作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程

• 先进计算与数据处理 • 上一篇    下一篇

基于用户使用历史与信誉评价的Web API 推荐

曹步清1,2,刘建勋1,唐明董1,谢芬方1   

  1. (1. 湖南科技大学计算机科学与工程学院,湖南湘潭411201;2. 武汉大学软件工程国家重点实验室,武汉430072)
  • 收稿日期:2014-06-12 出版日期:2015-06-15 发布日期:2015-06-15
  • 作者简介:曹步清(1979 - ),男,博士,主研方向:服务计算,云计算;刘建勋,教授、博士、博士生导师;唐明董,副教授、博士;谢芬方,硕士 研究生。
  • 基金资助:

    国家自然科学基金资助项目(61402168,61272063);武汉大学软件工程国家重点实验室开放基金资助项目(SKLSE2014-10-10)。

Web API Recommendation Based on User Usage History and Reputation Evaluation

CAO Buqing 1,2,LIU Jianxun 1,TANG Mingdong 1,XIE Fenfan 1   

  1. (1. School of Computer Science and Engineering,Hunan University of Science and Technology,Xiangtan 411201,China;2. State Key Laboratory of Software Engineering,Wuhan University,Wuhan 430072,China)
  • Received:2014-06-12 Online:2015-06-15 Published:2015-06-15

摘要:

随着网络上发布的Web API 服务越来越多,如何推荐给开发者用户感兴趣、信誉度高的Web API 服务,以构建高质量高可信的软件服务系统,成为一个具有挑战性的研究问题。为此,提出一种基于用户使用历史与信誉评价的Web API 服务推荐方法。计算用户使用历史记录与Web API 之间的相似度,获得Web API 的用户兴趣值。综合用户的Web API 评分,调用Web API 的Mashup 服务的评价贡献和Alexa 统计的Web API 访问流量,获得Web API 的信誉评价值。根据Web API 的用户兴趣值以及信誉评价值,实现Web API 的排名与推荐。实验结果表明, 该方法推荐的Web API 用户兴趣度DCG 值高于SR-Based 方法,服务信誉度DCG 值高于UI-Based 方法。

关键词: Web API 服务, 用户使用历史, 用户兴趣度, 信誉评价, 服务信誉度, Web API 推荐

Abstract:

With the release of more and more Web API services on Internet,it becomes a challenging research problem that how to recommend Web APIs that developer user are interested in and reputation degrees are high,to construct high quality and trustworthy software service system. This paper presents Web API service recommendation approach based on user usage history and reputation evaluation (WASR). It computes the similarity between user history records and Web API services,and gets user interest degree. Service reputation degree is computed by considering the user score of Web API,the score contributions of those Mashup services calling the Web API,and traffic flow of Web API based on statistical data by Alexa. It ranks and recommends Web API services according to the user interest degree and service reputation degree of Web APIs. Experimental results show that this approach can recommend Web API services with higher DCG of user interest degree than those of SR-based approach,and higher DCG of service reputation degree than those of UI-based approach.

Key words: Web API service, user usage history, user interest degree, reputation evaluation, service reputation degree, Web API recommendation

中图分类号: