Web service filtering and visualization with context aware similarity to bootstrap clustering

Abstract

Web service clustering is an efficient approach to address some challenges in service computing area such as discovering and recommending. To cluster the Web services, we need to filter the similar services. Key operation of filtering process is measuring the similarity of services. There are several methods used in current similarity calculation approaches such as keyword, information retrieval, ontology and hybrid methods. However, these approaches do not consider the context when measuring the similarity. So these approaches failed to capture the semantic of terms, which exist under a certain domain. In this paper, we propose context aware similarity method, which uses search results from search engines and support vector machine. Then, we apply Associated Keyword Space (ASKS) algorithm which is effective for noisy data and projected results from a three-dimensional (3D) sphere to a two dimensional (2D) spherical surface for 2D visualization to filter the services. Experimental results show our filtering approach is able to filter services based on domain and plot the result on sphere. Also our approach performs better than the existing approaches. Further, our approach aids to search Web services by visualization of the service data on a spherical surface.