基于Pareto法则的BPA概率转换

The Probability Transformation of BPA Based on Pareto Principle

基于帕累托(Pareto)法则,该文认为复杂焦元信度的分配应该依赖于其单子命题的信度,信度大的更能决定复杂焦元的分配,信度为零的也能影响复杂焦元的分配.将复杂焦元分成2类结构:单子命题没有零值和单子命题有零值.对于前者,找出了复杂焦元的帕累托元素,只在帕累托元素上按其信度权重进行分配,这既避免了一些反常的情况又能更加突出信度大的单子命题.对于后者,采用忽略一部分或平均的办法来进行分配;设置了2个参数,既能控制焦元分配的冒险程度,又能控制转换的效果.提出了一种能够根据给定冒险程度来计算基本概率分配函数概率转换的新方法.最后通过实例分析,验证了该方法是有效的.

Based on the Pareto Principle,it is believed that the distribution of belief value of complex focal elements should depend on the belief value of its singleton propositions.A larger belief value can better determine the distribution of complex focal elements,and a zero trust value can also affect the distribution of complex focal elements.The complex focal elements are divided into two types of structures,one is that the singleton proposition has no zero value,and the other is that the singleton proposition has zero value.For the former,the Pareto elements of the complex focal elements are found,and only the Pareto elements are allocated according to their belief value weights,which avoids some abnormal situations and highlights the singleton propositions with large belief values.For the latter,the method of ignoring a part or averaging is used to allocate,two parameters are set,which can control the risk degree of focal element allocation and adjust the effect of transformation.The new method is proposed to calculate the probability transition of basic probability assignment function according to the given risk degree.Finally,the effectiveness of the method is verified by example analysis.