基于隐式反馈数据的在线旅游推荐研究

发布时间：2018-03-19 13:09

本文选题：在线旅游行业　切入点：推荐系统　出处：《浙江工商大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着互联网O2O的迅猛发展,在线旅游行业出现了信息过载的问题,面对种类和数量越来越多的旅游产品,在线旅游的用户也难以快速高效的找到自己需要的产品。个性化推荐系统可以解决这个难题,借助用户的历史兴趣帮助用户在纷繁复杂的信息中找到其喜欢的物品。推荐系统多应用于电子商务、电影和视频、音乐等领域,对在线旅游行业推荐系统的研究甚少。因此,本文将结合在线旅游网站的行业背景,将逻辑回归模型和协同过滤算法进行组合,构建推荐模型。模型主要利用在线旅游网站用户的隐式反馈数据,首先采用逻辑回归模型对用户的隐式反馈数据进行建模,将用户的浏览行为转化成用户对物品的兴趣度,再结合在线旅游的行业特点使用协同过滤推荐算法对用户进行推荐。此模型利用在线旅游网站的隐式反馈数据,弥补了显示反馈数据不足的弊端,充分挖掘用户的浏览行为,获取用户的兴趣度,实验证明此模型可以有效获取用户偏好,产生高质量的推荐。
[Abstract]:With the rapid development of the Internet O2O, the online tourism industry has appeared the problem of information overload. In the face of more and more kinds and quantity of tourism products, Users of online travel also find it difficult to find the product they need quickly and efficiently. Personalized recommendation systems can solve this problem. With the help of users' historical interest, users can find their favorite items in the complicated information. Recommendation systems are mostly used in e-commerce, film and video, music and other fields, so there is little research on online travel industry recommendation system. In this paper, we combine the logical regression model with collaborative filtering algorithm to construct the recommendation model, which mainly uses the implicit feedback data of the users of online tourism website, and combines the industry background of online tourism website with logical regression model and collaborative filtering algorithm. Firstly, the user's implicit feedback data is modeled by the logical regression model, and the browsing behavior of the user is transformed into the user's interest in the items. Combined with the characteristics of online tourism industry, collaborative filtering recommendation algorithm is used to recommend users. This model makes use of the implicit feedback data of online tourism website to make up for the deficiency of display feedback data, and fully excavates the browsing behavior of users. The experiments show that the model can obtain user preference effectively and produce high quality recommendation.
【学位授予单位】：浙江工商大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F724.6;F274;F592

【参考文献】