Book Chapter

Machine Learning in Environmental Exposure Assessment

Cai Changjie, Wang Qingsheng

2023,2022

Overview

Environmental exposure assessment seeks to quantify exposure to potentially toxic environmental stressors, especially airborne pollutants. Health effects are typically estimated by regressing health outcome data on estimates of pollution exposure. Environmental exposure assessment is concerned with measuring or estimating exposure for the population of interest. The formulation of exposure assessment as a supervised learning regression problem naturally led to the use of machine learning models. A wide variety of machine learning models have been used, including generalized additive models, multivariate adaptive regression splines, kernel‐based regularized least squares, support vector regression, deletion/substitution/addition, cubist regression, cluster‐based bagging of mixed‐effect models, random forest, gradient boosting, neural networks, Bayesian regularized neural networks, generative adversarial networks. Model comparison, tuning, averaging, and feature selection all require reliable evaluation procedures. Cross‐validation and the bootstrap are two such procedures that are commonly used to evaluate exposure models.

• Publisher: John Wiley & Sons,John Wiley & Sons, Incorporated,John Wiley & Sons, Ltd
• Subject: airborne pollutants / Chemistry & Chemical Engineering / environmental exposure assessment / General References / Industrial Safety / kernel‐based regularized least squares / machine learning exposure models / multivariate adaptive regression splines / neural networks / population of interest / Safety & Industrial Hygiene / support vector regression
• ISBN: 9781119817482, 111981748X