No abstract

We present the AgreementMaker system for matching real-world schemas and ontologies, which may consist of hundreds or even thousands of concepts. The end users of the system are sophisticated domain experts whose needs have driven the design and implementation of the system: they require a responsive, powerful, and extensible framework to perform, evaluate, and compare matching methods. The system comprises a wide range of matching methods addressing different levels of granularity of the components being matched (conceptual vs. structural), the amount of user intervention that they require (manual vs. automatic), their usage (stand-alone vs. composed), and the types of components to consider (schema only or schema and instances). Performance measurements (recall, precision, and runtime) are supported by the system, along with the weighted combination of the results provided by those methods. The AgreementMaker has been used and tested in practical applications and in the Ontology Alignment Evaluation Initiative (OAEI) competition. We report here on some of its most advanced features, including its extensible architecture that facilitates the integration and performance tuning of a variety of matching methods, its capability to evaluate, compare, and combine matching results, and its user interface with a control panel that drives all the matching methods and evaluation strategies.

No abstract

Abstract.Using a pay-as-you-go strategy, we allow for a community of users to validate or invalidate mappings obtained by an automatic ontology matching system using consensus for each mapping. The ultimate objectives are effectiveness-improving the quality of the obtained alignment (set of mappings) measured in terms of F-measure as a function of the number of user interactionsand robustness-making the system as much as possible impervious to user validation errors. Our strategy consists of two major steps: candidate mapping selection, which ranks mappings based on their perceived quality so that users are presented first with those mappings with lowest quality, and feedback propagation, which seeks to validate or invalidate those mappings that are perceived to be similar to the mappings already presented to the users for validation. The purpose of these two strategies is twofold: achieve greater improvements earlier and minimize overall user interaction. There are three important features of our approach: the use of a dynamic ranking mechanism to adapt to the new conditions after each user interaction, the presentation of each mapping for validation more than once-revalidation-because of possible user errors, and the immediate propagation of the user input on a mapping without first achieving consensus for that mapping. We study extensively the effectiveness and robustness of our approach as several of these parameters change, namely the error and revalidation rates, as a function of the number of iterations, to provide conclusive guidelines for the design and implementation of multi-user feedback ontology matching systems.

Abstract. An ontology matching system can usually be run with different configurations that optimize the system's effectiveness, namely precision, recall, or F-measure, depending on the specific ontologies to be aligned. Changing the configuration has potentially high impact on the obtained results. We apply matching task profiling metrics to automatically optimize the system's configuration depending on the characteristics of the ontologies to be matched. Using machine learning techniques, we can automatically determine the optimal configuration in most cases. Even using a small training set, our system determines the best configuration in 94% of the cases. Our approach is evaluated using the AgreementMaker ontology matching system, which is extensible and configurable.