User guided entity similarity search using meta-path selection in heterogeneous information networks

Abstract

With the emergence of web-based social and information applications, entity similarity search in information networks, aiming to find entities with high similarity to a given query entity, has gained wide attention. However, due to the diverse semantic meanings in heterogeneous information networks, which contain multi-typed entities and relationships, similarity measurement can be ambiguous without context. In this paper, we investigate entity similarity search and the resulting ambiguity problems in heterogeneous information networks. We propose to use a meta-path-based ranking model ensemble to represent semantic meanings for similarity queries, exploit the possibility of using using user-guidance to understand users query. Experiments on real-world datasets show that our framework significantly outperforms competitor methods.

iQbees: Interactive Query-by-example Entity Search in Semantic Knowledge Graphs

“…Yu et al [11] solve a slightly different problem where entities similar to a single query entity are computed, exploiting a small number of example results. Focusing on heterogeneous similarity aspects, they propose to use features based on so-called meta paths between entities and several path-based similarity measures, and apply learning-torank methods for which they require labelled test data.…”

Integrating Vertex-centric Clustering with Edge-centric Clustering for Meta Path Graph Analysis

“…VEPathCluster improves the clustering quality by utilizing four novel mining strategies: (1) edge-centric random walk model; (2) clustering-based multigraph model; (3) integration of vertex-centric clustering and edge-centric clustering; and (4) dynamic weight learning. VEPathCluster iteratively performs the following three tasks to achieve high quality clustering: (1) fix edge clustering and weight assignment to update vertex clustering; (2) fix vertex clustering and weight assignment to update edge clustering; and (3) fix vertex clustering and edge clustering to update weight assignment.…”

“…Meta path-based social network analysis is gaining attention in recent years [1][2][3][4][5][6]. PathSim [1] presented a meta path-based similarity measure for heterogeneous graphs.…”

“…PathSim [1] presented a meta path-based similarity measure for heterogeneous graphs. [2] proposed a meta pathbased ranking model to find entities with high similarity to a given query entity. HCC [3] is a meta-path based heterogeneous collective classification method.…”

“…Meta path-based social network analysis is gaining attention in recent years [1][2][3][4][5][6]. Existing efforts utilize a selection of meta paths between the same type of entities to improve the quality of similarity search, classification, clustering, link prediction and citation recommendation in heterogeneous networks.…”