Leveraging Modern Machine Learning Tools in High Dimensional Inference
Dr. Zhanrui Cai
Department of Statistics and Data Science
Carnegie Mellon University
In this talk, we will introduce projects that share the same spirit: borrowing the strength of modern machine learning algorithms to achieve better performance in high dimensional inference. The inference problem of interests include test of independence and predictive inference. Those testing problems have broad applications in variable selection, graphical models, and causal inference. Specifically, we propose a new framework that allows us to borrow the strength of the most advanced classification algorithms developed from the modern machine learning community, making the new methods applicable to high dimensional, complex data. Extensive simulations demonstrate the advantages of the newly proposed test compared with existing methods. The tests are also applied to both genetic and economic datasets to demonstrate the advantage of the new methods.