基于主题模型的Mashup标签推荐方法

发布时间：2018-06-08 05:34

本文选题：Mashup + Web　；参考：《计算机学报》2017年02期

【摘要】：Web 2.0时代,标签作为Web资源管理和检索的有效方式已成为近年的热点研究对象.开发者通常为新的Mashup人工指定若干与功能性相关的标签,以便于用户理解、检索以及实现Mashup资源的分类管理.然而,手动指定标签十分繁琐且费时,自动生成Mashup标签十分必要但缺乏有效方法.针对该问题,文中提出一种基于主题模型的方法进行Mashup标签的自动推荐.该方法首先建立Mashups与Web Application Programming Interfaces(APIs)的描述文档以及Mashups与APIs之间的组合关系模型,然后寻找与待推荐标签Mashup的描述文档主题分布相似的Web APIs,并将它们与该Mashup直接组合的APIs合并,采用一种带权重的PageRank算法,从中挑选出最重要的APIs,最后将它们已有标签推荐给该Mashup.同时,针对所提方法文中设计实现一种标签排序算法,该算法优先推荐那些与Mashup主题最相关的标签.根据使用从ProgrammableWeb收集的真实数据进行实验可知,文中所提出的方法明显优于其他自动化标签推荐方法.
[Abstract]:In the era of Web 2.0, tags as an effective way to manage and retrieve Web resources have become a hot research object in recent years. Developers often manually specify several functional tags for new mashups to facilitate user understanding, retrieval and classification management of mashup resources. However, manually specifying tags is cumbersome and time-consuming. It is necessary to generate mashup tags automatically, but there is no effective method. To solve this problem, a method based on topic model is proposed to automatically recommend mashup tags. This method first establishes the description document of Mashups and Application programming interfaces (API) and the combination relation model between mashups and APIs, then looks for the web API Isa which is similar to the theme distribution of the description document of mashups to be recommended, and combines them with the APIs directly combined with the mashups. A weighted PageRank algorithm is used to select the most important API from which the existing tags are recommended to the Mashup. At the same time, a label sorting algorithm is designed and implemented for the proposed method, which gives priority to those tags that are most relevant to the Mashup topic. According to the experimental results from real data collected from Programmable Web, the proposed method is obviously superior to other automatic label recommendation methods.
【作者单位】：湖南科技大学知识处理与网络化制造湖南省普通高校重点实验室;
【基金】：国家自然科学基金(61572187,61300129,61272063,61572186) 国家科技支撑计划(2015BAF32B01) 教育部留学回国人员科研启动基金(教外司留[2013]1792) 湖南省教育厅资助科研项目(16K030) 湖南省研究生科研创新项目(CX2016B573)资助~~
【分类号】：TP391.3

【相似文献】