概率统计讨论班——Navigating the Challenges of Causal Inference: A Focus on Longitudinal Data
报告人:Andrew Ying(Google, Data Scientist)
时间:2024年5月10日下午15:10-16:10
地点:海纳苑2幢105
摘要:Causal inference is a rapidly evolving discipline with applications across various fields. The gold standard for causal inference is a double-blinded randomized controlled trial (RCT), which ensures exchangeability between treatment and control groups. However, observational data are often used as an alternative due to the high cost and ethical concerns associated with RCTs.
Observational data may lack exchangeability due to confounding factors, presenting challenges for causal inference.
In this talk, I will introduce modern causal inference techniques that address these challenges, covering the potential outcome framework and standard identification assumptions.I will also focus on the unique challenges and opportunities associated with longitudinal data, which consists of repeated measurements over time. Longitudinal data can provide valuable insights into causal relationships but introduces additional complications. I will present some advanced statistical methods designed for longitudinal causal inference.
By the end of this talk, attendees will have a foundational understanding of causal inference, its applications, and considerations for working with longitudinal data.
联系人:赵敏智(zhaomz@zju.edu.cn)