基于效用的Web个性化服务模型

发布时间：2019-02-17 08:10

【摘要】：现代电子信息技术的快速发展使得互联网信息呈爆炸方式增长。然而,信息的急速增长却未能给用户提供方便,海量数据给用户获取知识带来了极大的障碍。这种障碍使得用户在信息的海洋中却无法找到期望有效知识。为解决此类问题,Web个性化服务技术应运而生。Web个性化服务是将Web技术与数据挖掘技术相结合以提高Web站点服务质量为目的的一种服务模式。个性化服务是一种“信息找人”的服务模式。将Web挖掘技术应用于个性化服务系统之中,结合Web文本挖掘、Web领域本体等技术进一步提高Web站点系统的服务质量。近年来,在国内外学者的共同努力之下Web个性化服务技术取得了一系列重大科研成果并形成一套经典的个性化服务模型。然而,现阶段个性化服务理论体系仍有许多需要完善的地方,比如系统“冷启动”问题、如何高效准确更新用户兴趣模型的问题、个性化推荐算法研究问题等。为解决以上问题,本文在传统个性化服务模型的基础之上,提出了一种基于效用的Web个性化服务模型。本模型引入了效用理论,在其基础之上提出了一种基于效用的用户兴趣模型更新算法,以探求一种高质量的用户兴趣模型根更新算法。对于传统个性化服务模型系统中普遍存在的“冷启动”问题,本文通过引入三方平台登录模块,以三方平台上丰富的用户网络资源来构建用户初始兴趣模型,使用户可以快速得到高质量个性化服务。并且针对经典个性化推荐算法-K-means算法的不足,本文提出了一种基于协同聚类的用户-兴趣项双聚类算法。通过模拟实验表明,本模型能够提供较高质量的个性化服务,可以满足Web站点的个性化服务建设需求,具有较高的理论研究价值与现实意义。
[Abstract]:The rapid development of modern electronic information technology makes the Internet information explosive growth. However, the rapid growth of information can not provide convenience to users. This barrier prevents users from finding the desired knowledge in the ocean of information. In order to solve this kind of problem, Web personalization service technology emerges as the times require. Web personalization service is a kind of service mode which combines Web technology and data mining technology to improve the service quality of Web site. Personalized service is a kind of service mode of "information seeking person". The Web mining technology is applied to the personalized service system. Combined with the Web text mining and Web domain ontology technology, the service quality of the Web site system is further improved. In recent years, with the joint efforts of scholars at home and abroad, Web personalized service technology has made a series of important scientific research results and formed a set of classic personalized service model. However, the theoretical system of personalized service still needs to be improved, such as "cold start" problem, how to update user interest model efficiently and accurately, and how to study personalized recommendation algorithm. In order to solve the above problems, this paper proposes a Web personalized service model based on utility based on the traditional personalized service model. The utility theory is introduced in this model, and a Utility based updating algorithm of user interest model is proposed to explore a high quality root updating algorithm of user interest model. For the common "cold start" problem in the traditional personalized service model system, this paper constructs the initial user interest model by introducing the three-party platform login module and using abundant user network resources on the tripartite platform. Enables the user to obtain the high quality personalization service quickly. Aiming at the shortcomings of the classical personalized recommendation algorithm-K-means algorithm, this paper proposes a user-interest double clustering algorithm based on cooperative clustering. The simulation results show that the model can provide high quality personalized service and meet the needs of Web site. It has high theoretical research value and practical significance.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09;TP391.1

【参考文献】