Abstract:The increasing prominence of the problem of censored data in various fields has made studying how to perform parameter estimation and variable selection in censored mixed-effects models one of the hotspots of current research. In this paper, considering the situation that the response variable is restricted by the bilateral limit, a double-penalty Bayesian Tobit quantile regression model was constructed to carry out parameter estimation and variable selection in the interval-censored mixed-effects model, and at the same time, the fixed-effects and random effects coefficients are compressed in Tobit's mixed-effects model, so as to reduce the estimation error of the model at the same time as the variable selection of the model is carried out. The posterior distribution of each unknown parameter was derived using the conditional Laplace prior and the mixed truncated normal distribution of interval-censored data, and then the Gibbs sampling algorithm for unknown parameter estimation was constructed. Through Monte Carlo simulation, it was found that the new method is more advantageous than the classical method in terms of variable selection and parameter estimation accuracy in various situations, such as different model sparsity, different data censoring ratios and different random error distributions, and the model is able to realize automatic variable selection. Finally, the new method is used to analyze the correlation between the crime rate and various economic indicators in China.
What problem does this paper attempt to address?
The paper mainly aims to address the following issues:
### Research Background and Objectives
- **Research Background**: In the practical application of statistics, the problem of data censoring due to various factors is becoming increasingly prominent. How to perform parameter estimation and variable selection in censored mixed-effects models has become a current research hotspot.
- **Objective Problem**: The paper proposes a Bayesian double-penalty quantile regression method to address the issues of parameter estimation and variable selection in longitudinal interval-censored data.
### Specific Problems Addressed
1. **Parameter Estimation and Variable Selection**: Traditional linear regression models and Tobit models have limitations in handling interval-censored data, especially when the response variable is subject to double-sided constraints. This paper aims to improve parameter estimation and achieve variable selection by constructing a double-penalty Bayesian Tobit quantile regression model that combines the conditional Laplace prior and the mixed truncated normal distribution of interval-censored data.
2. **Model Applicability in High-Dimensional Data**: As the data dimension increases, existing regression analysis methods face challenges in handling interval-censored data. The proposed method considers different model sparsity, different data censoring proportions, and different random error distributions, and demonstrates the advantages of the new method in these scenarios through Monte Carlo simulations.
3. **Automatic Variable Selection**: The proposed new method can achieve automatic variable selection, which helps reduce model estimation errors while improving the predictive ability of the model.
### Methodological Contributions
- **Model Construction**: The paper constructs a double-penalty quantile regression method based on the Bayesian framework, which is suitable for the Tobit model of longitudinal interval-censored data. This model combines the compression of fixed-effect and random-effect coefficients to reduce model estimation errors.
- **Algorithm Design**: A Gibbs sampling algorithm is proposed to estimate unknown parameters, performing parameter estimation and variable selection by iteratively updating the posterior distribution of each parameter.
### Empirical Analysis
- **Simulation Study**: Through Monte Carlo simulations, the performance of the unpenalized Bayesian Tobit quantile regression (P-BTQR), the double Lasso-penalized Bayesian Tobit quantile regression (PDL-BTQR), and the double adaptive Lasso-penalized Bayesian Tobit quantile regression (PDAL-BTQR) methods are compared under different quantiles, different data censoring proportions, and different random error distributions.
- **Empirical Case**: Finally, the paper uses the new method to analyze the correlation between China's crime rate and various economic indicators, further verifying the effectiveness and practicality of the method.
### Conclusion
In summary, the main objective of this paper is to propose an effective method for parameter estimation and variable selection when dealing with longitudinal interval-censored data, aiming to overcome the limitations of existing methods. The effectiveness and superiority of the proposed method are verified through theoretical analysis and empirical research.