Handling Missing Data in Extended Possibility-based Fuzzy Relational Databases

Handling missing data is widely studied to make proper replacement and reduce uncertainty of data. Several approaches have been proposed for providing the most possible results. However, few studies provide solutions to the problem of missing data in extended possibility-based fuzzy relational (EPFR) databases. This type of problem in the context of EPFR databases is difficult to resolve because of the complexity of the data involved. In this paper, we propose an approach of filling missing data and query processing of the databases. To obtain the rational predict of the missing data, we adopt a concept and measurement of proximate equality of tuples to define data operation and fuzzy functional dependency (FFD). We provide a method to predict the missing data and replace the data based on our proposal. The results of the missing value process preserve those FFDs that hold in the original database instance.