Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Library: VaR
See also: VaRest VaRopt VaRpred VaRdiagtable

Quantlet: VaRver
Description: verifies probability forecasts

Usage: stat = VaRver(tab)
Input:
tab n x 3 matrix, frequency table of forecasts and its success.
Output:
stat list containing various statistics suitable for the verification of probability forecasts: "stat.miscal": which is measure of miscalibration "stat.miscreg": the regression based on miscalibration "stat.varf": variance of forecasts "stat.resol": measure of resolution "stat.diffm": difference in conditional means "stat.sigfx1", "stat.sigfx0": conditional variances of forecasts "stat.me": mean error (ME) "stat.biasreg": regression based measure of bias "stat.mse": mean square error (MSE) Brier score "stat.ss": skill score "stat.potskil": measure of potential skills of a perfectly calibrated procedure

Example:
library("VaR")
x=read("kupfer")
x=x[1:1001]
y=diff(log(x))         ; returns
sig=VaRest(y)[,2]/qfn(0.99)
y=y[251:1000]
table=VaRdiagtable((-0.01~0.01),sig,y,10)
stat=VaRver(table)
setenv("outputformat","% 8.5f")
stat.mse
stat.me

Result:
MSE and ME of the forecasts:

Contents of mse
[1,]  0.23450
Contents of me
[1,] -0.06667



Author: Z. Hlavka, 20011802 license MD*Tech
(C) MD*TECH Method and Data Technologies, 05.02.2006