现在位置:首页 > 学术报告
 

 

Academy of Mathematics and Systems Science, CAS
Colloquia & Seminars

Speaker:

孙志猛 副教授, 中央财经大学统计与数学学院

Inviter:  
Title:
An Imputation Method for Missing Data of large-Scale Network
Time & Venue:
2018.10.19 9:00-10:00 N205
Abstract:
We develop a network imputation method for missing data with large-scale units in this paper. Under the assumption of independence, many researchers have studied traditional regression imputation methods. However, the individuals are usually connected with each other in reality and the assumption of independence fails to catch these valuable information. The connection informa-tion includes friendship in QQ or Wechat and the follower-followee relationship in Weibo and Twitter etc. By using the network structure information, we firstly propose a novel imputation method based on a maximum likelihood method. However, the resulting estimation procedure is inefficiently computational if the network is large and sparse. It is because the determinant and inverse of some related matrixes of large scale need to be computed. To solve the problem, we further propose a maximum approximate partial likelihood estimation (MAPLE) method to estimate the unknown parameters. Then, we develop a novel network imputation method by using the MAPLE. Lastly, we establish the asymptotic properties of the proposed estimator and conduct numerical studies to investigate their finite sample properties .
 

 

附件下载:
 
 
【打印本页】【关闭本页】
电子政务平台   |   科技网邮箱   |   ARP系统   |   会议服务平台   |   联系我们   |   友情链接