dac_test {tstests} | R Documentation |
The directional accuracy test of Pesaran and Timmermann (1992), and excess profitability test of Anatolyev and Gerko (2005).
dac_test(actual, forecast, ...)
actual |
a series representing the actual value of the series in the out of sample period. |
forecast |
the forecast values of the series in the out of sample period. |
... |
not currently used. |
The null hypothesis for the test of Pesaran and Timmermann (1992) is that the actual and predicted are independent (no sign predictability), whereas the test of Anatolyev and Gerko (2005) measures the significance of the excess profitability under the null hypothesis of no excess excess profitability. Both are Hausman type tests asymptotically distributed as standard Normal.
An object of class “tstest.dac” which has a print and as_flextable method.
The test will not work with constant forecasts.
Pesaran,M.H., Timmermann,A. (1992). “A simple nonparametric test of predictive performance.” Journal of Business & Economic Statistics, 10(4), 461–465.
Anatolyev,S., Gerko,A. (2005). “A trading approach to testing for predictability.” Journal of Business & Economic Statistics, 23(4), 455–461.
data(arma_forecast)
print(dac_test(arma_forecast$actual, arma_forecast$forecast))