统计学系系列讲座之343-47期

统计学系系列讲座之343

 

时 间:2019年6月18日(星期二)10:30-11:30

地 点:史带楼205室

主持人:沈娟 副教授 复旦大学管理学院统计学系

主 题:Integrating multi-source block-wise missing data in model selection

主讲人:Prof. Annie Qu University of Illinois at Urbana-Champaign

简 介:

Annie Qu(瞿培勇)教授是伊利诺伊大学香槟分校(UIUC)统计咨询中心主任,复旦大学讲座教授。

摘 要:

For multi-source data, blocks of variable information from certain sources are likely missing. Existing methods for handling missing data do not take structures of block-wise missing data into consideration. In this paper, we propose a Multiple Block-wise Imputation (MBI) approach, which incorporates imputations based on both complete and incomplete observations. Specifically, for a given missing pattern group, the imputations in MBI incorporate more samples from groups with fewer observed variables in addition to the group with complete observations. We propose to construct estimating equations based on all available information, and optimally integrate informative estimating functions to achieve efficient estimators. We show that the proposed method has estimation and model selection consistency under both fixed-dimensional and high-dimensional settings. Moreover, the proposed estimator is asymptotically more efficient than the estimator based on a single imputation from complete observations only. In addition, the proposed method is not restricted to missing completely at random. Numerical studies and ADNI data application confirm that the proposed method outperforms existing variable selection methods under various missing mechanisms. This is joint work with Fei Xue.

 

统计学系系列讲座之344

 

时 间:2019年6月19日(星期三)10:00-11:00

地 点:史带楼301室

主持人:夏寅 教授 复旦大学管理学院统计学系

主 题:Adversarial Risk Analysis

主讲人:Professor David Banks Department of Statistics, Duke University

简 介:

Dr. David Banks is currently the Director of the Statistical and Applied Mathematical Sciences Institute, and a professor in the Department of Statistical Science at Duke University.  He has held previous positions at UC Berkeley, the University of Cambridge, and Carnegie Mellon University.  He also served as the chief statistician of the U.S. Department of Transportation, and held positions at both the National Institute of Standards and Technology, and the U.S. Food and Drug Administration.  He obtained his PhD in 1984 at Virginia Tech, and has served as editor of the Journal of the American Statistical Association and Statistics and Public Policy.  His research areas include dynamic text networks, risk analysis, agent-based models, biosurveillance, computational advertising, and the estimation of cryptic populations.

摘 要:

Adversarial Risk Analysis (ARA) is a Bayesian alternative to classical game theory.  Rooted in decision theory, one builds a model for the decision-making of one's opponent, placing subjective distributions over all unknown quantities. Then one chooses the action that maximizes expected utility.  This approach aligns with some perspectives in modern behavioral economics, and enables principled analysis of novel problems, such as a multiparty auction in which there is no common knowledge and different bidders have different opinions about each other.   

 

统计学系系列讲座之345

时 间:2019年6月24日(星期一)10:00-11:00

地 点:史带楼205室

主持人:朱仲义 教授 复旦大学管理学院统计学系

主 题:Regression Analysis with individual-specific patterns of missing covariates

主讲人:林华珍教授 西南财经大学

简 介:

林华珍教授,西南财经大学统计研究中心主任, 教育部**学者特聘教授,国家杰出青年科学基金获得者,国家百千万人才工程获得者,享受国务院政府特殊津贴专家,教育部新世纪优秀人才。

主要研究方向为非参数方法、转换模型、生存数据分析、函数型数据分析、潜变量分析、ROC方法、偏态数据分析、捕获-再捕获数据分析,发表学术论文40余篇,其中包括发表在国际统计学四大顶级期刊AoS、JASA、JRSSB、Biometrika和计量经济学顶级期刊JOE上论文若干。先后七次主持国家自然科学基金项目。

摘 要:

It is increasingly common to collect data from heterogeneous sources in practice. Two major challenges complicate the statistical analysis of such data. First, only a small proportion of units have complete information across all sources. Second, the missing data patterns vary across individuals. Our motivating online-loan data have 93% missing covariates where the missing pattern is individual-specific. The existing regression analysis with missing covariates either are inefficient or require additional modeling assumptions on the covariates. We propose a simple yet efficient iterative least squares estimator of the regression coefficient for the data with individual-specific missing patterns. Our method has several desirable features. First, it does not require any modeling assumptions on the covariates. Second, the imputation of the missing covariates involves feasible one-dimensional nonparametric regressions, and can maximally use the information across units and the relationship among the covariates. Third, the iterative least squares estimate is both computationally and statistically efficient. We study the asymptotic properties of our estimator and apply it to the motivating online-loan data.

