報(bào)告人：李亞光 多倫多大學(xué) 博士后

報(bào)告時(shí)間：12月5號(hào)上午 9：00-9：30

報(bào)告地點(diǎn)：騰訊會(huì)議 ID 504 790 537

報(bào)告題目：A model‐based multithreshold method for subgroup identification

摘要：Thresholding variable plays a crucial role in subgroup identification for personalized medicine. Most existing partitioning methods split the sample based on one predictor variable. In this paper, we consider setting the splitting rule from a combination of multivariate predictors, such as the latent factors, principle components, and weighted sum of predictors. Such a subgrouping method may lead to more meaningful partitioning of the population than using a single variable. In addition, our method is based on a change point regression model and thus yields straight forward model‐based prediction results. After choosing a particular thresholding variable form, we apply a two‐stage multiple change point detection method to determine the subgroups and estimate the regression parameters. We show that our approach can produce two or more subgroups from the multiple change points and identify the true grouping with high probability. In addition, our estimation results enjoy oracle properties. We design a simulation study to compare performances of our proposed and existing methods and apply them to analyze data sets from a Scleroderma trial and a breast cancer study.

個(gè)人簡(jiǎn)介： 李亞光，統(tǒng)計(jì)學(xué)博士。2018年11月在中國(guó)科學(xué)技術(shù)大學(xué)取得博士學(xué)位，后在多倫多大學(xué)Dalla Lana公共衛(wèi)生學(xué)院從事博士后研究，先后訪(fǎng)問(wèn)過(guò)新加坡國(guó)立大學(xué)和約克大學(xué)。主要從事高維數(shù)據(jù)分析和個(gè)性化醫(yī)療等領(lǐng)域的研究。在SCIENCE CHINA-Mathematics和Statistics in Medicine等國(guó)際知名學(xué)術(shù)期刊上發(fā)表多篇論文。

報(bào)告人：張佳 西南財(cái)經(jīng)大學(xué) 博士后

報(bào)告時(shí)間：12月5號(hào)上午 9：30-10：00

報(bào)告地點(diǎn)：騰訊會(huì)議  ID 504 790 537

報(bào)告題目：High Dimensional Elliptical Sliced Inverse Regression in non-Gaussian Distributions

摘要：Sliced inverse regression (SIR) is the most widely-used sufficient dimension reduction method due to its simplicity, generality and computational efficiency. However, when the distribution of the covariates deviates from the multivariate normal distribution, the estimation efficiency of SIR gets rather low, and the SIR estimator may be inconsistent and misleading, especially in high-dimensional setting. In this paper, we propose a robust alternative to SIR - called elliptical sliced inverse regression (ESIR) for analyzing high-dimensional, elliptically distributed data. There are wide applications of the elliptically distributed data, especially in finance and economics where the distribution of the data is often heavy-tailed. To tackle the heavy-tailed elliptically distributed covariates, we novelly utilize the multivariate Kendall's tau matrix in a framework of generalized eigenvalue problem in sufficient dimension reduction. Methodologically, we present a practical algorithm for our method. Theoretically, we investigate the asymptotic behavior of the ESIR estimator under high-dimensional setting. Simulation results show that ESIR significantly improves the estimation efficiency in heavy-tailed scenarios. Analysis of the Istanbul stock exchange data set also demonstrates the effectiveness of our proposed method. Moreover, ESIR can be easily extended to other sufficient dimension reduction methods and applied to non-elliptical heavy-tailed distributions.

個(gè)人簡(jiǎn)介：張佳，經(jīng)濟(jì)學(xué)博士。2019年6月在西南財(cái)經(jīng)大學(xué)取得博士學(xué)位，同年進(jìn)入西南財(cái)經(jīng)大學(xué)從事博士后研究，博士后導(dǎo)師為常晉源教授。主要從事高維經(jīng)驗(yàn)似然和充分降維等領(lǐng)域的研究。在CSDA、JMVA和JSPI等國(guó)際知名學(xué)術(shù)期刊上發(fā)表多篇論文。

