Web APIs recommendation for Mashup development based on hierarchical dirichlet process and factorization machines

Abstract

Mashup technology, which allows software developers to compose existing Web APIs to create new or value-added composite RESTful Web services, has emerged as a promising software development method in a service-oriented environment. More and more service providers have published tremendous Web APIs on the internet, which makes it becoming a significant challenge to discover the most suitable Web APIs to construct user-desired Mashup application from these tremendous Web APIs. In this paper, we combine hierarchical dirichlet process and factorization machines to recommend Web APIs for Mashup development. This method, firstly use the hierarchical dirichlet process to derive the latent topics from the description document of Mashups and Web APIs. Then, it apply factorization machines train the topics obtained by the HDP for predicting the probability of Web APIs invocated by Mashups and recommending the high-quality Web APIs for Mashup development. Finally, we conduct a comprehensive evaluation to measure performance of our method. Compared with other existing recommendation approaches, experimental results show that our approach achieves a significant improvement in terms of MAE and RMSE.