From 599a1dfbbaf342e1efc8ccf90ee2bd7970531490 Mon Sep 17 00:00:00 2001
From: Giuseppe Ragusa <giuseppe.ragusa@gmail.com>
Date: Tue, 15 Oct 2024 20:11:24 +0000
Subject: [PATCH] Delete

---
 src/CR.jl      | 18 +++-------------
 test_sw_mc1.jl | 56 --------------------------------------------------
 2 files changed, 3 insertions(+), 71 deletions(-)
 delete mode 100644 test_sw_mc1.jl

diff --git a/src/CR.jl b/src/CR.jl
index 9751204..2382c98 100644
--- a/src/CR.jl
+++ b/src/CR.jl
@@ -10,18 +10,6 @@ function clusterize(X::Matrix, g::GroupedArray)
     return Symmetric(X2' * X2)
 end
 
-# function mean_by_cluster(X::Matrix, g::GroupedArray)
-#     X2 = zeros(eltype(X), g.ngroups, size(X, 2))
-#     idx = 0
-#     @inbounds for j in axes(X, 2)
-#         idx += 1
-#         @simd for i in axes(X, 1)
-#             X2[g.groups[i], idx] += X[i, j]
-#         end
-#     end
-#     return Symmetric(X2' * X2)
-# end
-
 function avar(k::T, X::Union{Matrix{F},Vector{F}};
               kwargs...) where {T<:Union{CR0,CR1,CR2,CR3},F<:AbstractFloat}
     f = k.g
@@ -45,7 +33,7 @@ end
 nclusters(k::CR) = (map(x -> x.ngroups, k.g))
 rescale!(k::T, S::Matrix) where {T<:CR0} = nothing
 
-function rescale!(k::T, S::Matrix) where {T<:CR1}
+function rescale!(k::T, S::AbstractMatrix) where {T<:CR1}
     # scale total vcov estimate by ((N-1)/(N-K)) * (G/(G-1))
     G = minimum(nclusters(k))
     N = length(k.g[1])
@@ -53,5 +41,5 @@ function rescale!(k::T, S::Matrix) where {T<:CR1}
     @. S *= ((N - 1) / (N - K)) * (G / (G - 1))
 end
 
-rescale!(k::T, S::Matrix) where {T<:CR2} = nothing
-rescale!(k::T, S::Matrix) where {T<:CR3} = nothing
+rescale!(k::T, S::AbstractMatrix) where {T<:CR2} = nothing
+rescale!(k::T, S::AbstractMatrix) where {T<:CR3} = nothing
diff --git a/test_sw_mc1.jl b/test_sw_mc1.jl
deleted file mode 100644
index 42b5b9c..0000000
--- a/test_sw_mc1.jl
+++ /dev/null
@@ -1,56 +0,0 @@
-using CovarianceMatrices
-using LinearAlgebra
-using Random
-using StatsFuns
-## Monte Carlo for the regression model with a single
-## stochastic regressor, in which ut and xt are independent Gaussian
-## AR(1)’s with AR coefficients ρu = ρx = √0.7.
-
-function simulate(T, beta_0, beta_1)
-    # Generate the data
-    rho_u = sqrt(0.7)
-    rho_x = sqrt(0.7)
-    sigma_u = 1
-    sigma_x = 1
-    u = zeros(2*T)
-    x = zeros(2*T)
-    for t = 2:2*T
-        u[t] = rho_u * u[t-1] + sigma_u * randn()
-        x[t] = rho_x * x[t-1] + sigma_x * randn()
-    end
-    y = beta_0 .+ beta_1 .* x + u
-    return y[T+1:end], [ones(T,1) x[T+1:end]]
-end
-
-
-
-T = 600
-ν = floor(Int64, 0.4*T^(2/3))
-reject = zeros(10000)
-
-for j in 1:10000
-
-  y, x = simulate(T, 0, 0)
-  β̂ = x\y
-  û = y - x*β̂
-  z = x.*û
-  Ω̂ = CovarianceMatrices.avar(CovarianceMatrices.EWC(ν), z)
-  Z = x'x
-  Ω̂ᶜʰᵒˡ = cholesky(Ω̂).L
-  V̂ᶜʰᵒˡ = sqrt(T).*(Z\Ω̂ᶜʰᵒˡ)
-  V̂ = V̂ᶜʰᵒˡ*V̂ᶜʰᵒˡ'
-  se = sqrt.(diag(V̂))
-
-  ## sqrt.(diag(T*inv(Z)*Ω̂*inv(Z)))
-
-
-  ## V̂ = cholesky(A)\cholesky(Ω̂).L
-  ## V̂*V̂'
-  ## inv(inv(A)*Ω̂*inv(A))/100^2
-  
-  t = β̂[2] ./ se[2]
-
-  reject[j] = abs(t) >= StatsFuns.tdistinvcdf(ν, .975)
-end
-
-mean(reject)