电子商务中滞销商品推荐技术研究

发布时间：2019-01-08 17:57

【摘要】： 随着电子商务和网络技术的迅猛发展,信息和商品的数量急剧增加,在此背景下推荐系统和个性化推荐技术应运而生。推荐系统能够帮助客户寻找其感兴趣的商品,具有巩固老客户,吸引新客户,提高用户的满意度的作用,从而提高销售量。但是,商品数量和品种的增多的同时,也出现了有些具有价值的商品不易被用户发现而滞销的问题,以至物资浪费,利益损失。挖掘滞销商品的价值,不仅能够降低企业的损失,还能凭借低成本创造更大的利益。因此,滞销商品的推荐技术研究在电子商务中具有重要意义。本文首先介绍了几种常用的个性化推荐技术及方法。重点介绍了基于用户的协同过滤推荐技术的基本思想和相似邻居的计算,以及关联规则的应用。其次,通过分析历史销售数据来预测和挖掘滞销商品及与之密切相关的中间商品,提出一种借助中间商品激励用户购买滞销商品的滞销商品推荐模型,并提出按照时间权重的中间商品兴趣度进行项目的个性化推荐。此推荐模型能够有效预测滞销商品,并根据客户特征进行个性化推荐,为滞销商品的推荐提供一种行之有效的方法。此外,在此模型基础上,提出一种利润最大化的滞销商品推荐算法,其基本思想是通过商品的关联性确定最大利润推荐项集,再根据利润函数为每个用户推荐利润最高的项集,其中推荐项集由中间商品和滞销商品组成。最后,本文通过实验验证了滞销商品预测方法,基于利润最大化的滞销商品推荐算法的可行性。
[Abstract]:With the rapid development of electronic commerce and network technology, the quantity of information and commodity increases rapidly. Under this background, recommendation system and personalized recommendation technology emerge as the times require. Recommendation system can help customers to find the goods they are interested in. It can consolidate the old customers, attract new customers, improve customer satisfaction, and thus increase the sales volume. However, with the increase of the quantity and variety of goods, there are also some problems of unsalable goods which are not easy to be discovered by users, even material waste and loss of profits. Excavating the value of unsalable goods can not only reduce the loss of enterprises, but also create greater benefits by means of low cost. Therefore, the recommendation technology of unsalable goods is of great significance in e-commerce. This paper first introduces several commonly used personalized recommendation techniques and methods. The basic idea of user-based collaborative filtering recommendation technology, the computation of similar neighbors, and the application of association rules are introduced. Secondly, by analyzing the historical sales data to predict and excavate the unsalable goods and the intermediate commodities closely related to them, a recommendation model of unsalable commodities is proposed to encourage customers to buy unsalable goods with the help of intermediate commodities. And put forward according to the time weight of the intermediate commodity interest degree to carry on the item personalized recommendation. This recommendation model can effectively predict unsalable goods and make personalized recommendation according to customer characteristics, which provides an effective method for the recommendation of unsalable goods. In addition, on the basis of this model, a recommendation algorithm for unsalable goods with maximum profit is proposed. The basic idea is to determine the maximum profit recommendation item set through the correlation of goods, and then to recommend the highest profit set for each user according to the profit function. The recommended item set consists of intermediate commodities and unsalable commodities. Finally, this paper verifies the feasibility of unsalable goods prediction method and unsalable goods recommendation algorithm based on profit maximization.
【学位授予单位】：沈阳航空工业学院
【学位级别】：硕士
【学位授予年份】：2010
【分类号】：F713.36

【参考文献】