A0150
Title: Fusion learning: Combining inferences from diverse data sources with heterogeneous data
Authors: Regina Liu - Rutgers University (United States) [presenting]
Dungang Liu - University of Cincinnati (United States)
Min-ge Xie - Rutgers University (United States)
Abstract: Nowadays, advanced data collection technology makes inferences from diverse data sources easily accessible. Fusion learning combines inferences from multiple sources or studies to make more effective inferences than from any individual source or study alone. The tasks are focused on: 1) Whether/When to combine inferences? 2) How to combine inferences efficiently if you need to? A general framework for nonparametric and efficient fusion learning for inference on multi-parameters, which may be correlated, is present. The main tool underlying this framework is the new notion of depth confidence distribution (depth-CD), which is developed by combining data depth, bootstrap and confidence distributions. It is shown that depth-CD is an omnibus form of confidence region, whose contours of level sets shrink toward the true parameter value, and thus an all-encompassing inferential tool. The approach is shown to be efficient, general and robust. Specifically, it achieves high-order accuracy and Bahadur efficiency under suitably chosen combining elements. It readily applies to heterogeneous studies with a broad range of complex and irregular settings. This property also enables the approach to utilize indirect evidence from incomplete studies to gain efficiency for the overall inference.
