연사: 김재경 교수(KAIST/IBS)
일시: 5월 27일 화요일 16:00~18:00
장소: 과학기술1관 코워킹스페이스
제목: Advancing Static and Time-series data: Random Matrix Theory, Causal Inference and Mathematical Modeling
초록: In this talk, I will discuss methods for extracting meaningful information from static and time-series data. For static data, Principal Component Analysis (PCA) is widely used to detect signals in noisy datasets. However, determining the appropriate number of signals often relies on subjective judgment. I will introduce an approach based on random matrix theory to objectively select the optimal number of signals. For time-series data, causal inference techniques such as Granger causality are commonly employed. Unfortunately, these methods often yield high false-positive rates. I will present a novel mathematical model-based approach to causal inference. This approach accurately detect network structures from molecules to climates level. Finally, I will talk about how to use mathematical modeling approach to analyze time-series data with an example of wearable data.