摘要
文章阐述了基于移动互联网的精细化营销平台的设计方案,该平台对移动互联网用户的行为进行分析,掌握用户的喜好及聚类分群,为流量经营等精分产品的实现奠定基础。我们引入页面tittle、keywords分析技术,采用K-Means、SVM和Naive Bayes,利用简单贝叶斯模型,完成对未含类主题变量文档的标注。通过SVM算法进行特征向量选取和训练,并利用EM算法获得极大似然估计的最优解。
This paper describes the design of precise marketing platform based on mobile Internet. It analyzes of the behavior from mobile Internet users and user preferences which lays the foundation for the realization of accurate flow management. We introduce the page tittle, keywords analysis, K - Means, SVM and Naive Bayes. By using the simple Bayesian model, we accomplish the mark of documentations lack of certain topics. By means of the algorithms of SVM and EM, we obtain the ootimal solution of the maximum likelihood estimation.
出处
《成都航空职业技术学院学报》
2015年第3期45-49,共5页
Journal of Chengdu Aeronautic Polytechnic
基金
四川省自然科学重点项目<云计算在信息化建设中的应用研究>(编号:15ZA0399)成果之一
