This is a user interface to control the auxiliary parameter in fitting multi-kink quantile regression.
fit.control(toll=1e-4,h=1,it.max=50,K.max=6,stop.if.error=TRUE,dev0=NULL,visual=FALSE,
visualBoot=FALSE,pow=c(1,1),digits=NULL,grid=NULL,n.boot=20)A list with the arguments as components to be used by 'mkqr.fit' and 'mkqr.bea'.
Positive convergence tolerance.
Positive factor (from zero to one) modifying the increments in kink parameter updates during the iterative process.
Positive integer for the maximal number of iterations.
Positive integer for the maximal given number of kink points.
Logical indicating if the estimation algorithm shoud be stopped if some kink point estimators belong to the non-admissible set. Default is FALSE which suggests to remove the non-addimissble change points automatically.
Initial objective value or deviance. Default is NULL which implies that the initial values is unknown.
Logical indicating if the results of estimation process should be printed at each iteration.
Logical indicating if the results of estimation should be printed at each iteration in the bootstrap restarting process.
The powers of the pseudo covariates employed by the algorithm.
If specified it means the desidered number of decimal points of the kink estimators to be used during the iterative algorithm.
It measures how close between the two adjacent change points should be merged, default is NULL.
Positive integer indicating the times of bootstrap re-sampling in the bootstrap restarting algorithm, default is 20.
Chuang Wan
Wei Zhong, Chuang Wan and Wenyang Zhang. (2020) Estimation and inference for multi-kink quantile regression. working paper.
mkqr.bea,mkqr.fit