基于协同过滤的景点推荐WebGIS平台设计与实现

发布时间：2018-05-09 19:06

本文选题：时空标签 + 协同过滤　；参考：《西安科技大学》2017年硕士论文

【摘要】：景点推荐服务平台在促进旅游业发展、推动地区经济增长、改善游客出游体验等方面发挥着不可或缺的作用。为了弥补当前主流旅游电子商务平台景点推荐功能缺失的不足以及改善个性化景点推荐应用缺乏的现状,本文以微博数据作为研究与应用的基础数据,以提出的自学习协同过滤算法与交集相似度计算方法作为景点推荐引擎构建的理论支撑,以WebGIS技术、数据库技术以及前端开发技术等作为平台设计实现的技术支持,通过构建时空标签数据模型与景点推荐模型,进行推荐算法的评测以及平台程序的编码与测试,完成了南京市景点推荐服务平台的设计与实现。具体研究内容与结果如下:(1)在时空标签数据模型构建中,从微博数据特征的角度阐述了采用微博数据作为研究与应用基础数据的可行性,并对微博数据的获取途径进行了说明;详细介绍了微博数据的聚合处理过程以及景点、游客、相似景点三个方面的时空标签数据模型。(2)在景点推荐模型构建中,为了改善协同过滤存在的数据稀疏和新用户问题,提出了基于文本分词与标签提取的自学习协同过滤算法;为了解决传统相似度度量方法只适用于量化数值的问题,提出了基于特征标签的交集相似度计算方法;然后对应于基于项目、用户以及自学习的协同过滤构建了各自的景点推荐模型。(3)在景点推荐算法评测中,分别介绍了评测数据、评测指标以及评测流程;通过对评测结果在准确率、召回率以及兴趣度方面的对比分析,得出在基于标签的景点推荐中,自学习的协同过滤明显优于基于项目的协同过滤和基于用户的协同过滤,良好的改善了数据稀疏和新用户问题。(4)在基于WebGIS的景点推荐服务平台设计与实现中,基于自学习的协同过滤算法和交集相似度计算方法构建了景点推荐引擎,采用GeoDataBase和MongoDB存储景点空间数据和属性数据,通过ArcGIS Server和WCF REST发布数据服务,调用ArcGIS API、jQuery类库等进行功能实现,利用Html、CSS、Javascript进行平台用户界面的布局与设计,完成了南京市景点推荐服务平台的设计与实现。
[Abstract]:Recommendation service platform plays an indispensable role in promoting tourism development, promoting regional economic growth and improving tourist experience. In order to make up for the deficiency of the recommendation function of the mainstream tourism e-commerce platform and to improve the status quo of the lack of personalized recommendation application, this paper takes Weibo data as the basic data for research and application. The self-learning collaborative filtering algorithm and the intersection similarity calculation method are used as the theoretical support for the construction of the recommendation engine of scenic spots, and the WebGIS technology, database technology and front-end development technology are used as the technical support for the platform design and implementation. By constructing spatio-temporal label data model and scenic spot recommendation model, evaluating the recommendation algorithm and coding and testing the platform program, the design and implementation of Nanjing Scenic spot recommendation Service platform are completed. The specific research contents and results are as follows: (1) in the construction of spatio-temporal tag data model, the feasibility of using Weibo data as the basic data for research and application is expounded from the point of view of Weibo's data characteristics. This paper introduces in detail the process of data aggregation and processing of Weibo's data, as well as the spatio-temporal label data model of scenic spots, tourists and similar scenic spots. In order to improve the problem of data sparsity and new users in collaborative filtering, this paper discusses the construction of recommendation model for scenic spots. A self-learning collaborative filtering algorithm based on text segmentation and label extraction is proposed, and in order to solve the problem that the traditional similarity measurement method is only applicable to quantization value, an intersection similarity calculation method based on feature labels is proposed. Then, corresponding to the project, user and self-learning collaborative filtering, we construct their recommendation model. In the evaluation of the recommendation algorithm, we introduce the evaluation data, the evaluation index and the evaluation process. Through the comparative analysis of the accuracy, recall and interest of the evaluation results, it is concluded that the self-learning collaborative filtering is better than the project-based collaborative filtering and user-based collaborative filtering in the tag-based recommendation of scenic spots. In the design and implementation of the recommendation service platform based on WebGIS, the recommendation engine is constructed based on self-learning collaborative filtering algorithm and intersection similarity calculation method. Using GeoDataBase and MongoDB to store spatial data and attribute data of scenic spots, publishing data services through ArcGIS Server and WCF REST, calling ArcGIS API jQuery class library, etc. The design and implementation of Nanjing Scenic spot recommendation Service platform are completed.
【学位授予单位】：西安科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P208

【参考文献】