基于智能感知分析用户行为的个性化推荐系统研究

发布时间：2018-10-05 17:37

【摘要】：近年来,计算机科学技术的迅猛发展,极大地推动了互联网技术在人们生活中的普及率。人们在日常生活中正在不断地通过互联网来进行信息的获取、查询和各种网络服务。然而,在互联网技术飞速发展的过程中,在Internet快速发展的同时,网络安全问题成为了阻碍其发展的重要因素,安全问题还给电子商务技术的发展带来了巨大的挑战。随着电子商务技术的发展,越来越多的企业开始深入分析网络用户的行为,通过分析用户的行为来精确地掌握用户的兴趣爱好,从而根据他们的兴趣爱好来完成个性化商品推荐服务,在这个过程中可以充分提高用户的交易成功率,同时提高企业的盈利水平。很多政府和非营利性网站也意识到分析用户行为对网络维护和发展的重要性,所以越来越多的科研工作者开始投入对网络用户行为的研究,对网络用户行为进行精确分析越来越重要。很多电子商务网站都是通过获取用户的点击行为来进行用户行为数据的收集的。随着互联网技术的不断发展,越来越多的广告充斥在网络上。所有用户在使用互联网的过程中,自身的浏览记录都会被网络自动地进行识别和记录,这也就给用户行为分析提供了非常准确的行为基础,在研究用户行为上,对数据的采集和网站数据提取是非常重要的。本课题的研究是基于智能感知技术来精确分析用户行为,在文中对网络用户行为的概念和相关的技术核心都进行了分析,同时根据用户行为的不同进行了分类,对用户行为进行分类可以有效提高用户行为分析的准确度。通过对用户行为进行建模来了解模型的输入和输出,从而提高对用户行为分析系统模型的认识。同时,在使用基于协同过滤算法完成内容分析的基础上,在用户行为分析中加入了度量项,采用均方误差来提高计算的精准度。通过系统测试可以发现,使用加权组合算法可以有效减小算法的时间复杂度,有效弥补了协同过滤算法计算上存在的冷启动问题,大大提高了个性化推荐系统的精准度。
[Abstract]:In recent years, the rapid development of computer science and technology has greatly promoted the popularity of Internet technology in people's lives. In daily life, people are constantly using the Internet to access information, query and various network services. However, in the process of rapid development of Internet technology, with the rapid development of Internet, network security has become an important factor hindering its development. Security issues have also brought great challenges to the development of e-commerce technology. With the development of electronic commerce technology, more and more enterprises begin to analyze the behavior of network users in depth, by analyzing the behavior of users to accurately grasp the interests and hobbies of users. Thus, according to their interests and hobbies to complete personalized commodity recommendation service, in this process can fully improve the transaction success rate of users, while improving the profitability of enterprises. Many government and non-profit websites are also aware of the importance of analyzing user behavior to the maintenance and development of the network, so more and more researchers are beginning to study the behavior of Internet users. It is more and more important to accurately analyze the behavior of network users. Many e-commerce websites collect user behavior data by obtaining user click behavior. With the continuous development of Internet technology, more and more advertisements fill the network. In the process of using the Internet, all users will automatically identify and record their browsing records, which provides a very accurate behavioral basis for user behavior analysis. It is very important to collect data and extract website data. The research of this topic is based on intelligent perception technology to accurately analyze user behavior. In this paper, the concept of network user behavior and related technical core are analyzed, at the same time, according to the different user behavior are classified. Classification of user behavior can effectively improve the accuracy of user behavior analysis. In order to improve the understanding of the user behavior analysis system model, the input and output of the model can be understood by modeling the user behavior. At the same time, based on the collaborative filtering algorithm to complete the content analysis, a metric is added to the user behavior analysis, and the mean square error is used to improve the accuracy of the calculation. Through the system test, it can be found that the weighted combination algorithm can effectively reduce the time complexity of the algorithm, effectively make up for the cold start problem in the calculation of collaborative filtering algorithm, and greatly improve the accuracy of personalized recommendation system.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.3

【相似文献】