Measuring Semantic Similarity between Services Using Hypergraphs
Hamdi Yahyaoui, Mohamed R. Al-Mulla, Eiman Boujarwah · 2021
We propose in this paper a new HyperGraph-based Similarity Algorithm (HGSA) to measure the semantic similarity between services with focus on Web services. HGSA relies on a term-similarity function that considers several parameters such as node’s local density, depth, number of relation factors, and link strength. This function is leveraged to build a similarity matrix between Web services. Computing the similarity between Web services is reduced to applying the Hungarian method to this matrix. We study the complexity of the proposed algorithm and provide experiments that show that it outperforms other related approaches in terms of accuracy.