TheilU(a, p, type = c(2, 1), na.rm = FALSE)NA values should be stripped before the computation proceeds. If set to TRUE complete cases of cbind(x, y) will be used. Defaults to FALSE.
type = 1 is taken from Theil (1958, pp. 31-42). The argument a represents the actual observations and p the corresponding predictions. He left it open whether a and p should be used as absolute values or as observed and predicted changes.
Theil (1966, chapter 2) proposed U type = 2 as a measure of forecast quality: "...where $A_i$ and $P_i$ stand for a pair of predicted and observed changes. ..."
As $U_1$ has some serious disadvantages (see Bliemel 1973) it is recommended to use $U_2$.Thiel, H. (1966): Applied Economic Forecasting. Chicago: Rand McNally.
Bliemel, F. (1973): Theil's Forecast Accuracy Coefficient: A Clarification, Journal of Marketing Research Vol. 10, No. 4 (Nov., 1973), pp. 444-446
GiniTheilU(1:10, 2:11, type=1)
TheilU(1:10, 2:11, type=2)
Run the code above in your browser using DataLab