報告題目：Assisted Estimation of Gene Expression Graphical Models
主講人：張慶昭副教授（廈門大學）
時間：2019年11月29日（周五）10:00 a.m.
地點：北院卓遠樓305會議室
主辦單位：統(tǒng)計與數(shù)學學院
摘要：In the study of gene expression data, network analysis has played a uniquely important role. To accommodate the high dimensionality and low sample size and generate interpretable results, regularized estimation is usually conducted in the construction of gene expression Gaussian Graphical Models (GGMs). Gene expressions are regulated by regulators. To better decipher the interconnections among gene expressions, conditional GGMs (cGGMs), which accommodate gene expressions as well as their regulators, have been constructed. In practical data analysis, the construction of both GGMs and cGGMs is often unsatisfactory, mainly caused by the large number of model parameters and limited sample size. In this article, we recognize that, with the regulation between gene expressions and regulators, the sparsity structures of the GGMs and cGGMs satisfy a hierarchy. Accordingly, we propose a joint estimation which reinforces the hierarchical structure and use GGMs to assist the construction of cGGMs and vice versa. Consistency properties are rigorously established, and an effective computational algorithm is developed. In simulation, the assisted construction outperforms the separation construction of GGMs and cGGMs. Two TCGA datasets are analyzed, leading to findings different from the direct competitors. Beyond gene expression data, the proposed approach can be potentially applied to a variety of other high dimensional network analysis.
主講人簡介：
張慶昭，現(xiàn)為廈門大學經(jīng)濟學院統(tǒng)計系和王亞南經(jīng)濟研究院副教授、博士生導(dǎo)師。2013年獲得中國科學院數(shù)學與系統(tǒng)科學研究院概率論與數(shù)理統(tǒng)計博士學位，先后在中國科學院大學和美國耶魯大學進行博士后研究。主要研究方向為高維數(shù)據(jù)分析、多源數(shù)據(jù)融合、函數(shù)數(shù)據(jù)分析、統(tǒng)計學習和數(shù)據(jù)挖掘等，在JASA、Biometrics、Statistica Sinica等統(tǒng)計學頂級期刊和一流期刊發(fā)表論文30余篇。國際統(tǒng)計學會推選會員，主持國家自科面上、青年各1項，教育部基金1項。
滇公網(wǎng)安備53010202000350號滇ICP備17008047號 Copyright ? 2013 Yunnan University of Finance and Economics All Rights Reserved
