植入引导式层次聚类Mashup服务推荐算法  

Mashup service recommendation algorithm based on implantable guided hierarchical clustering algorithm

作　　者：张拥华[1] 李原[1] 曾凡仔[2] 周朕[1] 

机构地区：[1]湖南工业职业技术学院,长沙410208 [2]湖南大学信息科学与工程学院,长沙410012

出　　处：《计算机应用研究》2016年第2期424-427,共4页Application Research of Computers

基　　金：湖南省教育厅科学研究课题(12C1032)

摘　　要：针对传统Mashup服务推荐算法在关键字聚合搜索和网络构建等方式中计算复杂度过高的问题,提出一种基于语义标签的植入引导式层次聚类Mashup服务推荐算法。首先,为提高聚类算法的收敛精度,提高算法运行效率来满足大型数据搜索对算法简化的需求,采用数据预处理和植入易于获取具有代表性的样本数据对聚类进行引导,防止层次聚类算法顶层集分类失败导致的算法聚类失败;其次,利用改进的聚类算法结合实际的Mashup服务数据库,设计了植入引导式层次聚类Mashup服务推荐算法。最后,通过仿真对比表明,基于语义的植入式半监督层次聚类Mashup服务推荐算法的精度要好于对比算法,验证了所提算法的有效性。In order to solve the problem of high complexity in keyword searching and network construction, this paper de- signed a semantic label based implanted guide hierarchical clustering Mashup service recommendation algorithm. Firstly, in order to improve the convergence precision and the operating efficiency of clustering algorithm, which met the demand of the algorithm simplify in the large data searching, it used the data preprocessing and the representative sample to guide the cluste- ring, which would prevent the hierarchical clustering algorithm failure. Secondly, it used the improved clustering algorithm combined with the actual Mashup service database to design the implanted guide hierarchical clustering based Mashup service recommendation algorithm. Finally, through the simulation contrast show that the Mashup service recommendation algorithm based on implantable guided hierarchical clustering is better than the comparison algorithm in accuracy, which validates the ef- fective of the proposed algorithm.

关 键 词：引导 层次聚类 MASHUP 标签 服务推荐 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] TP301.6[自动化与计算机技术—控制科学与工程]