基于核密度估计的基本概率指派生成方法  

New method to determine BPA based on kernel density estimation

作　　者：黄杰 尉永清[3] 伊静[1,4] 刘孟迪 Huang Jie;Wei Yongqing;Yi Jing;Liu Mengdi(School of Information Science&Engineering,Shandong Normal University,Jinan 250358,China;Shandong Provincial Key Laboratory for Distributed Computer Software Novel Technology,Jinan 250014,China;Dept.of Basic Education,Shandong Police College,Jinan 250014,China;School of Computer Science&Technology,Shandong Jianzhu University,Jinan 250014,China)

机构地区：[1]山东师范大学信息科学与工程学院,济南250358 [2]山东省分布式计算机软件新技术重点实验室,济南250014 [3]山东警察学院公共基础部,济南250014 [4]山东建筑大学计算机科学与技术学院,济南250014

出　　处：《计算机应用研究》2020年第7期2037-2040,2044,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(61373148);山东省自然科学基金资助项目(ZR2014FL010);山东省教育厅基金资助项目(J15LN34);山东省社科规划项目(17CHLJ18,17CHLJ33,17CHLJ30)。

摘　　要：D-S合成方法作用的对象是基本概率指派(basic probability assign,BPA),如何生成BPA是D-S理论应用中重要且有待解决的首要步骤。针对生成BPA提出一种基于核密度估计(kernel density estimation,KDE)的BPA生成方法:训练数据用于构建基于最优化窗宽的核密度估计的数据属性模型;然后利用训练数据的核密度模型计算测试数据的密度—距离—分布值Tri-D(density-distance-distribution),通过嵌套式的方法分配Tri-D值获取测试数据对应的BPA;最后D-S合成BPA得到最终判断,通过分类准确率来判断BPA生成方法的有效性。实验通过在UCI数据集上与其他方法的分类准确率对比验证了提出方法的有效性。The determination of BPA or the action object of D-S fusion method is an open problem in process of D-S theory application.This paper proposed a BPA determination method based on KDE.The method used training data to construct a data attribute model with optimized bandwidth based on the optimized kernel density estimation,then calculated the Tri-D value of test data by using the kernel density model of training data.The next step was obtaining BPA of test data by using the nested method to assign Tri-D.Finally,it fused BPA by D-S method to get the final result,and judged the validity of the BPA generation method by the classification accuracy rate.An illustrative case regarding the classification accuracy compared with other methods on UCI data sets shows the effectiveness of the method.

关 键 词：基本概率指派 核密度估计 Tri-D 窗宽 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]