hi,
i am buliding a prediction system.this is how it should work
1)load input data (4 variables) x1:x4 from 12 files as 2-d array.
2)load corresponding single o/p y array(yd)
3)now, i want that there should be feedback from o/p to i/p(e.g 2 o/p feedback(F=2),corresponding to order 2 of A in ARX model),and no input feedback,or delay.
the program should work in fillowing manner:(F=2)
time input desired
t0 x1 x2 x3 x4 y1
t1 x1 x2 x3 x4 y2
as per feedback, i need to give input to'SI Estimate ARX Model ' vi from row F+1. so from row F+1 , train data set is
input desired
t2 x1 x2 x3 x4 y1 y2 y3
t3 x1 x2 x3 x4 y2 y3 y4
and so on.
If i use MISo array, i m using 4 inputs,but size is 4*1,because inputs is row by row,and not a 4* m.so, its still 1-D,while signal order cluster i hv defined dosn't match with this.
if i add a row of zeros, it will accept input, but still the coefficients will be correct? also, if i want to give single input as response i.e for line t2, y3, it still don'st accept it,as ir accepts array,and help says,
that its length should be more than Aorder+B order.
I have attached code, any help?
thnks
