Research Article
Knowledge Graph Enhanced Web API Recommendation via Neighbor Information Propagation for Multi-service Application Development
@INPROCEEDINGS{10.1007/978-3-031-24383-7_2, author={Zhen Chen and Yujie Li and Yuying Wang and Xiaowei Liu and Yifan Xing and Linlin Liu and Dianlong You and Limin Shen}, title={Knowledge Graph Enhanced Web API Recommendation via Neighbor Information Propagation for Multi-service Application Development}, proceedings={Collaborative Computing: Networking, Applications and Worksharing. 18th EAI International Conference, CollaborateCom 2022, Hangzhou, China, October 15-16, 2022, Proceedings, Part I}, proceedings_a={COLLABORATECOM}, year={2023}, month={1}, keywords={Multi-service application development Web API recommendation Knowledge graph Neighbor information propagation}, doi={10.1007/978-3-031-24383-7_2} }- Zhen Chen
Yujie Li
Yuying Wang
Xiaowei Liu
Yifan Xing
Linlin Liu
Dianlong You
Limin Shen
Year: 2023
Knowledge Graph Enhanced Web API Recommendation via Neighbor Information Propagation for Multi-service Application Development
COLLABORATECOM
Springer
DOI: 10.1007/978-3-031-24383-7_2
Abstract
In cloud era, Web APIs have been the best carrier for service delivery, capability replication and data output in multi-service application development. Currently, the number of Web APIs on the Internet is huge and growing exponentially. To enable accurate and fast Web API selection for developers, researchers have proposed a variety of Web API recommendation methods. However, existing methods cannot solve the inherent data sparsity problem well. In addition, existing methods use context information indirectly by finding neighbors or discretely through embedding techniques, while rich semantic information in the Web API ecosystem is ignored. To solve the above problems, we firstly crawl and analyze Web API data to construct a Web API knowledge graph, which laid a data foundation for alleviating the data sparsity problem. Then, we propose a knowledge graph-enhanced Web API recommendation model, so as to improve recommendation accuracy by capturing high-order structural information and semantic information. Typically, multivariate representations of user and Web API are made by the neighbor information propagation in Web API knowledge graph. The proposed model supports end-to-end learning for beneficial feature extraction. Finally, experiments results demonstrate the proposed model outperforms baselines significantly, thereby promoting the development of Web API economy.
