Four models of stepwise multiple range regression for
estimating tractor fuel consumption was given in Table 4.
The effect of the variable of chassis type was not considerable
while in ANN it was effective such that caused to increase
of network accuracy. It may be explained by the
nominal type of this variable (2WD, 4WD and MFWD). This
variable was entered to the model as an ordinal number
which cannot nobly represent the effect of this variable
compared to the ANN. Also the effect of engine speed