日時：2025年12月26日(金) 13:30 ~　60分

場所：Webのみ　（※対象は特に限定しない）

講演者：Juraj Bodik 先生（University of Lausanne）　　　　

題目： Causality and extremes

概要：
Extreme events in time series carry essential causal information, yet almost all methods focus on the body of the distribution. Additive models Y=f(X)+e are poorly suited to settings where volatility and tail behavior matter. We introduce a mathematical framework for (Granger-type) causality in extremes, identifying causal links from extreme events. In the end, we also discuss Conditionally Parametric Causal Models (CPCM), extending beyond additive noise models to capture changing variance and heavy tails.

お問い合わせは以下のアドレスにメールでご連絡ください。
dser-center@biwako.shiga-u.ac.jp