Description Usage Arguments Details Value Author(s) References See Also Examples

Takes a vector of coefficients (valid for orthogonal variables), their standard errors, the significance level the variables were selected at, and the sample size, to return bias-corrected coefficient estimates to account for the bias induced by model selection.

1 | ```
biascorr(b, b.se, p.alpha, T)
``` |

`b` |
a Kx1 vector of coefficients. |

`b.se` |
a Kx1 vector of standard errors of the coefficients in 'b'. |

`p.alpha` |
numeric value between 0 and 1, the significance level at which selection was conducted. |

`T` |
integer, the sample size of the original model selection regression. |

The function computes the bias-corrected estimates of coefficients in regression models post general-to-specific model selection using the approach by Hendry and Krolzig (2005). The results are valid for orthogonal regressors only. Bias correction can be applied to the coefficient path in `isat`

models where the only additional covariate besides indicators is an intercept - see Pretis (2015).

Returns a Kx3 matrix, where the first column lists the original coefficients, the second column the one-step corrected coefficients, and the third column the two-step bias-corrected coefficients.

Felix Pretis, http://www.felixpretis.org/

Hendry, D.F. and Krolzig, H.M. (2005): 'The properties of automatic Gets modelling'. Economic Journal, 115, C32-C61.

Pretis, F. (2015): 'Testing for time-varying predictive accuracy using bias-corrected indicator saturation'. Oxford Department of Economics Discussion Paper.

`isat`

, `coef.gets`

, `plot.gets`

, `isatvar`

, `isattest`

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 | ```
###Bias-correction of the coefficient path of the Nile data
#nile <- as.zoo(Nile)
#isat.nile <- isat(nile, sis=TRUE, iis=FALSE, plot=TRUE, t.pval=0.005)
#var <- isatvar(isat.nile)
#biascorr(b=var$const.path, b.se=var$const.se, p.alpha=0.005, T=length(var$const.path))
##Bias-correction of the coefficient path on artificial data
#set.seed(123)
#d <- matrix(0,100,1)
#d[35:55] <- 1
#e <- rnorm(100, 0, 1)
#y <- d*1 +e
#ys <- isat(y, sis=TRUE, iis=FALSE, t.pval=0.01)
#var <- isatvar(ys)
#biascorr(b=var$const.path, b.se=var$const.se, p.alpha=0.01, T=length(var$const.path))
``` |

Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.