报告人：Prof Vali Asimit

时    间：2025年10月22日下午2:30-3:30

地    点：览秀楼105学术报告厅

报告摘要: Accurate risk estimation is central to actuarial science, insurance, and finance, yet traditional approaches often fail in modern high-dimensional settings where the number of assets, portfolios, or risk factors is of the same order as—or larger than—the available sample size. In such regimes, the sample mean vector and sample covariance matrix yield unstable and systematically biased estimates, leading to underestimation of portfolio volatility and mispricing of risk-based capital requirements. The high-dimensional estimation paradigm addresses these challenges by adopting tools from random matrix theory, shrinkage estimation, and large-sample asymptotics in which both the dimension and sample size grow together. Similar difficulties arise in regression models, where conventional estimators become unreliable and robust alternatives are needed. Within this framework, corrected risk estimators remain consistent under high-dimensional scaling and often admit simple analytical bias adjustments, such as inflation factors or optimal linear shrinkage toward structured targets—closely related in spirit to the well-known credibility premium in actuarial science. While the risk aggregation problem is discussed in broad terms, a fully fleshed out methodology for estimating the mean vector for a large set of risks is provided for which some applications from portfolio theory will be provided.

报告人简介: Prof Vali Asimit joined Bayes (formerly Cass) in January 2011 in Actuarial Science, became a Senior Lecturer in Actuarial Science in August 2013 and become a Reader in Actuarial Science. Vali is a Professor in Actuarial Analytics since August 2019. He pursued his doctoral research on "Dependence Modelling with Applications in Finance and Insurance" at the University of Western Ontario. As part of his academic work he has published and acted as referee for international statistical and actuarial journals. Vali received the 2010 Fortis Award for the best Insurance: Mathematics and Economics (IME) journal paper presented at the 14th International Congress of IME. He is an Associate Editor of the Insurance: Mathematics and Economics journal, one of the two top actuarial journals in the world.