*Joint work with Wei Liu, Wei Lan.

 

统计学系系列讲座之346

 

时 间:2019年6月21日(星期五)10:00-11:00

地 点:史带楼302室

主持人:张新生 教授 复旦大学管理学院统计学系

主 题:Network-based Survival Models for Predicting Drug Resistance and Responses to Cancer Therapies

主讲人:Professor Yongzhao Shao New York University School of Medicine

简 介:

Dr. Yongzhao Shao is currently a professor in Department of Population Health and Department of Environmental Medicine at New York University. His research areas include biostatistics, mathematical statistics, and statistical modeling in environment and public health. He is an elected fellow at American Statistical Association (ASA)

摘 要:

Recent advancement in immunotherapies and targeted therapies have saved many lives of patients with late-stage cancers. However, currently, only a portion of patients have durable responses to these treatments and drug resistance still exists widely for many cancers. An important task in cancer precision medicine is to build effective statistical models to predict which patient will likely respond to the given therapy and/or will experience drug resistance. As is well known, tumor-associated microenvironment plays important roles in tumor progression and drug resistance. We have developed a multicellular gene network approach to investigating the prognostic role of macrophage-tumor cell interactions in tumor progression and drug resistance in advanced brain tumors (gliomas). Multicellular gene networks connecting macrophages and tumor cells were constructed from samples of RNA-seq data in mice gliomas treated with BLZ945. Subsequently, a differential network-based Cox regression model was built to identify the risk signature using a cohort of 310 glioma samples. A large independent validation set of 690 glioma samples from The Cancer Genome Atlas (TCGA) database was used to test the prognostic significance and accuracy of the gene signature in predicting prognosis and targeted therapeutic response of glioma patients. The multicellular gene network approach developed herein indicates profound roles of the macrophage-mediated tumor microenvironment in the progression and drug resistance of gliomas. The identified macrophage-related gene signature has good prognostic value for predicting resistance to targeted therapeutics and survival of glioma patients, implying that combining current targeted therapies with potentially new macrophage-targeted therapy may be beneficial for the long-term treatment outcomes of glioma patients. This talk is based on joint work with Drs. Xiaoqiang Sun of Zhong-shan School of Medicine and Xiaohua Zhang of University of Macau.

 

统计学系系列讲座之347

 

时 间:2019年6月25日(星期二)13:30-14:30

地 点:史带楼205室

主持人:张新生 教授 复旦大学管理学院统计学系

主 题:Some Recent Developments on Individualized Decision Rules

主讲人:Professor Yufeng Liu University of North Carolina

简 介:

Dr. Yufeng Liu is professor in statistics in Department of Statistics and Operations Research, Department of Genetics, and Department of Biostatistics at University of North Carolina at Chapel Hill. His research interests include statistical learning techniques for complex and high dimensional data, graphical models, and individualized decision rules. He is an elected fellow at American Statistical Association (ASA), Institute of Mathematical Statistics (IMS), and an elected member at International Statistical Institute (ISI).

摘 要:

With the emergence of precision medicine, estimation of optimal individualized decision rules has attracted tremendous attentions in many scientific areas. One main goal is to develop the most effective tailored decision for each individual. To that end, one needs to incorporate individual characteristics to detect a proper individual decision rule, by which suitable decisions can be made to optimize each individual's outcome. In this talk, I will present some new statistical learning techniques which directly target the optimal individualized decision rule. Theoretical and numerical comparisons with several existing methods will be presented to demonstrate the effectiveness of the proposed methods.

 

统计学系 

2019-6-13