Keywords‐driven web APIs group recommendation for automatic app service creation process. (27th September 2020)

• Record Type: Journal Article
• Authors: Gong, Wenwen; Lv, Chao; Duan, Yucong; Liu, Zengguang; Khosravi, Mohammad R.; Qi, Lianyong; Dou, Wanchun
• Other Names: Zhou Junlong guestEditor.; Kritikakou Angeliki guestEditor.; Zhu Dakai guestEditor.; Lastra Jose L. Martinez guestEditor.; Hu Shiyan guestEditor.
• Is Part Of: Software, practice & experience. Volume 51:Number 11(2021)
• Page Start: 2337
• Page End: 2354
• Publication Date: 2020-09-27
• Subjects: APIs group recommendation -- app creation -- diverse composition -- query keywords; Computer software -- Periodicals; Computer programming -- Periodicals; Computer programs -- Periodicals; 005.3
• DOI: 10.1002/spe.2902
• Languages: English
• ISSNs: 0038-0644

Abstract

Summary: With the ever‐increasing popularity of web application programming interfaces (APIs) sharing communities, it is becoming a promising way for software developers to design and create their interesting Apps through composing a set of selected web APIs that can collectively fulfill the App functions expected by the App developer. However, the App developer's web APIs selection decision‐makings are often nontrivial due to the massive candidate APIs as well as their diverse functions. Furthermore, it is difficult to guarantee that the selected web APIs are compatible enough. Moreover, traditional web APIs recommendation approaches only return a recommended APIs list, which are often not sufficient to accommodate the App developer's undetermined and fuzzy personalized preferences. Considering the above challenges, a novel keywords‐driven web APIs recommendation approach called keywords‐driven and compatibility‐aware multiple API group recommendation is proposed in this article for green and compatible software, which cannot only satisfy the App developer's functional requirements, but also return a group of web APIs recommended lists. Each returned list includes a set of compatible web APIs. Finally, we design a series of experiments based on a real‐world web APIs dataset, that is, PW dataset crawled from www.programmableWeb.com . Experimental reports compared with other competitive approaches in existing literatures indicate the effectiveness and efficiency of our proposal in this work.