Using semantic similarity model to improve OGC web services matching accuracy

Abstract

In order to seek out available service from massive geographic information services with high accuracy and satisfy the needs of service users, this paper takes advantage of the characteristics of metadata documents of OGC geographic information services, extracting many description vocabularies from metadata description files according to different tags' weight. Then, a semantic similarity measurement model is proposed to figure out the semantic similarity between the search keyword and OGC Web Service described by semantic distance similarity, semantic structure similarity and attribute information similarity. Using the above model, several experiments are implemented. The results show that: 1. the computing mode of the integrated semantic similarity proposed in this paper is in line with the public cognition and the expert evaluation. 2. comparing with the keyword-based search method, the semantic similarity model could increase recall ratio over 40%, up to 90%; and increase precision over 6%, up to 90%.