语义挖掘结合神经网络的电商网页推荐方案  

E-commerce Webpage Recommendation Scheme Base on Semantic Mining and Neural Networks

作　　者：王斐玉 文华 王红梅[2] 

机构地区：[1]新疆交通职业技术学运输管理学院,乌鲁木齐831401 [2]新疆工程学院计算机工程系,乌鲁木齐830052

出　　处：《控制工程》2018年第1期149-153,共5页Control Engineering of China

基　　金：2015年交通运输职业教育科研项目(2015A12);2015年新疆交通职业技术学院院级课题(J-15-12)

摘　　要：基于Web挖掘的网页推荐技术已在电子商务网站中得到广泛应用，针对现有推荐方案不能满足用户的应用需求问题。提出一种结合语义Web挖掘和BP神经网络的电子商务网页推荐方案。首先，对用户搜索的Web日志进行处理，提取出5种特征：内容优先级，时间消耗优先级，电子商务用户对候选网站的显式／隐式反馈，推荐语义和输入偏差量。然后，将这些特征作为BP神经网络的输入特征，经过分类识别最终输出网页的优先级。最后，根据优先级对网页进行排序并推荐给用户。以图书销售网页作为样本进行实验，结果表明，该方案能够快速准确的识别出用户所需的网页。Web mining has achieved many applications in e-commerce webpages, for the issue that the existing recommendation scheme cannot meet the user＇s application requirements, an e-commerce webpage recommendation scheme based on semantic web mining and BP neural network is proposed. First, the Weblog of user search is processed to extract 5 features： content priority, time priority, e-commerce user＇s explicit / implicit feedback, recommendation semantic and input bias. Then, these features are used as the input features of BP neural network, and output the webpage priority by the classification and identification. Finally, it sorts and recommends the webpage to the user according to the priority. Experiments are carried out with the book sale webpage as a sample, the results show that the proposed scheme can fast and accurately identify the user＇s required webpages.

关 键 词：电子商务 网页推荐 语义Wleb挖掘 BP神经网络 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]