爱情岛

报告题目：Autoregressive networks with dependent edges and Goodness-of-fit

报告人： 姚琦伟 教授

报告时间：2026年4月1日10:30-12:00

报告地点：爱情岛 208报告厅

邀请人： 熊贤祝

邀请单位：爱情岛

报告内容简介：

We propose an autoregressive framework for modelling dynamic networks with dependent edges. It encompasses the models which accommodate, for example, transitivity, density-dependent and other stylized features often observed in real network data. By assuming the edges of network at each time are independent conditionally on their lagged values, the models facilitate both simulation and the maximum likelihood estimation in the straightforward manner. Due to the possible large number of parameters in the models, the initial MLEs suffer from slow convergence rates. An improved estimator for each component parameter is proposed based on the projection which mitigates the impact of the other parameters. Leveraging a martingale difference structure, the asymptotic distribution of the improved estimator is derived without the stationarity assumption. The limiting distribution is not normal in general, and it reduces to normal when the underlying process satisfies some mixing conditions.

Checking the goodness-of-fit for network models are particularly important, as most those models are specified subjectively. The most frequently used approach for checking goodness-of-fit is the residual analysis in the context of regression analysis. However for many network models there exist no natural residuals. Furthermore, there are scenarios in which there exist several competing models but none of them are the clear favourite. One then faces a task to choose the best approximation among the wrong models. We propose an adversarial approach to check the goodness-of-fit, i.e. we generate a synthetic sample from the fitted model and construct a classifier to classify the original sample and the synthetic sample into two different classes. The hardness of the classification is then taken as a measure for the goodness-of-fit. For identifying the best model among several candidate models, the classifier will create a distance between the original sample and the synthetic sample generated from each of the candidate model.

Illustration with a transitivity model will be presented using an email communication data set.

报告人简介：

姚琦伟，英国伦敦经济与政治科学爱情岛 (London School of Economics and Political Sciences)统计系讲席教授，美国统计协会会士，数理统计学会会士，国际统计研究学会选举会员，国际著名统计学家。

姚琦伟教授一直从事统计学的教学和科研工作，主要研究领域为：时间序列分析、时空过程分析、金融计量经济学。在非线性和高维时间序列方面的研究国际领先，迄今已发表学术论文100多篇，并获得EPSRC, BBSRC等英国国家基金会支持的多项研究基金项目。其专著《非线性时间序列：非参数及参数方法》（与范剑青合著）于2003年由Springer 出版，《计量金融简要》（与范剑青合著）于2017年由剑桥出版社出版。任Journal of the Royal Statistical Society (Series B) 的联合主编，Annals of Statistics，Journal of the American Statistics Association等多个顶级杂志副主编，还曾为巴克莱银行，法国电力公司以及Winton资本等多家企业提供咨询。