作者: Buqing Cao , Min Shi , Xiaoqing Liu , Jianxun Liu , Mingdong Tang
DOI: 10.1007/978-3-319-49178-3_30
关键词: World Wide Web 、 Web API 、 Topic model 、 Measure (data warehouse) 、 Factorization 、 Exploit 、 Mashup 、 Service discovery 、 Computer science 、 Dimension (data warehouse)
摘要: The rapid growth in the number of Web APIs, coupled with myriad functionally similar makes it difficult to find suitable APIs develop Mashup applications. Even if existing recommendation methods show improvements service discovery, accuracy them can be significantly improved due overlooking impact sparsity and dimension relationships between on accuracy. In this paper, we propose a method for creation by combining relational topic model factorization machines technique. This firstly uses characterize among Mashup, their links, mine latent topics derived relationships. Secondly, exploits train predicting link relationship recommend adequate relevant top-k target creation. Finally, conduct comprehensive evaluation measure performance our method. Compared other approaches, experimental results that approach achieves significant improvement terms precision, recall, F-measure.