TVQRLB_OP: Quantile Regression Model with Time-Varying Covariates under...

In ZexiCAI/TVQRLB: Quantile Regression Model with Time-Varying Covariates under Length-Biased Sampling

Description

Fits a quantile regression model to the provided dataset where the covariates are viewed as time-dependent and the sampling is length-biased. The parameters are obtained by minimizing the Euclidean norm of certain estimating equations. For the standard error estimation, a more computationally efficient algorithm named "Orthogonal Procrustes" method, based on matrix singular value decomposition, is used.

Usage

TVQRLB_OP(dataset, betao, bootstrap_time, qtile, B_size = 10000)

Arguments

dataset	The survival data.
betao	The initial estimate for the parameter.
bootstrap_time	The bootstrapping time for estimation of the "meat" matrix V.
qtile	The quantile level used to conduct the quantile regression.
B_size	The resampling size for the OP method. Default is 10000.

Value

This function returns a list of lists with each list containing three elements:

• qtile, the quantile level specified to fit the model

• beta_est, the estimated value of parameter

• sd_OP, the standard error of parameter estimates

References

Cai, Z. and Sit, T. (2018+), "Quantile regression model with time-varying covariates under length-biased sampling," Working Paper.

Examples

TVQRLB_OP(fixedCP.cen20, c(-1, 0.5, 1.5), 1000, 0.5, 10000)
TVQRLB_OP(fixedCP.cen40, c(-1, 0.5, 1.5), 1000, 0.5, 10000)
TVQRLB_OP(fixedCP.cen60, c(-1, 0.5, 1.5), 1000, 0.5, 10000)
TVQRLB_OP(exponCP.cen20, c(-1, 0.5, 1.5), 1000, 0.5, 10000)
TVQRLB_OP(exponCP.cen40, c(-1, 0.5, 1.5), 1000, 0.5, 10000)
TVQRLB_OP(exponCP.cen60, c(-1, 0.5, 1.5), 1000, 0.5, 10000)

ZexiCAI/TVQRLB documentation built on Dec. 30, 2019, 6:02 p.m.