基于多类支持向量机的基本信度分配确定方法

Method of determining BBA based on multi-class support vector machine

在多传感器信息融合中,证据理论是一种重要方法,但其基本信度分配的确定已成为实际应用的瓶颈问题。针对此,提出了一种基于多类支持向量机的基本信度分配确定方法,以Platt后验概率为基础,将两类后验概率扩展到多类,并对后验概率形式的软判决进行了改进,从而构造了信度分配函数。通过仿真对比分析验证了该方法的合理性和有效性。

Evidence theory was an important method in the field of multi-sensor information fusion,but its BBA was dif-ficult to determine so that it became a bottleneck in practical applications.Therefore,a novel method of determining BBA based on multi-class support vector machine was proposed.On the basis of Platt posteriori probability,the two-class pos-teriori probabilities were extended to multi-class in the method.And then the soft decision in the form of posteriori prob-ability was improved and the BBA function was constructed.The rationality and effectiveness of the proposed method are verified by the simulation analysis finally.