The log-Burr XII regression model for grouped survival data

Area of Science:

• Biostatistics
• Survival Analysis
• Statistical Modeling

Background:

• Grouped survival data, common in life tables, often presents ties.
• Existing models may require adaptation for handling numerous ties effectively.
• Assessing data influence is crucial for robust statistical inference.

Purpose of the Study:

• To evaluate the log-Burr XII regression model for grouped survival data with many ties.
• To assess the performance of maximum likelihood and jackknife estimation methods.
• To introduce and utilize diagnostic measures for identifying influential observations.

Main Methods:

• Utilizing life table methodology with data grouped into k intervals.
• Fitting discrete lifetime regression models to the grouped survival data.
• Employing maximum likelihood and jackknife methods for parameter estimation.
• Applying global and local influence measures, including total local influence, for diagnostics.

Main Results:

• The log-Burr XII regression model demonstrates applicability to grouped survival data with ties.
• Monte Carlo simulations assess the finite sample properties of maximum likelihood estimators.
• Diagnostic measures effectively identify influential observations within the model.
• Analysis of a real dataset validates the practical utility of the proposed regression model.

Conclusions:

• The log-Burr XII regression model provides a viable approach for analyzing grouped survival data, particularly when ties are prevalent.
• The diagnostic measures are essential for ensuring the reliability of model results.
• The study confirms the utility of the model and estimation techniques in practical applications.