报告人：李木易教授（厦门大学）
香港大学统计学博士(2011)，现任厦门大学WISE与经济学院统计系双聘副教授,博士生导师。主要研究方向为时间序列分析理论与应用、金融统计，研究成果发表在Journal of Econometrics、Journal of Business and Economic Statistics, Journal of Time Series Analysis，《管理科学学报》等期刊。主持国家自然科学基金3项（青年1项、面上2项）以及福建省社科规划项目1项、福建省自科基金项目1项、教育部计量经济学重点实验室（厦门大学）实验教学项目等。获2019年厦门大学教学比赛翻转课堂组二等奖。担任中国概率统计学会第十一届理事，全国工业统计学教学研究会青年统计学家协会第一届理事。为Journal of Econometrics, Journal of Time Series Analysis, Statistics Sinica, Econometric Review, Computation Statistics and Data Analysis, Science in China, China Economic Review, European Journal of Finance等期刊匿名审稿人。
报告摘要：We consider goodness-of-fit tests for vector autoregressive models with uncorrelated but not necessarily independent innovations. In this situation, the conventional multivariate portmanteau tests are severely over-sized due to the misspecification of critical values obtained from distribution. To address this issue, we propose a naive random weighting (NRW) approach to approximate the null distribution when the errors satisfy the specific dependence assumption. When this assumption is further relaxed, we further propose a block-wise random weighting (BRW) procedure and justify its first-order asymptotic validity. Monte Carlo experiments under various scenarios are conducted to evaluate the finite-sample performance of the proposed bootstrapping procedures and a real example is analyzed to demonstrate the usefulness of our methods.
报告时间：11月6日（本周五）9：30-10：30
报告地点：竞慧西楼209