A k-NN-based approach using MapReduce for meta-path classification in heterogeneous information networks

Abstract

Classification of the nodes along with the interconnected semantic edges in a Heterogeneous Information Network (HIN) has a lot of significance in identifying the class labels which involves the application of knowledge and dissemination of knowledge from one node to the other. In this paper, the authors applied PathSim similarity measure for finding k-nearest neighbors along with the use of the well-known MapReduce paradigm to classify the meta-paths in a Heterogeneous Information Network. Applying MapReduce simplified the classification approach which deals with huge data present in the Heterogeneous Information Networks. Experiments were carried out on movie theater dataset, and the results are accurate and successful.