Association Rules to Recover the Missing Data Value for an Attribute in a Database

摘要

Missing data values for particular attributes are a common problem in certain databases. When analyzing the data, ignoring the records with the missing data values may not be a feasible solution since the ignored records might have essential information for the analysis. There are several statistical based methods to remedy the problem of missing values. In our paper, we propose a method that uses association rules, a data mining technique, to predict the missing value for a particular attribute in a given database. Using our proposed method for each attribute with missing values, all missing values can be predicted with a certain confidence level.