一种缺失数据的填补方法  

A complement method for missing data

作　　者：张其文[1] 李明[1] 

机构地区：[1]兰州理工大学计算机与通信学院,甘肃兰州730050

出　　处：《兰州理工大学学报》2006年第2期102-104,共3页Journal of Lanzhou University of Technology

基　　金：甘肃省自然科学基金(ZS022-A25-035);甘肃省教育厅科研基金(0416B-04)

摘　　要：分析了在含有遗失值的数据集上如何进行有效的数据填补,以便更客观地反映数据集中数据所隐含的内在联系.通过采用粗糙集理论的有关思想和方法,用相似关系代替粗糙集理论中的不可分辨关系,提出了基于相似关系的填充算法,提高了遗失数据填补的精确度.数据实验表明了该方法的有效性和可行性.The way of effective data complement for a data set with missing values was analyzed so as to reflect more objectively internal relationship among data in data set. Through making use of relevant idea and method of rough set theory, substituting similar relation for indiscernibility relation in the letter, a complement algorithm, and presents based on similar relation was presented, so that the accuracy of missing-data complementation was improved. It was shown by data experiment that this method was valid and feasible.

关 键 词：信息系统 粗糙集 遗失值 数据补全 相似类 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]