Cluster detection based on spatial associations and iterated residuals in generalized linear mixed models

Area of Science:

• Statistics
• Spatial Analysis
• Biostatistics

Background:

• Generalized linear mixed effect models (GLMMs) are often used for spatial clustering analysis.
• Current methods primarily rely on Bayesian approaches or frequentist hypothesis testing for cluster detection.

Purpose of the Study:

• To develop a novel frequentist method for assessing spatial properties within GLMMs.
• To provide a strategy for detecting spatial clusters using parameter estimates of spatial associations.
• To evaluate spatial model improvement through iterated residuals.

Main Methods:

• A frequentist approach is proposed for analyzing spatial associations in GLMMs.
• Parameter estimates of spatial associations are utilized to detect spatial clusters.
• Iterated residuals are employed to assess spatial aspects of model improvement.

Main Results:

• The proposed frequentist method consistently and efficiently detects spatial clusters.
• The method accurately identifies both the locations and magnitudes of spatial clusters.
• Simulations and a case study validate the effectiveness of the approach.

Conclusions:

• The developed frequentist method offers a robust alternative for spatial cluster detection in GLMMs.
• This approach enhances the assessment of spatial properties and model fit.
• The findings contribute to more accurate spatial data analysis in various scientific fields.