Redundancy Avoidance in Entity Resolution Based On Social Networks Paradigm

Abstract

Entity resolution (ER) aims at identifying and merging records in one or more datasets that refer to the same real-world entity. The ER problem is becoming more challenging in the context of Big Data. We study the ER problem by transforming it into a Social Network where data records can be treated as real-world entities capturing the existing relationships (e.g. friendship, householder). A framework to handle the transformation of the data model is presented and evaluated on several datasets. The framework is tested using four state-of-the-art ER, including (1) k-mean, (2) Levenshtein, (3) Jaro Winkler, and (4) Soundex on SNA in terms of time and accuracy performance metrics.