Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.



Re: Neural network Vs DOE
Posted:
Feb 16, 2013 9:06 PM


Nihad AGHBALOU <nihadagh@gmail.com> wrote in message <85e273106fa342798fd86ed87c727921@googlegroups.com>... > Hello, > i have in learning base 6 colums and 3 line it contrain different values of x1 x2 and x3, for each triple(x1,x2,x3)i have one target y, knowing that the a priori model show is > y=23+19.84*x18.91*x2+4.41*x3 7.53*x1*x2+1.41*x2*x3+5.16*x2*x3+0.66*x1*x2*3 > how can i progress? > i'm realy blocked thearat > thk
Do you mean +1.41*x1*x3 ? and +0.66*x1*x2^3 ?
What do you want to do?
Design a neural net that will approximate the I/O relationship of the polynomial?
Start with the 3x10 input matrix and 1x10 target matrix. Then plug them in to the example for fitnet
help fitnet
doc fitnet.
Hope this helps.
Greg



