MOMA - A Mapping-based Object Matching System

Object matching or object consolidation is a crucial task for data integration and data cleaning. It addresses the problem of identifying object instances in data sources referring to the same real world entity. We propose a flexible framework called MOMA for mapping based object matching. It allows the construction of match workflows combining the results of several matcher algorithms on both attribute values and contextual information. The output of a match task is an instance-level mapping that supports information fusion in P2P data integration systems and can be re-used for other match tasks. MOMA utilizes further semantic mappings of different cardinalities and provides merge and compose operators for mapping combination. We propose and evaluate several strategies for both object matching between different sources as well as for duplicate identification within a single data source. Object matching (also known as object consolidation, duplicate identification, record linkage, entity resolution or reference reconciliation) is a crucial task for data integration and data cleaning [9, 18, 27] and its history goes back over 20 years [3, 19]. It addresses the problem of identifying object instances referring to the same real world entity. The instances may reside in different, typically heterogeneous data sources or may already be stored in a single data source, e.g., in a structured database or a search engine store. The instances to be consolidated may be physically materialized or dynamically be requested from sources, e.g. by database queries or keyword searches. The high importance and difficulty of the object matching problem has triggered a huge amount of research on different variations of the problem. Most previously proposed approaches focus on matching relational records and apply what we call attribute matching, i.e., they use the values of selected attributes to determine the similarity between instances. These approaches have also been called " fuzzy join " or " fuzzy match " [6, 7] and considered different similarity functions (e.g., different types of string similarity) [10] and their efficient implementation. More recently, authors have recognized the value of considering additional information for object consolidation, in particular semantic relationships or mappings. For instance, [4] uses the co-authorship relationship to match author instances , i.e., two authors are considered as highly similar if they have the same co-authors. Several studies proposed graph matching algorithms to consider such contextual information, e.g., [4, 8, 11, 35]. The use of relationship information is especially promising when the effectiveness of …