Empirical likelihood simultaneous confidence band for conditional variance function
摘要
Simultaneous confidence bands (SCBs) are powerful global inference tools for establishing a band around the true conditional variance function with a specified confidence level for all covariate values. In this paper, we propose empirical likelihood-based SCBs to address this issue. The proposed empirical likelihood-based bands provide advantages such as enhanced coverage accuracy, straightforward implementation, and automatic avoidance of estimating variances and studentizing. The results of the simulation and empirical examples demonstrate that these methods are effective in estimating the volatility of a process and can be used to generate forecasts of future volatility.