報(bào)告人：王楊  上海交通大學(xué)  博士后

報(bào)告時(shí)間：12月5號(hào)上午 10：00-10：30

報(bào)告地點(diǎn)：騰訊會(huì)議 ID 504 790 537

報(bào)告題目：A Kernel Regression Model for Panel Count Data with Nonparametric Covariate Functions

摘要：Local kernel pseudo-partial likelihood is used for estimation in panel count model with nonparametric covariate functions. Estimator of the derivative of nonparametric covariate function is derived first and nonparametric function estimator is then obtained by integrating the derivative function. Under some regularity conditions, uniform consistency rates and pointwise asymptotic normality are obtained for the local derivative estimator. Moreover, the baseline function estimator is shown to be uniformly consistent. The demonstration of the asymptotic results relies strongly on the modern empirical theory, which not require the Poisson assumption. Simulation studies also show that the local derivative estimator performs well in finite-sample regardless of whether or not the Poisson assumption holds. We also apply the proposed methodology to analyze a clinical study on childhood wheezing.

個(gè)人簡(jiǎn)介：王楊，2014年本科畢業(yè)于信陽(yáng)師范學(xué)院數(shù)學(xué)系；2016年碩士畢業(yè)于浙江大學(xué)數(shù)學(xué)系；2016年至今，在上海交通大學(xué)統(tǒng)計(jì)系攻讀博士學(xué)位；2019年至2020年，在美國(guó)內(nèi)布拉斯加州醫(yī)學(xué)中心訪(fǎng)學(xué)交流，師從面板計(jì)數(shù)數(shù)據(jù)領(lǐng)域的專(zhuān)家張殷教授。博士期間的研究方向包括面板計(jì)數(shù)數(shù)據(jù)、非參數(shù)統(tǒng)計(jì)分析、生存分析、臨床試驗(yàn)研究。其中博士課題為醫(yī)學(xué)面板計(jì)數(shù)數(shù)據(jù)的非參數(shù)統(tǒng)計(jì)分析。隨著電子病歷數(shù)據(jù)的廣泛應(yīng)用，面板計(jì)數(shù)數(shù)據(jù)在臨床研究應(yīng)用中很常見(jiàn)。面板計(jì)數(shù)數(shù)據(jù)作為一種只能在固定的隨訪(fǎng)時(shí)間點(diǎn)上觀(guān)測(cè)的復(fù)發(fā)性事件數(shù)據(jù)，通常被用來(lái)研究試驗(yàn)因素的固定效應(yīng)，但是在臨床試驗(yàn)中，試驗(yàn)因素的效應(yīng)往往是隨時(shí)間變化或者是非線(xiàn)性。此時(shí)，固定效應(yīng)模型就會(huì)導(dǎo)致效應(yīng)估計(jì)的偏差，而非參數(shù)效應(yīng)模型能夠較好的分析這種時(shí)變效應(yīng)或者非線(xiàn)性效應(yīng)。因此，針對(duì)醫(yī)學(xué)面板計(jì)數(shù)數(shù)據(jù)建立非參數(shù)效應(yīng)模型尤為關(guān)鍵，能夠?yàn)榕R床工作者提供更為準(zhǔn)確的醫(yī)學(xué)指導(dǎo)。博士期間以第一作者完成兩篇文章，一篇于2019年被Statistica Sinica接受，另一篇在Biometrics二審中。與醫(yī)生等合作完成6篇文章，其中4篇已經(jīng)發(fā)表，2篇在審稿中。

報(bào)告人：張樹(shù)雄  北京師范大學(xué)  博士

報(bào)告時(shí)間：12月5號(hào)上午 10：30-11：00

報(bào)告地點(diǎn)：騰訊會(huì)議 ID 504 790 537

報(bào)告題目：On large deviation probabilities for empirical distribution of the branching random walk with heavy tails

摘要：

個(gè)人簡(jiǎn)介：張樹(shù)雄, 概率論專(zhuān)業(yè)博士生。2016.09-2021.07年在北京師范大學(xué)學(xué)習(xí)（碩博連讀)，師從何輝教授。研究方向?yàn)榉种﹄S機(jī)游動(dòng)的大偏差理論，布朗運(yùn)動(dòng)，超過(guò)程等。