Functional varying-coefficient Cox model and its application

Area of Science:

• Biostatistics
• Survival Analysis
• Statistical Modeling

Background:

• Increasing data complexity necessitates more flexible survival analysis models.
• Existing functional Cox models offer a foundation but require enhancements for multifaceted data.
• The need for models that can simultaneously address varying-coefficient and functional covariates is critical.

Purpose of the Study:

• To introduce a novel functional varying-coefficient Cox model.
• To develop corresponding estimation algorithms for the proposed model.
• To enhance the adaptability of survival models for complex, high-dimensional data.

Main Methods:

• Building upon the functional Cox model framework.
• Developing and proposing novel estimation algorithms.
• Simultaneous handling of varying-coefficient and functional covariates.

Main Results:

• The proposed functional varying-coefficient Cox model demonstrates enhanced adaptability.
• Simulation studies confirm the model's performance.
• A practical application using Alzheimer's Disease Neuroimaging Initiative (ADNI) data illustrates its utility.

Conclusions:

• The novel functional varying-coefficient Cox model effectively analyzes complex survival data.
• The model offers significant improvements in flexibility and adaptability.
• The proposed methodology is practical and applicable to real-world datasets, such as neuroimaging data.