基于迭代训练的Web Service混合协同过滤推荐模型

发布时间：2018-07-05 07:27

本文选题：Web服务推荐 + QoS　；参考：《计算机研究与发展》2013年S2期

【摘要】：伴随着互联网技术的日益发展,海量数据的集成融合促进了大数据技术的广泛应用,尤其以面向服务为核心的Web Service技术被普遍用来提供新型互联网服务,这使得针对服务提供商及个人用户设计一种基于Web Service的个性化服务推荐系统变得十分必要.因此,提出一种基于混合协同过滤技术进行服务质量(QoS)预测的服务推荐模型.该模型利用迭代训练的思想,不断提升服务质量预测值的准确率,并通过基于预测树(PTree)的性能优化策略,有效地降低了迭代过程的运行时间.基于一个包含150万条Web Service调用信息的数据集,开展了一系列的对比分析实验.实验结果表明,相比于其他一些推荐模型,所提出的基于迭代训练的混合协同过滤推荐模型在消耗同等资源的情况下,能够有效地降低预测值的误差,提升模型整体的预测准确率.
[Abstract]:With the development of Internet technology, the integration of mass data promotes the wide application of big data technology, especially the service-oriented web service technology is widely used to provide new Internet services. This makes it necessary to design a personalized service recommendation system based on Web Service for service providers and individual users. Therefore, a service recommendation model based on hybrid collaborative filtering for QoS prediction is proposed. The model uses the idea of iterative training to improve the accuracy of quality of service (QoS) prediction value, and the performance optimization strategy based on prediction tree (PTree) can effectively reduce the running time of iterative process. Based on a data set of 1.5 million Web Service calls, a series of comparative analysis experiments are carried out. The experimental results show that the proposed hybrid collaborative filtering recommendation model based on iterative training can effectively reduce the error of prediction value under the same resource consumption compared with other recommendation models. Improve the prediction accuracy of the model as a whole.
【作者单位】：计算机软件新技术国家重点实验室(南京大学);南京大学计算机科学与技术系;
【基金】：江苏省自然科学基金项目(BK2010373) 江苏省普通高校研究生科研创新计划资助项目(CXZZ11_0045) 计算机软件新技术国家重点实验室自主课题项目(ZZKT2013B11)
【分类号】：TP391.3;TP393.09

【参考文献】