主 题：面向流数据非平稳密度的非参数估计Nonparametric estimators of nonstationary densities of streaming data
主讲人：澳大利亚墨尔本大学数学与统计人妻偷情 Aurore Delaigle教授
主持人：人妻偷情
常晋源教授
时间：2026年7月6日14：00-15：00
地点：光华校区光华楼1003会议室
主办单位：人妻偷情
科研处
主讲人简介：
Aurore Delaigle, Fellow of the Australian Academy of Science, is a Professor and ARC Future Fellow in the Department of Mathematics and Statistics at the University of Melbourne, Australia. Her research interests include nonparametric statistics, deconvolution and functional data analysis. Following her undergraduate degree in mathematics at Université catholique de Louvain, Belgium, she completed a PhD in statistics at the same institution on kernel estimation in deconvolution problems. In her early career, she undertook a postdoctoral fellowship at University of California, Davis, before joining University of California, San Diego as an assistant professor. She was also a Reader at the University of Bristol. In 2014, she was promoted to Professor at the University of Melbourne. While at UC San Diego, she was awarded a Hellman Fellowship (2006–07). In 2013, she was awarded the Moran Medal from the Australian Academy of Science, for her contribution to "contemporary statistical problems". From 2013 to 2018, she is an ARC Future Fellow, investigating new nonparametric statistical methods. She is a Fellow of the Institute of Mathematical Statistics for her work in "non-parametric function estimation, measurement error problems, and functional data". She is also an elected member of the International Statistical Institute. In 2018 she became a Fellow of the American Statistical Association and in May 2020 she was elected Fellow of the Australian Academy of Science.
Aurore Delaigle是澳大利亚科人妻偷情院士，现任澳大利亚墨尔本大学数学与统计人妻偷情教授，并担任澳大利亚研究委员会（ARC）未来研究员（ARC Future Fellow）。她的研究方向主要包括非参数统计、反卷积和函数型数据分析。
Delaigle教授在比利时鲁汶天主教大学获得数学学士学位，随后于该校取得统计学博士学位，其博士研究主要聚焦于反卷积问题中的核估计方法。在职业生涯早期，她曾在美国加州大学戴维斯分校从事博士后研究，之后加入加州大学圣地亚哥分校担任助理教授，并曾在英国布里斯托大学担任副教授（Reader）。2014年，她晋升为澳大利亚墨尔本大学教授。
在加州大学圣地亚哥分校工作期间，她获得海曼研究员奖（Hellman Fellowship，2006–2007）。2013年，因在“当代统计学问题”领域的杰出贡献，她获得澳大利亚科人妻偷情颁发的莫兰奖章（Moran Medal）。2013年至2018年，她担任澳大利亚研究委员会未来研究员，致力于新型非参数统计方法的研究。
凭借在非参数函数估计、测量误差问题和函数型数据分析领域的重要贡献，她当选为国际数理统计学会（IMS）会士，并成为国际统计学会（ISI）当选会员。2018年，她当选美国统计协会（ASA）会士；2020年5月，正式当选澳大利亚科人妻偷情院士（Fellow of the Australian Academy of Science）。
内容提要：
We consider nonparametric estimation of the nonstationary density of streaming data collected continuously over time. Those data are typically not entirely accessible at all times, and analyzing them requires dynamic approaches that do not require repeated access to past data. Several nonparametric estimators of nonstationary densities have been suggested in the literature, which all require choosing important tuning parameters at each time. We study theoretical properties of those estimators and propose a data-driven dynamic selection of their tuning parameters, which can be implemented iteratively and requires only sequential access to consecutive blocks of the most recent data, and which includes a selection of the sizes of the blocks. We illustrate the procedure through simulated and real streaming data.
本次讲座重点阐释针对随时间持续收集的流数据，其非平稳密度的非参数估计问题。这类数据通常无法在任何时刻都被完全获取，因此对其进行分析需要采用动态方法，且这些方法不能依赖于反复调取过往的历史数据。尽管现有文献中已经提出了一些针对非平稳密度的非参数估计量，但它们都要求在每一个时间点上选取重要的调节参数。为此，本讲座主要围绕这些估计量的理论性质研究进行展开，并提出了一种数据驱动的动态参数选取方法。该方法可以迭代实施，仅需按顺序访问最近的连续数据块，并且包含了对数据块大小的选取策略。讲座将通过模拟数据和真实的流数据对这一流程进行演示说明。
