基于机器学习的电子商务推荐系统的研究与应用

Research and Application of the E-Commerce Recommender System Based on Machine Learning

摘要: 随着互联网的快速发展，新型的商业运营模式——电子商务使得我们的生活越来越便捷。本文所考虑的推荐系统是电子商务技术中的一种，其是利用电子商务网站向客户提供商品信息和建议，帮助用户决定应该购买什么产品，模拟销售人员帮助客户完成购买过程。协调过滤推荐技术是推荐系统技术中的一种，其基本思想是用户可以根据兴趣进行分类，类似的用户有着非常相似的利益，可以通过协作用对目标用户接收信息，过滤其他用户使用的建议，且其算法一般可以分为基于记忆和基于模型两类。本文主要研究推荐系统中协调过滤技术的基于模型的算法，即矩阵补全问题在推荐系统中的应用。在本文中，我们构造了矩阵补全问题的非凸连续松弛模型，并运用加速的迭代阈值算法求解模型，然后运用真实数据检验模型和算法在推荐系统中的有效性。

Abstract: With the rapid development of the Internet, the new business operation model—electronic commerce makes our life more and more convenient. The recommender system considered in this paper is one of electronic commerce technologies, which uses e-commerce websites to provide customers with commodity information and suggestions, help users decide what products to buy, and simulate sales personnel to help customers complete the purchase process. Collaborative filtering recommended technology is one of the recommender system technologies, its basic idea is that the user can be classified according to interest, similar users have very similar interests, can receive information through collaboration to the target user, filtering other users use suggestions, and the algorithm can be divided into two categories based on memory and based on the model. This paper focuses on the model-based algorithm of collaborative filtering technique in recommender system, namely the application of matrix completion problem in recommender system. In this paper, we construct a non-convex continuous relaxation model for the matrix completion problem, and use the accelerated iterative threshold algorithm to solve the model, and then use the real data to test the effectiveness of the model and the algorithm in the recommender system.