数据科学与统计前沿问题研讨会
——华东师大&西安交大校级交流研讨
会议日程
会议地点:西安交通大学数学楼2-1会议室
时间:2024年6月6日
日期 |
时间 |
内容 |
主持人 |
6月6日 |
9:00-9:20 开幕式 |
孙建永院长致辞 |
孟德宇 教授 |
合影留念 |
9:20-10:00 学术报告 |
报告人:周勇教授 (华东师范大学) 题目:复杂商务场景下的统计学习与管理决策 |
10:00-10:30 学术报告 |
报告人:马慧娟副教授 (华东师范大学) 题目:Quantile Regression Models for Compliers in Randomized Experiments with Noncomplianc |
10:30-11:00 学术报告 |
报告人:陈律副教授 (华东师范大学) 题目: Stackelberg reinsurance game between the insurer and the reinsurer |
11:00-11:10 |
茶歇 |
11:10-11:40 学术报告 |
报告人:张澍一助理教授 (华东师范大学) 题目:基于U统计量的经验风险最小化的分布式统计推断 |
姜丹丹 教授 |
11:40-12:10 学术报告 |
报告人:郁淼淼助理教授 (华东师范大学) 题目: A functional-data based online algorithm to detect dynamic and correlated data streams with an application to oil mixing |
12:10-14:00 |
午餐 |
14:00-18:00 |
讨论金融科技与大数据技术学术年会筹备事宜 |
18:00-19:30 |
晚餐 |
报告摘要
报告人:周勇教授
报告题目: 复杂商务场景下的统计学习与管理决策
报告摘要: 大规模复杂商务场景离不开电子商务、互联网金融和移动支付,而这些商业活动会产生大量有价值的复杂数据,这些数据的有效处理和分析涉及到运用大数据方法挖掘数据信息和预测未来市场走势等以提升商业效益和效率,以及商务决策的科学性。复杂商务场景下数据的复杂性,为统计学习与管理决策带来了机遇和挑战。处理大数据的手段,共分为两大类,一种是人工智能和机器学习方法,一种是发展新的统计方法,两者有区别也有联系。然而机器学习得到的结果往往缺乏可解释性,同时也很难进行统计推断,例如显著性检验及区间估计等,但人工智能技术可以实现复杂问题的社会计算,统计机器学习成为必然。本讲座介绍我们最近三年在大数据分布计算、半监督学习、隐私保护相关研究成果,及其在金融、社会学及医疗管理等领域的相关研究。
报告人:马慧娟副教授
报告题目: Quantile Regression Models for Compliers in Randomized Experiments with Noncompliance
报告摘要: Understanding the causal effect of a treatment in randomized experiments with noncompliance is of fundamental interest in many domains. Utilizing the instrumental variable (IV) framework, compliers are the only subpopulation that closely relevant to the assessment of causal treatment effect. In this paper, we study flexible quantile regression models for compliers with and without treatment. We establish unbiased es-timating equations by investigating the relationship between observed data and latent subgroup indicators. A novel iterated algorithm is proposed to solve the discontinu-ous equations that involve unknown parameters in a complicated way. The complier average treatment effect and quantile treatment effects can be estimated. The consis-tency and asymptotic normality of the proposed estimators are established. Numerical results, including extensive simulation studies and real data analysis of the Oregon
health insurance experiment, are presented to show the practical utility.
报告人:陈律副教授
报告题目: Stackelberg reinsurance game between the insurer and the reinsurer
报告摘要: We propose a continuous-time framework to analyze optimal reinsurance, in which an insurer and a reinsurer are two players of a stochastic Stackelberg differential game. This allows us to determine optimal reinsurance from joint interests of the insurer and the reinsurer, which is rarely considered in the continuous-time setting. In the Stackelberg game, the reinsurer moves first and the insurer does subsequently to achieve a Stackelberg equilibrium toward optimal reinsurance arrangement. Under mean variance maximization criteria, we study the game problem and find the equilibrium can be achieved by the pair of a variance reinsurance premium principle and a proportional reinsurance treaty, or that of an expected value reinsurance premium principle and an excess-of-loss reinsurance treaty. Moreover, we consider the optimal structure when multiple reinsurers involved in a reinsurance chain as participants of the game.
