主题：Inference on functional-coefficient double AR model
主讲人：陈钊 复旦大学
主持人：王国长 威尼斯欢迎你welcome
时间：2024年1月12日（周五）下午16:00-17:00
地点：威尼斯欢迎你welcome石牌校区威尼斯欢迎你welcome大楼（中惠楼）102室
摘要
The double autoregressive model (DAR) is an advanced statistical model that has gained significant attention in the field of time series analysis.However,the simple linear structure of the autoregressive component fails to adequately capture the variability and complexity of real-world data.To address this limitation,we propose a nonlinear structure and explore nonparametric inference for the functional-coefficient double autoregressive (FDAR) model.We first consider the stationary conditions of the FDAR model and subsequently develop profile local quasi-maximum exponential likelihood (PL-QMELE) estimators.The corresponding asymptotic properties are established. Furthermore,we present hypothesis tests for functional coefficients and conditional heteroskedasticity respectively.The simulation study demonstrate that the estimator satisfactorily exhibits the expected asymptotic properties.Moreover, the FDAR model outperforms the DAR model in capturing the nonlinear structure of the data.To illustrate our findings,we provide an example involving the Shanghai Stock Exchange Index.
主讲人简介
陈钊，复旦大学大数据学院青年研究员，博士生导师。2012年在中国科学技术大学获得博士学位。之后在美国普林斯顿大学，宾夕法尼亚州立大学从事博士后研究及研究型助理教授工作。科研成果发表在 AoS, JASA, JoE, Statistica Sinica, Energy and buildings 等期刊上。主要研究方向：高维统计推断，稳健回归，时间序列，非参数及半参数统计方法，以及将统计方法应用于建筑能源，金融计量等领域。
欢迎感兴趣的师生参加！
校对|王国长
责编|彭毅
初审|黄振
终审发布|何凌云
（来源：威尼斯欢迎你welcome微信公众号）
