学术报告
报告题目：Estimation of Spatial Autoregressive Regression with Lasso-Selected Weights
报告人：尹建鑫 副教授
中国人民大学
时间：2016年3月31日9:30-10:30
地点：统计研究院426教室
摘要:This paper studies the spatial autoregressive regression model. Different from the existing literature where the interaction effects between cross sectional units are modeled as a linear function of spatial weights, we allow those interaction effects to be estimated from the data. With a sample of N observations and N(N-1) interaction effect coefficients to be estimated, we apply the Lasso-type variable selection technique to select the spatial weights and apply the instrumental variable estimation technique to estimate the coefficients. Under the assumption that the number of the different true values of the interaction effect coefficients is finite and fixed and some sufficient conditions, we show that proposed procedure yields a root-N consistent estimator and establish the asymptotic normal distribution of the proposed estimator. We also show that the proposed estimator has oracle property and provide consistent covariance matrix estimator.
欢迎广大师生参加！
统计研究院
2016年3月18日