QQ客服 官方微博 反馈留言

题名	协同过滤算法在移动电子商务推荐系统中的应用研究; Research on Application of Collaborative Filtering Algorithm in the Mobile E-commerce Recommendation System
作者	杨波
答辩日期	2014 ; 2013
导师	陈海山
关键词	移动电子商务 协同过滤算法 推荐系统 E-commerce Collaborative Filtering Algorithm Recommendation System
英文摘要	协同过滤算法是电子商务推荐系统中应用最为成功的推荐算法，很多电子商务企业都采用协同过滤算法作为企业推荐系统的核心算法。推荐技术可以有效解决信息过载问题，增加网站的粘度和交叉销售能力。随着移动互联网的发展，移动电子商务也在逐步兴起，并逐渐占据了大量的市场份额，推荐系统也是必不可少的。与传统电子商务相比，移动电子商务具有移动性，可接受性，安全性，定位性等特点。因此将协同过滤算法应用到移动电子商务推荐系统中时，必须做相应的改进才能适用移动电子商务的特殊环境。 本文对电子商务个性化推荐系统进行了深入研究，认真分析了协同过滤推荐算法本身存在的数据稀疏性问题和冷启动问题。通过SlopeOne算法降低了数据的稀疏性，提高了推荐系统的执行效率，使推荐系统具有较好的推荐质量。冷启动问题可以通过用户的注册信息进行特征分类，根据不同的用户特征对用户进行推荐。 针对移动电子商务领域的特殊性，本文在协同过滤算法中引入了定位技术和时间信息。通过定位技术，移动电子商务在用户的交互过程中能达到很高的个性化程度，从而满足用户对服务的差异化需求。遗忘函数的应用则进一步提高了用户兴趣迁移中的预测精度，可以推荐给用户当前最感兴趣的项目。本文最后通过实验验证了这种改进思路的可行性，证明这种改进算法的确更适合在移动电子商务平台上运用。; Collaborative filtering is the most successful e-commerce recommendation algorithm, many e-commerce companies are using collaborative filtering as the core algorithm of the enterprise recommendation system. Recommendation technology can effectively solve the problem of information overload and increase the viscosity of the website and the ability to cross-sell. With the development of mobile Internet, mobile e-commerce is gradually rising and gradually occupies a significant market share, so recommendation system is also essential. Compared with the traditional e-commerce, mobile e-commerce has the mobility, acceptability, safety, location and other characteristics. It does the corresponding improvement in order to apply the special environment of the mobile e-commerce. The dissertation has in-depth study of e-commerce recommendation system and analyzes the data sparsity and cold start problems of the collaborative filtering algorithm carefully. The Slope One algorithm reduces the data sparsity and improves the efficiency and quality of the recommendation system. Cold start problems can be characterized through the user's registration information classification, recommended depending on the user characteristics. Due to the particularity of mobile e-commerce, positioning technology and time information is introduced into the collaborative filtering algorithm. By positioning technology, mobile e-commerce can achieve a high degree of personalization to meet the different demands of the users. Time information applications to further improve the prediction accuracy of the migration of user interest; can recommend to the user the right item. The experiments verify the feasibility of the improvement algorithm, to prove that this improved algorithm is more suitable for mobile e-commerce platform.; 学位：工程硕士; 院系专业：软件学院_软件工程; 学号：24320101152263
语种	zh_CN
出处	http://210.34.4.13:8080/lunwen/detail.asp?serial=39062
内容类型	学位论文
源URL	[http://dspace.xmu.edu.cn/handle/2288/78190]
专题	软件学院－学位论文
推荐引用方式 GB/T 7714	杨波. 协同过滤算法在移动电子商务推荐系统中的应用研究, Research on Application of Collaborative Filtering Algorithm in the Mobile E-commerce Recommendation System[D]. 2014, 2013.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。