Profile local linear estimation of generalized semiparametric regression model for longitudinal data

Area of Science:

• Statistics
• Biostatistics
• Econometrics

Background:

• Longitudinal data analysis requires flexible models to capture complex covariate effects.
• Existing methods may not adequately handle time-varying coefficients or heterogeneous sampling times.
• Semiparametric models offer a powerful framework for such data.

Purpose of the Study:

• To develop a generalized semiparametric regression model for longitudinal data.
• To estimate both constant and time-varying covariate effects.
• To provide robust statistical inference for complex longitudinal datasets.

Main Methods:

• Utilizing local linear estimating equations for nonparametric components.
• Employing profile estimating functions for parametric components.
• Implementing k-fold cross-validation for bandwidth selection and proposing a link function selection criterion.

Main Results:

• The proposed method automatically adjusts for heterogeneity in sampling times.
• Large sample properties of estimators and confidence intervals are established.
• Formal hypothesis testing procedures for covariate effects and time-varying effects are developed.

Conclusions:

• The developed semiparametric model provides a flexible and robust approach for longitudinal data analysis.
• The estimation and hypothesis testing procedures demonstrate good finite sample performance.
• The methods are validated through simulation studies and a real-world data example.