En http://www.savetman.com/curvefit/index.html puedes aproximar mediante un polinomio de grado entre 1 y 10.
Degree of approximation, m = 4
Coefficient of Term Std Dev
a[0]= 1.44057 - .0136362
a[1]= -8.80058E-07 x 2.03956E-07
a[2]= 1.50948E-12 x2 5.46828E-13
a[3]= -1.02355E-18 x3 4.6168E-19
a[4]= 2.31477E-25 x4 1.19638E-25
Please be careful about the signs of your coefficients!
Re:Polinomio de regresión o red neuronal
(Puntos:3, Informativo)( Última bitácora: Lunes, 29 Octubre de 2012, 18:48h )
Degree of approximation, m = 4
Coefficient of Term Std Dev
a[0]= 1.44057 - .0136362
a[1]= -8.80058E-07 x 2.03956E-07
a[2]= 1.50948E-12 x2 5.46828E-13
a[3]= -1.02355E-18 x3 4.6168E-19
a[4]= 2.31477E-25 x4 1.19638E-25
Please be careful about the signs of your coefficients!
Standard Deviation = 0.0311395
x-value y-value weight y-computed difference
0 1.5 1.0 1.44057 .0594263
4000 1.4521 1.0 1.43708 .0150225
6000 1.44 1.0 1.43535 .00465255
8000 1.428 1.0 1.43363 -0.0056293
8200 1.40 1.0 1.43346 -0.0334581
60000 1.359 1.0 1.39299 -0.0339862
103000 1.340 1.0 1.36485 -0.0248494
200000 1.3165 1.0 1.31712 -0.000623313
400000 1.2914 1.0 1.27049 .0209138
600000 1.2765 1.0 1.26486 .011635
1000000 1.2578 1.0 1.27792 -0.0201233
1500000 1.2428 1.0 1.23418 .00861515
2000000 1.232 1.0 1.2336 -0.00159562
(Lo siendo por el "alineaminto" de las columnas)