关联规则在移动电子商务推荐系统中的应用研究

发布时间：2018-10-08 14:12

【摘要】：移动电子商务是在无线平台上实现的电子商务。近年来移动电子商务由于其方便快捷的支付方式以及随时随地提供服务的优势，得到了迅猛发展；同时随着4G网络和智能手机的普及，移动电子商务具有极大发展潜力。相比于运行在PC端的传统电子商务，移动平台在屏幕上所能展示的商品信息相对有限，如何让用户迅速找到自己感兴趣和需要的商品，避免信息超载，而不是迷失在大量商品信息中，已成为移动电子商务发展的一个亟待解决的问题。推荐系统是解决信息超载问题的一个有效方法，它通过分析用户的兴趣偏好、个性化需求等，使用关联规则、协同过滤等推荐技术向用户推荐个性化信息。然而在实际应用中，已有的电子商务推荐系统仍存在着一些问题，如稀疏问题，可扩展，模型过拟合等问题，导致推荐效率较低，推荐质量不高，不能够满足用户的个性化需求，因此，对于移动电子商务推荐系统和推荐技术的研究具有比较大的实用价值。本文以实际的移动电子商务系统为应用背景，通过分析移动电子商务推荐系统的特征和当前推荐系统在移动电子商务和海量数据环境下存在的实时性和推荐效率不高等问题，对推荐系统的体系结构、功能模块以及工作流程进行了学习研究，设计了基于关联规则的移动电子商务推荐系统模型，该模型将推荐过程分为离线处理和在线推荐两大部分，离线处理又分为数据预处理和关联规则挖掘两个子模块。数据预处理通过数据库触发器和存储过程来实现数据的选择清理和格式的转换，，关联规则挖掘部分采用FP-growth算法实现频繁模式的挖掘，生成并导入关联规则库。在线推荐模块根据采集的用户信息与生成的规则库产生准确、实时的个性化推荐结果。该模型在推荐效率、推荐质量上进行了有益研究，推荐模型在移动电子商务系统中的应用有效地提高了推荐效率和实时性，能够很好地为用户推荐符合其兴趣偏好和需求的商品，从而提高商品销量和用户忠诚度。
[Abstract]:Mobile e-commerce is an e-commerce implemented on wireless platform. In recent years, mobile electronic commerce has developed rapidly because of its convenient and quick payment method and the advantage of providing services anytime and anywhere. With the popularity of 4G network and smart phone, mobile electronic commerce has great development potential. Compared with the traditional e-commerce running on the PC, the mobile platform can display relatively limited information on the screen, how to quickly find the products that users are interested in and need, and avoid information overload. Instead of getting lost in a lot of commodity information, it has become an urgent problem in the development of mobile e-commerce. Recommendation system is an effective method to solve the problem of information overload. It recommends personalized information to users by analyzing users' interest preferences, personalized requirements, using association rules, collaborative filtering and other recommendation techniques. However, in the practical application, there are still some problems in the existing E-commerce recommendation system, such as sparse problem, extensibility, model over-fitting and so on, which lead to low recommendation efficiency and low recommendation quality. Therefore, the research on mobile e-commerce recommendation system and recommendation technology is of great practical value. This paper takes the actual mobile electronic commerce system as the application background, through the analysis of the characteristics of the mobile electronic commerce recommendation system and the current recommendation system in the mobile electronic commerce and mass data environment of real-time and recommendation efficiency is not high, and so on. The architecture, function modules and workflow of the recommendation system are studied, and a mobile e-commerce recommendation system model based on association rules is designed. The model divides the recommendation process into two parts: offline processing and online recommendation. Off-line processing is divided into two sub-modules: data preprocessing and association rule mining. Data preprocessing realizes data selection and format conversion by database triggers and stored procedures, and association rules mining uses FP-growth algorithm to mine frequent patterns and generate and import association rules database. The online recommendation module generates accurate and real-time personalized recommendation results according to the collected user information and the generated rule base. This model has carried on the beneficial research in the recommendation efficiency, the recommendation quality, the recommendation model in the mobile electronic commerce system application has effectively improved the recommendation efficiency and the real-time performance. It can recommend the products according to their interests, preferences and needs to improve the sales volume and customer loyalty.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP391.3

【参考文献】