Category-Aware API Clustering and Distributed Recommendation for Automatic Mashup Creation

摘要: Mashup has emeraged as a promising way to allow developers compose existed APIs (services) create new or value-added services. With the rapid increasing number of services published on Internet, service recommendation for automatic mashup creation gains lot momentum. Since inherently requires with different functions, result should contain from various categories. However, most existing approaches only rank all candidate in single list, which two deficiencies. First, ranking without considering categories they belong may lead meaningless and affect accuracy. Second, are not always clear about need cooperate better creation. Without explicitly recommending relevant creation, it remains difficult select proper mixed lower user friendliness recommendation. To overcome these deficiencies, novel category-aware clustering distributed method is proposed Kmeans variant ( vKmeans ) based topic model Latent Dirichlet Allocation introduced enhancing categorization providing basis top , category relevance SCRR model, combines machine learning collaborative filtering, developed decompose requirements predict Finally, CDSR framework, predicting order within each category. Experiments real-world dataset have proved that approach significant improvement at precision rate but also enhances diversity results.