A Discriminative Approach to Ontology Mapping.
Michael Wick, Khashayar Rohanimanesh, Andrew McCallum, AnHai Doan · 2008
Techniques for automatically performing ontology mapping are vital for many real-world applications. Unfortunately, the problem is dicult because many types of evidence must be integrated to make good alignment decisions, and these decisions are co-dependent. In this paper, we propose a conditional random eld (CRF) for ontology mapping which combines probabilistic machine learning and dependencies among the prediction. We integrate multiple sources of evidence using clauses in rst-order logic, and learn corresponding weights directly from labeled training data. Our experiments show examples of impressive gains when tested on a commonly used mapping corpus; our method achieves an average of 11% (absolute) improvement in F1 when compared to other systems. We also show that our CRF is capable of generalizing from one mapping domain to another|making our supervised approach applicable for domains that lack labeled training data.