ar and arx seem to work well now, tests have been updated.

src/SystemIdentification.jl

@@ -37,6 +37,13 @@ type AR <: LinearModel
     a::Polynom
     na::Int
 end
+function Base.show(m::AR)
+    print("A(z) = 1")
+    for i = 1:length(m.a)
+        print(m.a[i], " + z^-$i")
+    end
+    print("\n")
+end
 
 """
 `a::Vector{Float64}`: The polynomial coeffs A(z) (not including the first 1)\n

src/armax.jl

@@ -27,7 +27,6 @@ function arx(y::Vector{Float64}, u::VecOrMat{Float64}, na, nb; λ = 0, doplot=fa
     prediction = A*w
     error = y_train - prediction
 
-
     si = na+1
     b = Polynom[w[si:si+nb[1]-1]]
     si += nb[1]

test/tests.jl

@@ -83,6 +83,7 @@ function run_tests()
         model, result = arx(collect(1:5.0),collect(1:5.0),1,1,λ = 1)
         @tassert result.method == :LS_reg
         @tassert result.λ == 1
+        model, result = arx(collect(1:5.0),[collect(2:6.0) ones(5)],2,[2,1])
     end
 end