Outlier detection method in GEEs

Area of Science:

• Statistics
• Biostatistics
• Longitudinal Data Analysis

Background:

• Generalized Estimating Equations (GEEs) are widely used for analyzing correlated data.
• Existing diagnostic tools for GEE model fit are limited, particularly for outlier detection.
• Accurate model fitting is crucial for reliable analysis of longitudinal studies.

Purpose of the Study:

• To develop and evaluate a novel outlier detection technique for GEE models.
• To assess the performance of the proposed method in identifying outliers in longitudinal data.
• To provide a practical tool for enhancing the robustness of GEE analyses.

Main Methods:

• Development of an outlier detection technique using the "working" score test statistic.
• Testing a mean-shift model to identify deviations from the expected data structure.
• Application of the method to simulated and real-world longitudinal datasets.

Main Results:

• The proposed method successfully identified known outliers in simulation studies.
• The technique demonstrated effectiveness in detecting outliers within a longitudinal dataset.
• The working score test statistic proved valuable for outlier assessment in GEEs.

Conclusions:

• The developed outlier detection method enhances the diagnostic capabilities for GEE models.
• This technique improves the accuracy and reliability of modeling correlated longitudinal data.
• The method offers a practical approach for identifying and handling outliers in GEE analyses.