Optimal discretization for geographical detectors-based risk assessment

Abstract

The geographical detectors model is a new spatial analysis method for the assessment of health risks. It is adapted to discrete risk factors. Meanwhile, the geographical detectors model also effectively analyzes the continuous risk factors by discretizing the continuous data into discrete data. The biggest difficulty is in deciding how to discretize continuous risk factors using the most appropriate discretization method. In this paper, we will discuss the selection of an optimal discretization method for geographical detectors-based risk assessment, and exemplify the process using neural tube defects (NTD) from the Heshun County, Shanxi Province, China.