PsiTSTransform(type, shift)

PsiTSTransform is a property supported by the time series functions that applies a transformation where:

  • type is the type of transformation to apply to the time series.  The argument type may be 0 (none), 1 (exponentiate) or 2 (square). 
  • shift is the data shift to apply to the time series. The shift argument is applied to the time series after the transformation has occurred.     

PsiTSIntegrate, PsiTSSeasonality and PsiTSTransform may only be used in conjunction with a Time Series and may be applied in any order.  These three properties are necessary to transform the generated vector to the original scale of the historical data. The Time Series which support these properties are:  PsiAR1, PsiAR2, PsiMA1, PsiMA2, PsiARMA11, PsiARCH1, PsiGARCH11, PsiEGARCH11 or PsiAPARCH11.  Additional properties supported by the Time Series functions are PsiLock, PsiStatic and PsiCorrMatrix.  Note that there is no truncation, shifting, censoring or individual seeding for a time series. 


PsiAR1(1,1,0,1,PsiTSTransform(2,2)) applies type = square and shift = 2 to each item in the series.