Rough Set Approach to Data Completion Based on Weighted Similarity
一种基于加权相似性的粗糙集数据补齐方法

In recent years,much attention has been given to the treatment of incomplete data. By now,many completion methods to incomplete data have been proposed in rough set theory. hhese methods usually compute the similarities between the object that contains missing values and other objects that do not contain missing values,and use the values of the most similar object to replace the missing values. However, there is a common problem for these methods. That is,these methods assume that the dependencies of decision attribute on all condition attributes arc the same, and the significances of all condition attributes are also the same,they ignore the differences between different condition attributes in a decision table. To solve this problem, in this paper we introduced a new notion of weighted similarity, which employs the dependencies of decision attribute on condition attributes and the significances of condition attributes as weights to compute the similarity. Based on the weighted similarity, we proposed a novel rough set data completion algorithm WSDCA.We compared WSDCA with the current data completion algorithms on UCI data sets. And experimental results demonstrate the effectiveness of our method to data completion.