Back To Index Previous Article Next Article Full Text

Statistica Sinica 32 (2022), 2023-2046

POISSON REGRESSION WITH
ERROR CORRUPTED HIGH DIMENSIONAL FEATURES

Fei Jiang and Yanyuan Ma

The University of California, San Francisco and Pennsylvania State University

Abstract: Features extracted from aggregated data are often contaminated with errors. Errors in these features are usually difficult to handle, especially when the feature dimension is high. We construct an estimator of the feature effects in the context of a Poisson regression with a high dimensional feature and additive measurement errors. The procedure penalizes a target function that is specially designed to handle measurement errors. We perform optimization within a bounded region. Benefiting from the convexity of the constructed target function in this region, we establish the theoretical properties of the new estimator in terms of algorithmic convergence and statistical consistency. The numerical performance is demonstrated using simulation studies. We apply the method to analyze the possible effect of weather on the number of COVID-19 cases.

Key words and phrases: Composite gradient descent, COVID-19, non-convex optimization, Poisson regression, measurement error.

Back To Index Previous Article Next Article Full Text