add convenience functions for ARMA and bump version

Project.toml

@@ -1,7 +1,7 @@
 name = "ARCHModels"
 uuid = "6d3278bc-c23a-5105-85e5-0d57d2bf684f"
 authors = ["Simon Broda <simon.broda@alumni.ethz.ch> and contributors."]
-version = "2.1"
+version = "2.2"
 
 [deps]
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"

docs/src/usage.md

@@ -324,7 +324,7 @@ Volatility parameters:
 ω  0.220848  0.0118061  18.7063    <1e-77
 ─────────────────────────────────────────
 ```
-In this case, an ARMA(1, 1) specification was selected.
+In this case, an ARMA(1, 1) specification was selected. As a convenience, the above can equivalently achieved using `selectmodel(ARMA, BG96)`.
 
 As a final example, a construction like the following can be used to automatically select not just the lag length, but also the class of GARCH model and the error distribution:
 

src/meanspecs.jl

@@ -101,7 +101,6 @@ end
     ARMA{p, q, T} <: MeanSpec{T}
 An ARMA(p, q) mean specification.
 """
-
 struct ARMA{p, q, T} <: MeanSpec{T}
     coefs::Vector{T}
     function ARMA{p, q, T}(coefs::Vector) where {p, q, T}
@@ -110,6 +109,29 @@ struct ARMA{p, q, T} <: MeanSpec{T}
     end
 end
 
+"""
+    fit(t::Type{<:ARMA}, data; kwargs...) -> UnivariateARCHModel
+
+Fit an `ARMA{p, q}` model to `data`.
+"""
+fit(t::Type{<:ARMA}, data; kwargs...) = fit(ARCH{0}, data; meanspec=t, kwargs...)
+
+"""
+    selectmodel(::Type{<:ARMA}, data; kwargs...)  -> UnivariateARCHModel
+
+Fit a number of `ARMA{p, q}` models to `data` and return that which
+minimizes the [BIC](https://en.wikipedia.org/wiki/Bayesian_information_criterion).
+
+# Keyword arguments:
+- `dist=StdNormal`: the error distribution.
+- `minlags=1`: minimum lag length to try in each parameter of `VS`.
+- `maxlags=3`: maximum lag length to try in each parameter of `VS`.
+- `criterion=bic`: function that takes a `UnivariateARCHModel` and returns the criterion to minimize.
+- `show_trace=false`: print `criterion` to screen for each estimated model.
+- `algorithm=BFGS(), autodiff=:forward, kwargs...`: passed on to the optimizer.
+"""
+selectmodel(t::Type{<:ARMA}, data; kwargs...) = selectmodel(ARCH{0}, data; meanspec=t, kwargs...)
+
 """
     ARMA{p, q}(coefs::Vector)
 Create an ARMA(p, q) model.

test/runtests.jl

@@ -224,6 +224,8 @@ end
     @test all(isapprox(coef(am), coef(fit(GARCH{1, 1}, BG96, meanspec=Intercept)), rtol=1e-4))
     @test coefnames(am)[end] == "X"
     @test all(isapprox(coef(am), coef(fit(GARCH{1, 1}, model.model)), rtol=1e-4))
+    @test sum(coef(fit(ARMA{1, 1}, BG96))) ≈ 0.21595383060382695
+    @test sum(coef(selectmodel(ARMA, BG96; minlags=2, maxlags=3))) ≈ 0.25404112526031397
 end
 
 @testset "VaR" begin