报告人:张澍一助理教授
报告题目:基于U统计量的经验风险最小化的分布式统计推断
报告摘要:基于U统计量的经验风险最小化问题涵盖了统计学的广泛应用领域,例如成对排序和生存分析等,而其中的一个关键问题是,U统计量的计算复杂度随着样本量的增加而呈多项式阶增长。在大数据下,这就带来了计算上的挑战,因为大量观测数据难以在单台机器上进行集中式计算。我们在分布式系统下开发了两种基于分治策略的高效通讯算法,方法一基于推广的替代似然思想,方法二将M估计推广到了成对损失而得到了一步更新估计,并在此基础上提出了迭代算法。我们证明了所提出的估计量与基准全局估计量渐近等价,并通过大量数值模拟进行验证。
报告人:郁淼淼助理教授
报告题目:A functional-data based online algorithm to detect dynamic and correlated data streams with an application to oil mixing
报告摘要:The occurrence of oil mixing during pipeline transportation causes potential risks and significant losses. Oil density shows dynamic and long-term correlations during transport. To solve the data dynamics problem, it is often possible to use decorrelation methods and time series methods, which are based on the inverse of the covariance matrix and autoregression modeling respectively. However, the long-term correlations render these methods ineffective. Our goal is to develop a flexible algorithm for the above density data that enables real-time monitoring of oil mixing and provides timely alerts. In this paper, the perspective of functional data is adopted to solve the dynamic and long-term correlations. Since the observed data contains the real numbers rather than the functions, we propose a method to map it to the functional values of the functional data. Subsequently, a slice statistic was introduced and the variance was found to vary with the observed time point. Therefore, after normalization, we designed a robust online monitoring statistic for monitoring and prove its asymptotic distribution, which helps to establish a control limit. Finally, both our simulations and real case studies demonstrate the superiority of our method.
报告人简介
周勇教授,国家杰出青年基金获得者,教育部长江学者特聘教授,中国科学院百人计划入选者,国务院政府特殊津贴专家,“新世纪百千万人才工程”国家级人选,国际数理统计学会(IMS)会士。华东师范大学经管学部教授,统计学院院长,统计交叉科学研究院院长。 曾任国务院学位委员会第七届统计学科评议组成员,教育部应用统计专业硕士教学指导委员会委员。现任中国优选法统筹法与经济数学研究会副理事长,中国管理科学学会常务理事。科技部重点研发计划项目首席科学家。
周勇教授主要从事大数据分析与建模、金融计量、风险管理、计量经济学、统计理论和方法等科学研究工作,取得许多有重要学术价值和影响的研究成果。先后承担并完成国家自然科学基金项目,国家杰出青年基金,自然科学基金委重点项目等科学项目10余项,科技部重点研发计划项目1项(首席科学家),曾获得省部级奖励二项。在包括国际顶级期刊《The Annals of Statistics》、《Journal of The American Statistical Association》,《Biometrika》,《JRSSB》及计量经济学顶刊《Journal of Econometrics》和《Journal of Business & Economic Statistics》《管理科学学报》等学术杂志上发表学术论文近200余篇。
马慧娟,华东师范大学统计学院与统计交叉科学研究院副教授。中国科学技术大学统计学博士,美国埃默里大学博士后。主要研究方向包括生存分析,分位数回归,因果推断等。在统计学期刊《Biometrika》,《Biometrics》,《Journal of Business & Economic Statistics》和《Statistica Sinica》等期刊发表论文二十余篇。曾主持国家自然科学基金青年项目一项和上海市浦江人才项目一项。现主持国家自然科学基金重点项目子项目一项,参与国家自然科学基金重点项目及科技部重点研发项目等。担任中国现场统计研究会生存分析分会理事。
陈律,华东师范大学统计交叉科学研究院副教授, 博士毕业于华东师范大学,加拿大滑铁卢大学博士后,研究领域为保险精算理论中的再保险投资决策、风险控制以及养老金管理等,先后在Mathematical Finance, Insurance: Mathematics and Economics, Astin Bulletin等国内外保险精算期刊发表十余篇文章。
张澍一,华东师范大学统计学院与统计交叉科学研究院助理教授。北京大学光华管理学院统计系博士,期间在美国爱荷华州立大学统计系联合培养,美国哈佛大学统计系博士后。主要研究方向为大数据统计分析、高维统计、统计交叉应用研究。在《Annals of Statistics》、《Journal of Machine Learning Research》、《Statistica Sinica》等期刊发表论文十余篇,主持国家自然科学基金青年项目1项、教育部人文社会科学研究一般项目1项。入选上海市领军人才(青年海外)、上海市浦江人才计划,担任中国现场统计研究会因果推断分会理事。
郁淼淼, 华东师范大学助理教授,曾任香港科技大学、华东师范大学博士后,研究方向包括大数据分析、隐私保护、质量过程控制。在包括权威期刊 Annals of Applied Statistics, Journal of Quality and Technology, IISE Transactions, Computers &Industrial Engineering, Statistica Sinica等杂志上发表学术论文十余篇。荣获上海市浦江人才计划、“超级博士后”等荣誉称号。