I´m trying to generalize a regressor vector in my NARX-model function. The regressor vector looks like this:
[-y(i-1) -w(i)*y(i-1) u(i-1) w(i)*u(i-1) u(i-2) w(i)*u(i-2)]
I want to be able to tell my function how many y- and u-values it should consider. So for example if I want to have a regressor vector with 2 y-values and 1 u-value it should generate this:
[-y(i-1) -w(i)*y(i-1) -y(i-2) -w(i)*y(i-2) u(i-1) w(i)*u(i-1)]
And so on for any amount of y- and u-values back in time. Any help is much obliged!
