Active Learning of Expressive Linkage Rules Using Genetic Programming

Abstract

A central problem in the context of the Web of Linked Data as well as in data integration in general is to identify entities in different data sources that describe the same real-world object. Many existing methods for matching entities rely on explicit linkage rules, which specify the conditions which must hold true for two entities in order to be interlinked. As writing good linkage rules by hand is a non-trivial problem, the burden to generate links between data sources is still high. In order to reduce the effort and expertise required to write linkage rules, we present the ActiveGenLink algorithm which combines genetic programming and active learning to generate expressive linkage rules interactively. The ActiveGenLink algorithm automates the generation of linkage rules and only requires the user to confirm or decline a number of link candidates. ActiveGenLink uses a query strategy which minimizes user involvement by selecting link candidates which yield a high information gain. Our evaluation shows that ActiveGenLink is capable of generating high quality linkage rules based on labeling a small number of candidate links and that our query strategy for selecting the link candidates outperforms the query-by-vote-entropy baseline.

Keywords: Entity Matching, Duplicate Detection, Link Discovery, Active Learning, Genetic Programming, Linkage Rules, ActiveGenLink