题目：An Introduction to Feature Screening in High Dimension and Model Averaging Prediction
报告摘要：This talk consists of two parts. In the first part, I will give an introduction to ultrahigh-dimensional feature screening. Specifically, three concrete approaches including: (a) the conditional quantile correlation-based sure independence screening (CQC-SIS) under varying coefficient models, (b) the copula partial correlation-based sure independence screening (CPC-SIS), and (c) a robust partial correlation-based sure independence (RPC-SIS) will be reviewed from our recent work. For each approach, I will present the motivation, methodology, and main theoretical results. Some empirical results would be shown if time permits. In the second part, I will briefly talk about some recent developments of frequentist model averaging prediction, mainly taking the Mallows&apos’s $C_p$ criteria-based model averaging (MMA) as an example.