旅游移动商务环境中基于情景的多维用户偏好模型及个性化推荐方法研究

发布时间：2018-08-21 07:42

【摘要】：旅游移动服务是一种情景依赖度较高的移动服务,用户在接受旅游产品的个性化推荐时,当前情景对用户偏好会产生一定程度的影响。关于旅游移动商务个性化推荐的研究已成为当前热点之一,目前旅游移动商务个性化推荐中基于情景的部分研究主要存在缺少维度权重和推荐结果类似的问题。虽然部分研究运用了情景要素对用户特征集进行扩展,却未充分考虑各情景要素本身对推荐结果及用户偏好的影响。还有部分研究仅使用时间、地点等物理环境维度的情景要素作为构建用户偏好模型的维度和产生推荐的依据,不同特征的用户在同一情景下获取的推荐结果相似,并未很好的实现个性化推荐。为了提高旅游移动商务的个性化与适应性程度,使用户可以通过旅游移动商务推荐系统更好地进行自我服务,本文重点针对以上两个问题开展了研究。本文从情景的角度出发,以景点推荐为例,对旅游移动商务环境中的个性化推荐方法展开研究,在分析旅游移动商务和个性化推荐方法研究现状的基础上,综合使用情景理论、个性化推荐方法,在当前情景、历史情景、用户历史行为记录和用户基本特征四个维度应用贝叶斯网络对用户偏好进行推理,构建了多维用户偏好模型。在此基础上,本文对已有个性化推荐方法进行了改进,通过实验验证证明了本文提出的旅游移动商务环境中的基于多维用户偏好模型的个性化推荐方法的推荐质量在一定程度上优于传统的个性化推荐算法。
[Abstract]:Tourism mobile service is a highly dependent mobile service. When users accept personalized recommendation of tourism products, the current situation will have a certain degree of impact on user preferences. The research on personalized recommendation of tourism mobile commerce has become one of the current hot spots. At present, there is a lack of dimensionality weight and similar recommendation results in the research of situation-based personalized recommendation of tourism mobile commerce. Although some of the studies use situational elements to extend the user feature set, they do not fully consider the impact of each situational element itself on the recommended results and user preferences. Some studies only use the scene elements of physical environment dimension such as time and place as the dimension of constructing user preference model and the basis of producing recommendation. The results obtained by users with different characteristics in the same situation are similar. Not very well to achieve personalized recommendation. In order to improve the degree of personalization and adaptability of tourism mobile commerce, and make users can better self-service through tourism mobile commerce recommendation system, this paper focuses on the above two problems. From the perspective of scene, this paper takes the recommendation of scenic spots as an example to study the personalized recommendation method in the tourism mobile commerce environment. On the basis of analyzing the present situation of the research on the tourism mobile commerce and the personalized recommendation method, this paper synthetically uses the scenario theory. In the four dimensions of current situation, historical situation, user history behavior record and user basic characteristics, Bayesian network is used to infer user preference, and a multi-dimensional user preference model is constructed. On this basis, the existing personalized recommendation methods have been improved. The experimental results show that the proposed personalized recommendation method based on multi-dimensional user preference model is superior to the traditional personalized recommendation algorithm to some extent.
【学位授予单位】：华东理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F274;F713.36;F590

【参考文献】