To testify the hypotheses about the role of characteristic Xzmw in activity F prediction, Decision Making Theory and Zadeh’s fuzzy logic are used.

 

1) For each sequence Sn with known activity Fn the characteristic Xzmw is calculated:

 

2) The simple regression is optimized based on the all {Xzmw(Sn); Fn} pairs:

 

3) Totally the following 22 conditions {q i}1£ i£ 22 of the regression analysis applicability are tested: linear, sign, rank correlations between Xzmw(Sn) and Fn values; the equality of distributions of these values, and others.

 

4) The significance level a i is estimated at which the condition q i is met.

 

 

 

 

back