计算机工程与应用2016，Vol.52Issue(6)：74-79,6.DOI:10.3778/j.issn.1002-8331.1509-0114
基于属性相关度的缺失数据填补算法研究
Algorithm study on missing data imputation based on attribute relevancy
摘要
Abstract
In view of less accurate in complement problem of incomplete information system, a missing data imputation algorithm is proposed based on attribute relevancy in aquaculture safety warning information system. According to study of the limited tolerance relation and decision rules, a new limited compatibility class is solved by the redefined limited compatibility relation. The relevancy of conditional attributes is introduced to construct a new extended matrix and impute data on the premise of effective guarantee deterministic to realize the completeness warning information system. Taking the missing data imputation of perch cultured as a case and using the data sets to fill verification, it shows the algorithm is superior to others on imputation accuracy and data reinforcement.关键词
不完备信息系统/限制相容关系/相关度/扩展矩阵/数据集Key words
incomplete information system/limited compatibility relation/relevancy/extended matrix/data set分类
信息技术与安全科学引用本文复制引用
毛玫静,鄂旭,谭艳,杨明婧..基于属性相关度的缺失数据填补算法研究[J].计算机工程与应用,2016,52(6):74-79,6.基金项目
辽宁省自然科学基金（No.2014020141）；辽宁省百千万人才基金择优资助项目（No.2012921058）；辽宁省社会科学规划基金重点项目（No.L14AGL001）。 （）
