Efficient and robust optimal design for quantile regression based on linear programming

Article Mathematics, Interdisciplinary Applications

Risk-Adapted Optimal Experimental Design

Drew P. Kouri et al.

Summary: This paper introduces a new optimality criterion, R-optimality, which aims to minimize the risk associated with large prediction variances. The effectiveness of this criterion is demonstrated through numerical methods and various examples.

SIAM-ASA JOURNAL ON UNCERTAINTY QUANTIFICATION (2022)

Add to Collection

Article Operations Research & Management Science

Calculating CVaR and bPOE for common probability distributions with application to portfolio optimization and density estimation

Matthew Norton et al.

Summary: The paper introduces the importance of conditional value-at-risk and value-at-risk in characterizing the tails of probability distributions, as well as the buffered probability of exceedance method. Research shows that minimizing buffered probability of exceedance for finding optimal portfolios has advantages over minimizing superquantiles, and introduces a method of using superquantiles for parametric density estimation.

ANNALS OF OPERATIONS RESEARCH (2021)

Add to Collection

Article Biology

Optimal subsampling for quantile regression in big data

Haiying Wang et al.

Summary: This study investigates optimal subsampling for quantile regression, deriving two versions of optimal subsampling probabilities and proposing algorithms based on these probabilities. An iterative subsampling procedure based on optimal subsampling probabilities is suggested for linearly transformed parameter estimation, with great scalability to utilize available computational resources. Numeric examples are provided to illustrate the proposed method using both simulated and real data.

BIOMETRIKA (2021)

Add to Collection

Article Statistics & Probability