Abstract:The assessment of semantic similarity between lexical terms plays a critical part in semantic-oriented applications for natural language processing and cognitive science. The optimization of calculation models is still a challenging issue for improving the performance of similarity measurement. In this paper, we investigate WordNet-based measures including distance-based, information-based, feature-based and hybrid. Among them, the distance-based measures are considered to have the lowest computational complexity due to simple distance calculation. However, most of existing works ignore the meronymy relation between concepts and the non-uniformity of path distances caused by various semantic relations, in which path distances are simply determined by conceptual hyponymy relation. To solve this problem, we propose a novel model to calculate the path distance between concepts, and also propose a similarity measure which nonlinearly transforms the distance to semantic similarity. In the proposed model, we assign different weights in accordance with various relations to edges that link different concepts. On basis of the distance model, we use five structure properties of WordNet for similarity measurement, which consist of multiple meanings, multiple inheritance, link type, depth and local density. Our similarity measure is compared against state-of-the-art WordNet-based measures on M&C dataset, R&G dataset and WS-353 dataset. According to experiment results, the proposed measure in this work outperforms others in terms of both Pearson and Spearman correlation coefficients, which indicates the effectiveness of our distance model. Besides, we construct six additional benchmarks to prove that the proposed measure maintains stable performance.

A Semantic-Rich Similarity Measure in Heterogeneous Information Networks.

A Semantic-Rich Similarity in Heterogeneous Information Networks

Recurrent Meta-Structure for Robust Similarity Measure in Heterogeneous Information Networks

DMSS: A Robust Deep Meta Structure Based Similarity Measure in Heterogeneous Information Networks

A Meta Graph-Based Top-k Similarity Measure for Heterogeneous Information Networks

HN-Sim: A Structural Similarity Measure over Object-Behavior Networks.

A Hybrid Semantic Similarity Measurement for Geospatial Entities

Howsim: A General And Effective Similarity Measure On Heterogeneous Information Networks

Effective Similarity Search on Heterogeneous Networks: A Meta-path Free Approach

Measuring Distance-Based Semantic Similarity Using Meronymy and Hyponymy Relations

Relevance Measure In Large-Scale Heterogeneous Networks

Semantic Enhanced Top-k Similarity Search on Heterogeneous Information Networks

A New Hybrid Semantic Similarity Measure Based on WordNet

HBGSim: A structural similarity measurement over heterogeneous big graphs

Measuring Semantic-Based Structural Similarity in Multi-Relational Networks

Distant Meta-Path Similarities for Text-Based Heterogeneous Information Networks

A WordNet-based hybrid semantic similarity measurement

On Efficient Network Similarity Measures.

HeteRank: A general similarity measure in heterogeneous information networks by integrating multi-type relationships.

A Semantic Path-Based Similarity Measure for Weighted Heterogeneous Information Networks.

E-rank: A Structural-Based Similarity Measure in Social Networks.