added tests

examples/mixture_normal_var.jl

@@ -3,7 +3,7 @@ using MarketRisk
 ω = [0.75, 0.25]
 μ = [0.0, -0.1]
 Σ = [0.2, 0.4]
-αs = [0.001, 0.01, 0.05, 0.1]
+αs = 0.001
 h = 10 / 250
 nus = [10., 5.]
 

src/expected_tail_loss.jl

@@ -11,15 +11,15 @@ function ExpectedTailLoss(μ, Σ, h, α, distribution)
  return ParametricExpectedTailLoss(μ, Σ, h, α, distribution)
 end
 
-function compute(etl::ParametricExpectedTailLoss{M, S, H, A, D}) where {M, S, H, A, D <: Normal}
+function compute(etl::ParametricExpectedTailLoss{M, S, H, A, D}) where {M<:Real, S<:Real, H<:Real, A<:Real, D<:Normal}
  @unpack μ, Σ, h, α, dist = etl
 
  xα = quantile(dist, α)
 
  return 1 / α * pdf(dist, xα) * sqrt(h) * Σ - h * μ
 end
 
-function compute(etl::ParametricExpectedTailLoss{M, S, H, A, D}) where {M, S, H, A, D <: Normal}
+function compute(etl::ParametricExpectedTailLoss{M, S, H, A, D}) where {M<:Real, S<:Real, H<:Real, A<:Array{Real}, D <: Normal}
  @unpack μ, Σ, h, α, dist = etl
 
  etl = Float64[]

src/value_at_risk.jl

@@ -48,7 +48,7 @@ function compute(var::ParametricValueAtRisk{V, M, H, A, D}) where {V, M, H<:Real
  return var 
 end
 
-function compute(var::ParametricValueAtRisk{V, M, H, A, MixtureModel{B, D, Normal, F}}) where {V, M, H, A<:Real, B, D, F}
+function compute(var::ParametricValueAtRisk{V, M, H, A, MixtureModel{B, D, Normal, F}}) where {V, M, H, A, B, D, F}
  @unpack μ, Σ, h, α, dist = var 
 
  function opt_func(x, α)
@@ -57,7 +57,7 @@ function compute(var::ParametricValueAtRisk{V, M, H, A, MixtureModel{B, D, Norma
 
  D_opt(f) = x -> ForwardDiff.derivative(f, float(x))
 
- return - find_zero((x -> opt_func(x, αᵢ), D_opt(opt_func)), 0.0, Roots.Newton())
+ return - find_zero((x -> opt_func(x, α), D_opt(x -> opt_func(x, α))), 0.0, Roots.Newton())
 end
 
 function compute(var::ParametricValueAtRisk{V, M, H, A, MixtureModel{B, D, Normal, F}}) where {V, M, H, A<:Array{Float64}, B, D, F}

test/etl.jl

@@ -2,24 +2,29 @@ using MarketRisk
 using Test
 
 # Test normal ETL
-μ_n = 0.0
-Σ_n = 0.3
+μ_en = 0.0
+Σ_en = 0.3
 
-h_n = 10 / 250
-α_n = 0.01
+h_en = 10 / 250
+α_en = 0.01
 
-etl_n = normal_etl(μ_n, Σ_n, h_n, α_n)
+N_en = Normal()
 
-@test etl_n ≈ 0.1599129
+etl_en = ExpectedTailLoss(μ_en, Σ_en, h_en, α_en, N_en)
+
+@test compute(etl_en) ≈ 0.15991285322074647
 
 # Test t-Student ETL
-μ_t = 0.0
-Σ_t = 0.3
+μ_et = 0.0
+Σ_et = 0.3
+
+ν_et = 5
 
-h_t = 10 / 250
-α_t = 0.01
+h_et = 10 / 250
+α_et = 0.01
 
-etl_t = normal_etl(μ_t, Σ_t, h_t, α_t)
+T_et = TDist(ν_et)
 
-@test etl_t ≈ 0.1599129
+etl_et = ExpectedTailLoss(μ_et, Σ_et, h_et, α_et, T_et)
 
+@test compute(etl_et) ≈ 0.23441282083322523

test/runtests.jl

@@ -2,4 +2,3 @@ using Test
 
 @testset "VaR tests" begin include("var.jl") end
 @testset "ETL tests" begin include("etl.jl") end
-

test/var.jl

@@ -17,7 +17,7 @@ N_n = Normal()
 
 var_n = ValueAtRisk(m_n, s_n, h_n, α_n, N_n)
 
-@test compute(var_n) ≈ 1.490889
+@test compute(var_n) ≈ 1.4908892538487946
 
 # Test t-Student VaR
 μ_t = 0.0
@@ -31,7 +31,7 @@ T_t = TDist(ν_t)
 
 var_t = ValueAtRisk(μ_t, Σ_t, h_t, α_t, T_t)
 
-@test compute(var_t) ≈ 0.15638784
+@test compute(var_t) ≈ 0.15638781416305675 
 
 
 # Test mixture normal VaR
@@ -45,17 +45,5 @@ m_nm = MixtureModel(Normal, [(h_nm * μ_nm[i], sqrt(h_nm) * Σ_nm[i]) for i in 1
 
 var_nm = ValueAtRisk(μ_nm, Σ_nm, h_nm, α_nm, m_nm)
 
-@test compute(var_nm) ≈ 0.21616624
+@test compute(var_nm) ≈ 0.21616624418759633
 
-# Test mixture t-Student VaR
-ω_tm = [0.75, 0.25]
-μ_tm = [0.0, - 0.1]
-Σ_tm = [0.2, 0.4]
-α_tm = 0.001
-h_tm = 10 / 250
-
-m_tm = MixtureModel(Normal, [(h_tm * μ_tm[i], sqrt(h_tm) * Σ_tm[i]) for i in 1:2], ω_tm)
-
-var_tm = ValueAtRisk(μ_tm, Σ_tm, h_tm, α_tm, m_tm)
-
-@test compute(var_tm) ≈ 0.3451456