Learning to Match Schemata using Predictors

We propose a learning algorithm that utilizes an innovative set of features to re-rank a list of top-K matches and improves upon the ranking of the best match. We provide a bound on the size of an initial match list, tying the number of matches in a desired level of confidence for finding the best match. We also propose the use of schema matching predictors as features in the learning task, and tailored nine new predictors for this purpose. A large scale empirical evaluation with real-world benchmark schema sets show the effectiveness of the proposed algorithmic solution for reranking top-K schema matches.