Path-Based Ranking and Clustering

Abstract

As newly emerging network models, heterogeneous information networks have many unique features, e.g., complex structures and rich semantics. Moreover, meta path, the sequence of relations connecting two object types, is an effective tool to integrate different types of objects and mine the semantic information in this kind of networks. The unique characteristics of meta path make the data mining on heterogeneous network more interesting and challenging. In this chapter, we will introduce two basic data mining tasks, ranking and clustering, on heterogeneous information network. Furthermore, we introduce the HRank method to evaluate the importance of multiple types of objects and meta paths, and present the HeProjI algorithm to solve the heterogeneous network projection and integration of clustering and ranking tasks.