一种K值自适应和局部搜索的KNN矩阵修复方法  

An Advanced K-Nearest Neighbors Matrix Completion Method Based on K-Value Self-Adaption and Local Search

作　　者：王社会[1,2] 杨俊安[1,2] 

机构地区：[1]电子工程学院404教研室,合肥230037 [2]安徽省电子制约技术重点实验室,合肥230037

出　　处：《电子信息对抗技术》2014年第6期60-63,共4页Electronic Information Warfare Technology

基　　金：安徽省自然科学基金(1308085QF99)

摘　　要：数据在采集和传输过程中由于多种原因会造成矩阵残缺,因此在数据分析之前需要对残缺矩阵进行修复。常见的KNN修复方法 k值选取不合理,且需在整个矩阵中搜索近邻,影响算法的修复效果。在其基础上提出了一种k值自适应的局部KNN矩阵修复方法,合理考虑了k值的选取和近邻项的搜索范围。实验证明了该方法能有效提高矩阵修复的正确率,且算法的时效性有所提高。Since the data may inevitably miss some elements in the process of information acqui-sition and transmission due to various reasons, it is necessary to full them. But the general KNN completion method is not reasonable on the selection of the value of k, and it needs to search for neighbors of the entire matrix. An advanced KNN matrix completion is presented by K-value self-adaption and local search, which reasonably considers the value of k and the search range of neighbors. The experiment indicates this method is effective, and the timeliness is also en-hanced.

关 键 词：数据缺失 矩阵修复 KNN理论 自适应K 局部最近邻 

分 类 号：TN971.1[电子电信—信号与信息处理]