Interest in shrinkage estimators goes back half a century but has rapidly increased recently with many new directions of research that cover a vast range of applications in different disciplines. Ongoing research on information-theoretic estimation and inference methods is similarly inter-disciplinary, involving information theory, engineering, mathematical statistics, econometrics and the natural sciences.
Modern shrinkage estimators like adaptive Lasso and Bridge methods are penalized estimators with various penalty designs. These methods assist in model selection, oracle efficient estimation and provide substantial computational advantages. In these respects they may be viewed as methods that utilize information better than existing techniques. They have been found to be particularly helpful in large dimensional systems where shrinkage can deliver sparse estimation, sparsisity in the limit, and improved finite sample performance. The range of applications is vast and includes microeconometric models of labor supply, time series econometric applications to economic growth, aggregate cointegrated systems, and market wide financial volatility, social interaction network estimation, and gene selection in biology.
This one day conference will address these various themes, the inter-connections between shrinkage estimation methodology and info-metrics, and explore recent advances in shrinkage methods and applications.
Shrinkage Estimators and Info-Metrics
Regularization Methods and Info-Metrics
Model Discovery and Info-Metrics
Mehmet Caner (NC State U) – Co-Chair Amos Golan (American U) – Co-Chair George Judge (UC Berkeley) Robin Lumsdaine (American U) Peter Phillips (Yale) – Co-Chair Eric Renault (Brown)
Andrew Barron (Yale) Victor Chernozhukov (MIT) Harrison Zhou (Yale)
Other Confirmed Speakers
Mehmet Caner (NCSU) Marine Carrasco (Montreal) Bruce Hansen (Wisconsin) George Judge (UC Berkeley) Eric Renault (Brown) Song Song (UC Berkeley) Pascale Valery (HEC-Montreal) Liao Zhipeng (UCLA)