Title: Evaluating Resampling Methods For Validating Data-Driven Causal Structures
Causal inference has developed methods for estimating causal effect coefficients under a variety of different conditions. In particular, any individual method will assume that the data-generating model containing the estimated causal effect coefficient has a specific structure. Background knowledge is often insufficient to verify whether the assumed structural conditions are met, but the data provides another source of information about causal structure. Numerous methods exist for estimating causal structure from observational data, however validating the estimated structure is also difficult. In this talk I present simulation results evaluating the use of resampling methods for validating estimated causal structure models.
Dr. Kummerfeld's primary research interest is in statistical and machine learning methods for discovering causal relationships, with a special focus on discovering causal latent variable models. His work includes (1) developing novel algorithms for discovering causal relationships and latent variables, (2) proving theorems about the properties of causal discovery and latent variable discovery algorithms, (3) performing benchmark simulation studies to evaluate features of the algorithms that are difficult or impossible to evaluate by other means, and (4) applying these novel algorithms to health data in order to inform the development of new treatments.