Correlated Topic Model for Web Services Ranking

Abstract

With the increasing number of published Web servicesproviding similar functionalities, it’s very tedious for aservice consumer to make decision to select the appropriate oneaccording to her/his needs. In this paper, we explore severalprobabilistic topic models: Probabilistic Latent Semantic Analysis(PLSA), Latent Dirichlet Allocation (LDA) and CorrelatedTopic Model (CTM) to extract latent factors from web servicedescriptions. In our approach, topic models are used as efficientdimension reduction techniques, which are able to capture semanticrelationships between word-topic and topic-service interpretedin terms of probability distributions. To address the limitation ofkeywords-based queries, we represent web service description asa vector space and we introduce a new approach for discoveringand ranking web services using latent factors. In our experiment,we evaluated our Service Discovery and Ranking approachby calculating the precision (P@n) and normalized discountedcumulative gain (NDCGn).