A Knowledge Graph Based Framework for Web API Recommendation

The rapid increase in the number of functionally similar web APIs, demands some web API recommender system (RS), to reduce the myriad of web APIs, during web API selection. The use of side information to resolve cold start and data sparsity limitations in RS, as a means of improving their accuracy, is gaining much popularity these days. Knowledge graphs (KGs) have shown to be very valuable source of side information, because they allow hybrid graph-based recommendation methods, comprising both collaborative and content information. While several knowledge graphs like DBpedia and Google knowledge graph exist, they do not include data for web API recommendation. In addition, constructing and leveraging such rich information for recommendation is challenging because it requires the ability to constructively encode different relations. In this paper, we present a knowledge graph framework, built from web API and mashups data, from www.programmableweb.com, that can be used as side information when recommending web APIs. Specifically, we first present a detailed knowledge graph schema, and then show how our knowledge graph can be used as side information to improve web API recommendation. We conjecture that using our proposed knowledge graph for web API recommendation not only improves recommendation accuracy, but also, diversity.