onlineregression12:   One-predictor-variable nonlinear regression   ⇒   y = f(x)

Formulas

y= axb         log(y)= log(a) + b[log(x)]

u= log(x)        v= log(y)         v= log(a) + bu

v= α + βu         log(a)= α        b= β        a= 10α

β= [Σ(uv)-(ΣuΣv)/n]/{Σ(u2)-[(Σu)2/n]}

α= [Σv-(βΣu)]/n        y = 10αxβ

Reference
Return
SAMPLE INPUT DATA:

Array size n:   9

Paired values of arrays y and x (you can cut and paste sample arrays into form):

y:     124,254,378,101,678,540,490,267,350

x:     25,46,78,22,99,89,83,52,73

SAMPLE OUTPUT:

a = 2.88228            b = 1.15428            y = 2.882 x1.154

Thank you for running onlineregression12.   Please call again.   [070417]