-
摘要: 在处理不确定信息的方法中,证据推理性能突出、应用广泛,它的鲁棒性分析十分必 要.首先基于证据推理的目的和要求,给出证据推理鲁棒性的定义,并提出冲突率的新概念,以 此为尺度对Dempster规则、加权分配冲突法和吸收法的鲁棒性进行了深入分析,证明了各方法 鲁棒性结果.仿真结果表明改进的组合规则鲁棒性能得到了明显增强.最后给出证据推理的一 般情形的鲁棒性分析.Abstract: Evidence reasoning has good performance in dealing with uncertain information and has been used in many fields. It is necessary to analyze its robustness. We first give the definition of robustness based on the aim of evidential theory and the define a new concept "conflict rate". Using the sectional conflict rate, we analyze the robust area of Dempster's rule, weigh distribution and absorptive method. Finally, computer simulation indicates that, two new rules are more robust than Dempster's rule. Moreover, we give a robust area of Dempster's rule according to the general identification frame.
-
Key words:
- Evidence reasoning /
- conflict rate /
- robustness /
- information fusion
计量
- 文章访问数: 2642
- HTML全文浏览量: 119
- PDF下载量: 1334
- 被引次数: 0