基于电子商务平台的保险推荐系统的设计与实现

发布时间：2018-03-30 01:08

本文选题：电子商务　切入点：保险推荐　出处：《复旦大学》2013年硕士论文

【摘要】：随着信息技术和互联网的高速发展,电子商务也随之迅速发展,人类逐渐从信息匮乏的时代渐渐走向了大数据的时代,信息繁琐和找不到人们需要的信息成为现在电子商务发展的瓶颈。在这个信息技术的新时代,无论对于信息的创造者还是信息的阅读者,都被淹没在大量的选择之中,信息创造者需要定向投递广告,信息阅读者需要找到符合自身愿望的产品,推荐系统的出现就是为了解决这一矛盾的重要工具。推荐系统,顾名思义就是为电子商务的用户来推荐其感兴趣的产品。它根据用户个人的属性,个人的喜欢,购买习惯来向其推荐相应的产品。本文的推荐是基于保险的,目前全世界对于保险的推荐还处于一个空白的阶段,如果只是套用现有的推荐算法,当用户属性不是很明确或者购买浏览行为不够多的情况下,比较难以对其进行个性化的推荐,推荐的效果就完全失去了意义。本文首先调研了国内外现有主流电子商务网站的推荐系统,在此基础上分析了X1保险电子商务网站的现状,对开发保险个性化推荐系统进行了比较详尽的需求分析,然后设计出了一个保险个性化推荐系统的系统框架,该系统框架包括推荐系统的实施层、推荐系统的引擎层、推荐系统的数据／知识层三个模块。其次,在推荐系统的详细设计阶段,量化了客户属性和产品属性,根据各个推荐算法的需要,设计出了数据库的模式,包括推荐算法需要的原始数据表和推荐结果表,重点设计了适合于各个场景和各个客户类型的个性化和非个性化的推荐算法,包括基于统计的推荐算法、基于关联规则的推荐算法、基于用户协同过滤的推荐算法、基于条目的协同过滤推荐算法和基于内容的推荐算法。然后为了验证推荐算法的正确性,对各个推荐算法进行了测试。最后,设计出了一个系统部署技术方案,将推荐系统和X保险电子商务网站进行了对接。
[Abstract]:With the rapid development of information technology and the Internet, electronic commerce has also developed rapidly. Human beings have gradually moved from the era of information scarcity to the era of big data. Information tedious and unable to find the information people need become the bottleneck of the development of electronic commerce. In this new era of information technology, both the creator of information and the reader of information are submerged in a large number of choices. Information creators need to deliver advertisements, information readers need to find products that meet their wishes. Recommendation system is an important tool to solve this contradiction. As its name implies, it recommends products of interest to users of electronic commerce. It recommends products to users according to their personal attributes, personal preferences and purchasing habits. The recommendation in this article is based on insurance. At present, the recommendation of insurance in the world is still in a blank stage. If we only apply the existing recommendation algorithm, when the user's attribute is not clear or the behavior of purchasing browsing is not enough, It is difficult to make personalized recommendation, and the effect of recommendation is completely meaningless. Firstly, this paper investigates the current recommendation system of mainstream e-commerce websites at home and abroad, and then analyzes the status quo of X1 insurance e-commerce website. This paper analyzes the requirement of developing insurance personalized recommendation system, and then designs a system framework of insurance personalized recommendation system, which includes the implementation layer of the recommendation system, the engine layer of the recommendation system, and the system framework of the insurance personalization recommendation system, which includes the implementation layer of the recommendation system, the engine layer of the recommendation system. Secondly, in the detailed design stage of the recommendation system, the customer attributes and product attributes are quantified, and the database model is designed according to the needs of each recommendation algorithm. Including the original data table and the recommendation result table which the recommendation algorithm needs, especially designed the personalized and non-personalized recommendation algorithm suitable for each scene and each customer type, including the recommendation algorithm based on statistics. Recommendation algorithm based on association rule, recommendation algorithm based on user collaborative filtering, recommendation algorithm based on item and recommendation algorithm based on content. Finally, a technical scheme of system deployment is designed, which connects the recommendation system with X Insurance E-commerce website.
【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.3

【相似文献】