A Review on Rule Based Method for Entity Resolution

Abstract

In real-world scenario entity may appear in multiple data sources so that the entity may have quite different descriptions. Hence, it is necessary to identify the records referring to the same real-world entity, which is named as Entity Resolution (ER).This paper highlights ER as one of the most important problems in data cleaning and arises in many applications like information integration and information retrieval. Traditional ER approaches are ineffective to find records based on pair wise similarity comparisons, which assumes that records referring to the same entity are very similar to each other than otherwise. However for certain circumstances this assumption does not always hold and similarity comparisons do not work very well when such assumption breaks. So to overcome traditional ER drawback a set of rules which could explain the complex matching conditions between records and entities is proposed such as rule discovery algorithm and rule based ER algorithm.