講座名稱：Adaptive Increment Learning for Estimating Optimal Individualized Treatment Regimes

講座時(shí)間：2020-08-23 9:30

講座人：朱文圣 教授

講座地點(diǎn)：騰訊會(huì)議直播（ID：880 164 821）

講座人介紹：

朱文圣，東北師范大學(xué)數(shù)學(xué)與統(tǒng)計(jì)學(xué)院教授、博士生導(dǎo)師、副院長。2006年12月博士畢業(yè)于東北師范大學(xué)，2013年12月起任東北師范大學(xué)數(shù)學(xué)與統(tǒng)計(jì)學(xué)院教授。2008-2010年在耶魯大學(xué)做博士后研究，2015-2017年訪問北卡羅來納大學(xué)教堂山分校。中國現(xiàn)場統(tǒng)計(jì)研究會(huì)計(jì)算統(tǒng)計(jì)分會(huì)副理事長、數(shù)據(jù)科學(xué)與人工智能分會(huì)秘書長，中國概率統(tǒng)計(jì)學(xué)會(huì)副秘書長，吉林省現(xiàn)場統(tǒng)計(jì)研究會(huì)秘書長等。主要從事統(tǒng)計(jì)學(xué)的方法與應(yīng)用研究，在Journal of the American Statistical Association、Test、NeuroImage、中國科學(xué)等雜志發(fā)表學(xué)術(shù)論文多篇，主持并完成國家自然科學(xué)基金項(xiàng)目多項(xiàng)。

講座內(nèi)容：

Personalized medicine has recently received increasing attention because of the significant heterogeneity of patient responses to the same medication. The estimation of individualized treatment regimes or individualized treatment rules is an important part of personalized medicine. Individualized treatment regimes are designed to recommend treatment decisions to patients based on their individual characteristics and to maximize the overall clinical benefit to the patients. However, most of the existing statistical methods are mainly focus on the estimation of optimal individualized decision rules for the two categories of treatment options and rely heavily on data from randomized controlled trials. There has been a relative lack of research work on the selection of multi-categorical treatment options in real-world settings. In this work, we address this challenge and propose a machine learning approach (AI-learning) to estimate optimal treatment regimes. This new learning approach allows for more accurate assessment of individual treatment response and alleviation of confounding. The increment value functions are proposed to compare matched pairs under a tree structure, and this approach allows for easy handling of the outcomes of various types of measuring treatment responses (including continuous, ordinal, and discrete outcomes). Through a large number of simulation studies, we demonstrate that AI-learning outperforms existing methods. Lastly, the proposed method is illustrated in an analysis of AIDS clinical trial data.

主辦單位：數(shù)學(xué)與統(tǒng)計(jì)學(xué)院