基于Stacking方法的多策略本体映射*

将概念相似度的计算问题看做分类问题，提出一种基于Stacking方法的多策略本体映射框架；利用Stacking方法组合多种概念相似度算法，进而提出基于Widrow-Hoff理论的元数据分类算法LMSMC。该框架中，第0层分类器使用各种概念相似度算法对源本体进行分类，第1层分类器使用LMSMC算法对元数据进行分类，从而实现组合多种算法的本体映射。实验表明该方法比单独使用相似度算法在查全率、查准率上均有所提高。

Multi-strategy ontology mapping based on Stacking method

This paper regarded the problem of concept similarity computing as the problem of classifying,and proposed the multi-strategy ontology mapping framework based on Stacking method,with the multiple concept similarity computing strategies being combined with Stacking method.Also presented a metadata classifying algorithm LMSMC based on Widrow-Hoff theory. In this framework, layer 0 classified the target ontology with the current concept similarity computing methods, and layer 1 classified the metadata with LMSMC algorithm, and then made the ontology mapping which combined multiple algorithms.Experiments indicate that the method performs better in improving precision and recall compared to individually using other algorithms.