摘要
针对传统Mashup服务推荐在网络构建方式的成本和计算复杂性过高问题,提出一种基于半监督层次聚类描述的Mashup服务推荐算法.首先,利用网爬工具收集ProgrammableWeb上的Mashup服务信息,并采用后缀剥离算法把Mashup服务的标签信息修改为名词形式,以此作为研究分析的数据集;其次,为提高聚类精度,提出一种半监督层次描述聚类算法,通过植入层次聚类算法顶层核心集方式,有效解决了传统层次描述聚类因顶层分类集构造失败而影响Mashup服务推荐算法的准确度.通过在聚类数据集和网爬Mashup服务数据库上的实验表明,该算法的Mashup服务推荐准确率优于对比算法.
According to the problem of high computational complexity and cost in network recommended for traditional Mashup service recommendation,the authors proposed the semisupervised ISHC hierarchy description based Mashup service clustering alogrithm.Firstly,the climbing tools were used to collect ProgrammableWeb Mashup service information,and the suffix stripping algorithm was used to modify the Mashup service label with noun form,which was used as the research and analysis data sets.Secondly,in order to improve the accuracy of clustering algorithm,an implantable semi supervised hierarchical clustering algorithm was proposed,to effectively prevent the hierarchical clustering algorithm topping set classification from failure to influence the Mashup service recommendation algorithm accuracy by implantating the core set approach of hierarchical clustering algorithm.The simulation results show that the recommendation accuracy of implantable semi supervised hierarchical clustering algorithm is better than that of comparison algorithm in Mashup services recommendation,which verifys the effectiveness of the proposed algorithm.
出处
《吉林大学学报（理学版）》
CAS
CSCD
北大核心
2015年第4期698-704,共7页
Journal of Jilin University:Science Edition
基金
宁夏回族自治区自然科学基金(批准号:NZ13048)
天津市高等学校人文社会科学研究项目(批准号:20102523)
