面向O2O用户行为分析的个性化推荐算法研究与应用

发布时间：2018-05-03 16:57

本文选题：相似性 + 用户行为分析　；参考：《河北大学》2017年硕士论文

【摘要】：近些年,随着互联网电子商务的发展,用户在电商平台上使用并产生的数据剧增,使用户面临严重的信息过载问题。个性化推荐系统是解决信息过载的有效方式之一。采用协同过滤推荐算法和基于内容的推荐算法来构建个性化推荐系统,并将两次推荐结果混合生成最终推荐结果。在基于用户的协同过滤推荐算法中,对比不同相似度度量方法,并采用修正的余弦相似性公式来计算用户之间的相似度。在基于内容的推荐算法中,采用余弦相似定理计算兴趣特征向量和项目特征向量之间的相似度。基于内容和协同过滤混合的推荐算法综合考虑了用户个人行为因素和邻近用户组因素,提高了推荐的有效性。其中,项目之间的相似度通过计算项目结构化数据之间的相似度得到。在研究面向O2O的用户行为分析的基础上,提出了面向O2O用户行为分析的混合推荐算法。以基于内容和协同过滤混合的推荐算法为基础,设计了面向O2O用户行为分析的混合推荐模块架构图,并通过理论与实际的结合验证了该混合推荐算法在基于O2O的生活信息服务系统上应用的有效性。通过调研与用户行为分析相结合的方式,归纳总结了在基于O2O的生活信息服务系统中对用户行为分析的研究状况,并展望了该领域的发展趋势和技术增长点。
[Abstract]:In recent years, with the development of electronic commerce on the Internet, the number of data generated by users on e-commerce platforms has increased dramatically, which makes users face serious problems of information overload. Personalized recommendation system is one of the effective ways to solve information overload. The collaborative filtering recommendation algorithm and the content-based recommendation algorithm are used to construct the personalized recommendation system, and the final recommendation results are generated by mixing the two recommendation results. In the user-based collaborative filtering recommendation algorithm, different similarity measures are compared, and the modified cosine similarity formula is used to calculate the similarity between users. In the content-based recommendation algorithm, the similarity between the feature vector of interest and the feature vector of item is calculated by using the cosine similarity theorem. The recommendation algorithm based on the mixture of content and collaborative filtering takes into account the factors of user's personal behavior and neighboring user groups and improves the effectiveness of recommendation. Among them, the similarity between items is obtained by calculating the similarity between structured data of items. Based on the research of O2O-oriented user behavior analysis, a hybrid recommendation algorithm for O2O-oriented user behavior analysis is proposed. Based on the hybrid recommendation algorithm based on content and collaborative filtering, the architecture diagram of hybrid recommendation module for O2O user behavior analysis is designed. The application of the hybrid recommendation algorithm in O2O based life information service system is verified by combining theory and practice. Through the combination of research and user behavior analysis, this paper summarizes the research status of user behavior analysis in the life information service system based on O2O, and looks forward to the development trend and technical growth point of this field.
【学位授予单位】：河北大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.3

【参考文献】