A Feature-Enhanced Service Function Vector Generation Method for Web API Clustering

The brevity of service function descriptions often results in feature sparsity and incompleteness of service function vectors generated by topic models. To improve the quality of service function vectors generated by topic modes, this paper proposes an enhanced service function vector generation method suitable for Web API clustering. We design a context-based feature word extraction method for service description text and construct a topic probability correction factor. The topic features of service function vectors are enhanced by applying feature word extraction and probability correction. Experiment results show that the service function vector with the enhanced topic can significantly improve the service clustering quality.