Types, methods and dispatch: Incomplete simulate function

[AgustinCB]

While working in the Julia version of the lectures (they're awesome, by the way) in this chapter: https://lectures.quantecon.org/jl/types_methods.html I noticed that the solution of the exercise seems to be incomplete. While the definition of AR(1) includes the use of sigma, the implemented version doesn't:

using Distributions

struct AR1_ex1{T <: Real}
    a::T
    b::T
    sigma::T
    phi::Distribution
end
function simulate(m::AR1_ex1, n::Integer, x0::Real)
    X = Array{Float64}(n)
    X[1] = x0
    for t in 1:(n-1)
        X[t+1] = m.a * X[t] + m.b + rand(m.phi)
    end
    return X
end

Because the test is made using sigma=1, this isn't immediately obvious when looking at the output.

My solution:

struct AR1{T <: Real}          
    a::T
    b::T
    σ::T
    ϕ::Distribution            
end
  
function simulate(m::AR1, n::Integer, x0::Real)
    X = Array{Float64}(n)
    X[1] = x0
    for t in 1:(n-1)
        X[t+1] = m.a * X[t] + m.b + m.σ * rand(m.ϕ)
    end
    return X
end

Let me know if I misunderstood the exercise!

[jstac]

Thanks @AgustinCB !

You're 100% right. I'll put in a fix that will show up over the next week or so.

Your comments also remind me that I want to use unicode variable names --- we held back at first because of issues with typesetting but those issues have been fixed upstream.

Please let us know if you spot more errors!