PROJECT TITLE :

Effective and Efficient Discovery of Top-k Meta Paths in Heterogeneous Information Networks

ABSTRACT:

Both academic institutions and private businesses have shown a significant amount of interest in heterogeneous information networks (HINs), which are essentially graphs with typed nodes and edges that have been labeled. We study the discovery of the k most important meta paths in real time, which can be used to support friend search, Product Recommendation, Anomaly Detection, and graph Clustering. Given two HIN nodes s and t, along with a natural number k, our research focuses on the discovery of these k most important meta paths. In this piece of work, we argue that the most important path may not necessarily be the path that is the shortest distance between s and t. Therefore, in order to redefine the unified importance function of the meta paths that connect s and t, we combine several ranking functions, which are based on frequency and rarity. Finding top-k meta paths using this importance function is a very time-consuming process, despite the fact that it can capture more information than other importance functions. As a result, we decide to incorporate this importance function into a multi-step framework. This allows the framework to filter some impossible meta paths between s and t in an effective manner. Additionally, in order to further improve the efficiency and performance of this framework, we combine a bidirectional searching algorithm with it. Based on the results of the experiment performed on a variety of datasets, our proposed method outperforms algorithms that are considered to be state-of-the-art in terms of its effectiveness while maintaining a reasonable response time.