Tagging augmented neural topicmodel for semantic sparse Web service discovery

摘要

Search engine based Web service discovery model suffers from the semantic sparsity problemrndue to the fact that Web services are described in short texts, which in turn leads to poorrnrecall. To address this issue, external information that enriches the semantics of theWeb servicernand improves discovery performance has been highly concerned. In light of this, we propose arnnovel Web service discovery approach that uses the neural topic model, which seamlessly integratesrntagging information and word embedding for semantic sparsity problem. More specifically,rninstead of clusteringWeb services as done in most existing service discovery approaches, we usernword embedding to map the words as continuous embeddings to embody external semantics ofrnthe service description.We also leverage the neural topic model in service discovery, which takesrncontinuous word distribution as the input and interprets theWeb service description as a hierarchicalrnmodel. Based on the neural topic model and word embedding, we propose an efficientWebrnservice query and ranking approach. Experiments conducted on a real-worldWeb service datasetrndemonstrate the effectiveness of the proposed approach.