SCALABLE RECORD LINKAGE AND DEDUPLICATION FOR ELECTION DATABASE

Abstract

Record linkage is the technique of finding same data from collection of database that has same contents. This technique is when tested on single database that is known as de-duplication. Matching data is very much important in so many fields, because they have so many valuable information which is not possible to acquire from anywhere or it may be too costly to get. In data cleaning method removing same or duplicate data is very much essential step which applied on singe database because, in data mining same data can badly affect result of that process. With rising nature of today’s database, in record linkage and de-duplication the difficulties of matching data become one of the challenges. For record linkage and de-duplication there are various indexing techniques have been developed in last few years. Their purpose is to reduce the number records by comparing matching records with non-matching one, at the same time must maintain high matching quality. This paper presents a system in which windowing Technique is applied on database, in which records are matched, their complication is analyzed and their work and flexibility is calculated using both fake and real data set