Sorry, I cannot understand your transfer function. BTW if you are totally new to J, it might be easier to write explicit form first, and make it run correctly before translating it into tacit form. So did you are a working solution in explicit form?
Alan Mac Hugh wrote: > I'm totally new to J, our lecture have this problem , below are its inputs > > ] Input =: 0 1 , 1 0 , 0 0 ,: 1 1 NB. This is how you define > > NB. a matrix in one line > > 0 1 > > 1 0 > > 0 0 > > 1 1 > > Input propagate 10 5 15 30 1 1 NB. The y arg is a random vector > > NB. of weights > > 0.880797 0.880088 0.731059 0.880797 > > where propagate is the transfer function Oj-=1/1 +e^-xi > > where xi = \xAD\xF4 0j Wij > > Oj is the matrix Input. > > This is my attempt at the problem, > ************************************************ > > getInput =: :[ > > getWeights=: :] > > getXi=: +\ (getInput* getWeights) > > d=: 1+ (^ getXi -~) > > propagate=: 1 % d > > ************************************************ > > all i want to to know is I am going about this right, it is a > dyadic function. I don't want any one to solve it ,as that wouldn't be > any > fun. If there some thing wrong just point me to right pace in the > documentation . > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > -- regards, bill ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
