Scalable Quantile Regression and Adaptive Optimal Decision Making

Dr Xu, Jinfeng   (Principal Investigator (PI))

30

2018-09-01

304301

Scalable Quantile Regression and Adaptive Optimal Decision Making

Biostatistics, Quantile Regression, Statistics, Survival Analysis

Probability & Statistics

Physical Sciences (P)

17308018

General Research Fund (GRF)

2018

Completed

1 Develop a computationally efficient single-pass algorithm for quantile regression with applications to massive survival data; Establish the asymptotic properties such as uniform consistency and weak convergence for the proposed non-averaged and averaged estimators in scalable quantile regression 2 Develop a scalable resampling-based inferential procedure for the proposed estimators in scalable quantile regression 3 Establish the distributional approximation theories for justifying the proposed resampling-based strategy for statistical inference in scalable quantile regression 4 Propose an adaptive scheme for estimating optimal decision making; Establish the consistency and asymptotic normality of the proposed adaptive estimator for optimal decision making 5 Develop scalable inferential procedures for the adaptive estimator and justify them by establishing distributional approximation theories 6 Investigate the efficiency of the non-averaged and averaged estimators for both quantile regression and optimal decision making 7 Explore how to optimally assign a decision to a newly arriving subject in the adaptive scheme to maximize the expected outcome of interest