基于用户画像与RBF的移动广告精准推送研究  

Research on accurate push of mobile advertising based on user profile and RBF

作　　者：李洁琼[1] LI Jieqiong(Xi’an Vocational and Technical College,Xi'an Shanxi 710077,China)

机构地区：[1]西安职业技术学院,西安710077

出　　处：《自动化与仪器仪表》2021年第1期38-42,共5页Automation & Instrumentation

基　　金：陕西省教育厅2020年度一般专项科学研究计划:基于用户画像模型的“矿道网”个性化推送方案研究(No.20JK0924)。

摘　　要：针对当前移动广告的精准推送需求,结合当前的智能技术,提出一种基于用户画像结合协同推荐的广告精准推送方法。在该方法中,基于网络爬虫技术对网页文本关键词的提取,采用VSM向量空间模型构建用户兴趣模型;然后针对个性化推荐中存在评分项目数据稀疏性问题,通过RBF神经网络的方式对评分矩阵进行预测,然后再通过协同推荐算法进行推荐。最后,搭建实验环境,以准确率、召回率和平均绝对误差作为评价指标,以中国电影网中的影评文章和Movielens数据集作为数据来源,分别对上述方案进行验证。结果表明,构建的方案,无论是在用户兴趣度估计方面,还是在推荐精准度方面,较RBF-协同算法和传统的基于内容推荐算法都具有很大的优势,说明本方案的可行性。in view of the current demand of mobile advertising precision push,combined with the current intelligent technology,this paper proposes an accurate advertising push method based on user profile and collaborative recommendation.In this method,web crawler technology is used to extract keywords from web pages,and VSM vector space model is used to build user interest model;then,aiming at the problem of data sparseness of scoring items in personalized recommendation,RBF neural network is used to predict the scoring matrix,and then the collaborative recommendation algorithm is used to recommend.Finally,the experimental environment is set up,and the accuracy rate,recall rate and average absolute error are used as evaluation indexes,and the film review articles and movielens data set in China film network are used as data sources to verify the above schemes.The results show that the scheme constructed in this paper has great advantages over RBF collaborative algorithm and traditional content-based recommendation algorithm in terms of user interest estimation and recommendation accuracy,indicating the feasibility of this scheme.

关 键 词：用户画像 移动广告 中国电影网 Movielens数据集 协同推荐 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]