The validity of the hypotheses about the role of the examined characteristic X in determining the site activity F is tested using Decision Making Theory and Zadeh’s fuzzy logic.
The valua U, which is called “the utility of usage of the characteristic X in prediction of the activity F of the given site”, is calculated for each characteristic Xzmw.
1) Simple regression is optimized:
;
f0
č f1 are regression coefficients.
2) Deviation values between the regressed and experimental activities are determined:
D n = Fzmw(Sn) - Fn.
3) A total number 22 conditions of regression analysis
are tested for the values {Fzmw(Sn), Fn, D n}:
- for correlations between Fzmw(S) and F (linear, sign, rank, etc.);
- for the equality of the Fzmw(Sn
- for deviations D
n.
This yields the total number 22 values of