Exploration of Data Mining Techniques in Record Deduplication

In todays business world, the database plays a vital role in decision making. As the organization grows, the size of the database also gets increased. This enormous growth in the database size leads to a problem of dirty data. Dirty data is the replicated data in the database which causes some issues like performance degradation, increasing operational cost and the lack of quality. This can be removed by the process of record deduplication. The record deduplication refers to identifying the same entity with different representations. Further cleaning and removing of replica in the repository become a mandatory work. Thus this paper surveys some of the record deduplication approaches. Also it compares with three approaches to record deduplication such as genetic programming, Modified BAT algorithm, and firefly algorithm approach with its limitation and advantages on all the three got discussed.