confint is used to estimate the bootstrap confidence intervals for objects with class
attribute tvlm, tvar, tvirf, tvsure and tvplm.
# S3 method for tvlm confint( object, parm, level = 0.95, runs = 100, tboot = c("wild", "wild2"), ... ) # S3 method for tvar confint( object, parm, level = 0.95, runs = 100, tboot = c("wild", "wild2"), ... ) # S3 method for tvsure confint( object, parm, level = 0.95, runs = 100, tboot = c("wild", "wild2"), ... ) # S3 method for tvvar confint( object, parm, level = 0.95, runs = 100, tboot = c("wild", "wild2"), ... ) # S3 method for tvirf confint( object, parm, level = 0.95, runs = 100, tboot = c("wild", "wild2"), ... ) # S3 method for tvplm confint( object, parm, level = 0.95, runs = 100, tboot = c("wild", "wild2"), ... )
| object | An object used to select a method. |
|---|---|
| parm | A specification of which parameters are to be given confidence intervals, either a vector of numbers or a vector of names. If missing, all parameters are considered. |
| level | Numeric, the confidence level required (between 0 and 1). |
| runs | (optional) Number of bootstrap replications. |
| tboot | Type of wild bootstrap, choices 'wild'(default), 'wild2'. Option 'wild' uses the distribution suggested by Mammen (1993) in the wild resampling, while 'wild2' uses the standard normal. |
| ... | Other parameters passed to specific methods. |
an object of class tvsure with BOOT, Lower and Upper different from NULL.
Chen, X. B., Gao, J., Li, D., and Silvapulle, P (2017) Nonparametric estimation and forecasting for time-varying coefficient realized volatility models, Journal of Business \& Economic Statistics, online, 1-13.
Mammen, E (1993) Bootstrap and wild bootstrap for high dimensional linear models, Annals of Statistics, 21, 255-285.
if (FALSE) { ##Calculation of confidence intervals for a TVLM model ##Generation of time-varying coefficients linear model set.seed(42) tau <- seq(1:200)/200 beta <- data.frame(beta1 = sin(2*pi*tau), beta2= 2*tau) X1 <- rnorm(200) X2 <- rchisq(200, df = 4) error <- rt(200, df = 10) y <- apply(cbind(X1, X2)*beta, 1, sum) + error data <- data.frame(y = y, X1 = X1, X2 = X2) ##Fitting the model and confidence interval calculation model.tvlm <- tvLM(y ~ 0 + X1 + X2, data = data, bw = 0.29) tvci <- confint(model.tvlm, level = 0.95, runs = 20) ##If a second confidence interval on the "same" object is calculated, ##for example with a different level, the calculation is faster tvci.80 <- confint(tvci, level = 0.8) }