System Identification Polynomial Estimation for New Inputs

I'm trying to model a black box system after being given an array of input data and an array of output data. I've been trying to use Matlab's System Identification Toolbox to find the polynomial model of the system.

I can use the command line to find the model by using "model = arx(data,[3 3 1])". Or I can get the same information with graphs using the toolbox.

I can find the model with it's Discrete-time ARX model: A(z)y(t) = B(z) u(t) +e(t) and it's A(z) and B(z).

What I don't know, is how I get the numerical predicted output values from the predicted model. I also don't know how to use new inputs in the predicted model to get new outputs. Can I display the plots that I can find from using the toolbox by using the command line instead?