MRE۲C: A method for constructing multi relational ensemble classifier based on two-step combining classifiers

In this work, we introduce MRE۲C method for classifying multi relational data. Multi-relational data are stored on relational databases where they consist of multiple relations that are linked together by entity-relationship links. MRE۲C creates multiple different feature subsets of relational database and then applies traditional classifiers as base classifiers. Final by using a proposed two-step combining classifier method, the results of base classifiers are combined. In first step, the proposed method uses local voting to create meta-features and then it learns meta learner to combine predication of base classifiers. Testing has been performed on two databases and six benchmark tasks. We compare our proposed method with other state-of-the-art multi relational classification methods which use different approaches to deal with multi relational setting. We showed that the proposed method achieves promising results in experiments.

کلیدواژه ها:

multi relational classification ، relational database ، Ensemble Learning ، meta learning ، طبقه بندی چند رابطه ای ، پایگاه داده های رابطه ای ، یادگیری تجمیعی ، یادگیر متا