电子商务网站使用推荐系统来分析用户个人的喜好、习惯,并向其推荐信息、商品。针对传统的推荐系统在实际中存在着数据稀疏性及挖掘潜在需求的问题,提出利用ART神经网络的聚类特性与产品本体来设计电子商务推荐系统。同时,当用户的偏好类别偏少时,提出以形式概念分析为基础的概念相似度方法来提高推荐质量。实验表明该方法有效地提高了推荐效率。
E-commerce websites analyze the interests and habits of users and recommend products by recommendation system now.But the traditional recommendation systems have some shortcomings,such as data sparsity and digging out the potential demand.Therefore,by utilizing both the clustering characteristic of ART and product ontology,the design method of EC recommendation system is put forward.A concept similarity measure method based on Formal Concept Analysis(FCA) is proposed to enhance quality of recommendation when lacking for sort of user's interests.The experimental results show that this method can effectively improve the performance of the recommendation system.