计算与应用数学拔尖博士生系列论坛——Differential network identification via lasso penalized fused D-trace loss

Date:2018-12-07

Speaker:Yichong Wu(Peking University)

Time:2018-12-07 12:00-13:30

Venue:Room 1560, Sciences Building No. 1

12:00-12:30 lunch12:30-13:30 Talk

Abstract: The relation and links of two genes in biological networks are of great importance in practical.

Specially, how the structure of a genetic network differs between two states is more attractive. We model

the differential network as the differences of two matrix under the Gaussian assumption. And then we

proposed a new loss function called fused D-trace loss to estimate the differential matrices. We also

develop a very efficient algorithm based on alternating methods and coordinate descent methods. The

convergency and consistency of the algorithm are mentioned. In the end, some results in simulation studied

and a real-data analysis are showed.

欢迎各位同学积极参加,报名链接https://www.wjx.top/jq/31735632.aspx,报名截止时间为12月6日下午16:00,我们将为报名同学提供午餐。