融合不同产品的一致度信息，进而确定继承因子；最后，以成败型产品作为实例，基于提出的方法，确定其继承因子，并与其它方法进行比较，验证了该方法的有效性和可靠性。
The Dempster-Shafer (D-S) theory was used to determine similarity factors of different similar products based on reliability information fusion, which makes full use of the information on similar products. Firstly, taking binominal products for example, by expressing the reliability information on multiple similar products as respective conjugate prior distribution, a mixture prior distribution as well as a mixture posterior distribution was built up. Then, considering the overlapping area of a prior distribution, the sample consistency for similar products was defined. Then combined with the D-S theory, the method for infusing reliability information on various similar products and determining similarity factors was proposed. Last but not least, a binominal example, given in the full paper, proves that the method is more effective than other methods