Page 30 - 75_01

DÍEZ MONTORO, R. & COL. AN. R. ACAD. NAC. FARM.

3.4. Data Analysis

The Statistica programme (Copyright © StatSoft,Inc:, 1993) was
used with specific non-linear regression equations. As the statistical cri-
terion (28,29) that allows a choice from different equations, SS and
Corrected Akaike´s Information Criterion (AICc) was used, expressed as

where N is the number of points, SS

the addition of residual squares, and p the number of parameters in

the equation. The fitting with the lowest AICc must be chosen. In or-
der to distinguish equations from monoexponential and biexponen-

tial models, AICc and ANOVA (F test) were used.

4. RESULTS
Results of z values for experiments 1-16 are shown in Table III.

5. DISCUSSION

5.1. Influence of m (Experiments 1, 2, 3, 4)

The results of experiments 1, 2, 3, 4 are fitted to the equation:

z=a·m/(m+b)+(c·m/(m+d)-a·m/(m+b))·(exp(-t·e)) Eq. 3

Its parameters, coefficient of correlation (r), sum of squares of
residuals (s) and AICc are:

a = 4399, b = 80.4, c = 51993, d = 341, e = 0.0858, r = 0.984, s =
5.59·106, AICc = 357.72.

Or to:

z=a·m/(m+b)+(c·m/(m+d)-a·m/(m+b))·(exp(-t·e))+f·m+(g·m/(m+h)-

f·m)·(exp(-t·j)) Eq. 4

Its parameters, coefficient of correlation, sum of squares of resid-
uals and AICc are:

a = 5577, b = 92.0, c = 52007, d = 340, e = 0.1053, f = 0.462, g = 1.295,
h = -69.6, j = -0.00269, r = 0.998, s = 7.785·105, AICc = 319.46

32