format

src/ipod.jl

@@ -446,7 +446,9 @@ From Maronna et al., Robust Statistics: Theory and Methods, Equation 4.49
 """
 location_variance(r::IPODRegression, sqr::Bool=false) = dispersion(r, sqr)
 
-StatsAPI.stderror(r::IPODRegression) = location_variance(r, false) .* sqrt.(abs.(diag(vcov(r))))
+function StatsAPI.stderror(r::IPODRegression)
+    return location_variance(r, false) .* sqrt.(abs.(diag(vcov(r))))
+end
 
 ## Loglikelihood of the full model
 ## l = Σi log fi = Σi log ( 1/(σ * Z) exp( - ρ(ri/σ) ) = -n (log σ + log Z) - Σi ρ(ri/σ)

src/penalties.jl

@@ -382,9 +382,7 @@ function proximal!(
         #            out[i] = zero(T)
         #        else
         x1 = sign(x[i]) * min(λ, max(0, abs(x[i]) - λ / step))
-        x2 =
-            sign(x[i]) *
-            min(λ * γ, max(λ, (abs(x[i]) * (γ - 1) - λ * γ / step) / (γ - 2)))
+        x2 = sign(x[i]) * min(λ * γ, max(λ, (abs(x[i]) * (γ - 1) - λ * γ / step) / (γ - 2)))
         x3 = sign(x[i]) * min(λ * γ, abs(x[i]))
         sols = [x1, x2, x3]
         ind = argmin(convex_approx.(p, sols, x[i], step))

src/regularizedpred.jl

@@ -447,27 +447,20 @@ struct CGDRegPred{T<:BlasReal,M<:AbstractMatrix{T},V<:Vector{T},P<:PenaltyFuncti
         if !isseparable(penalty)
             error(
                 "Coordinate Gradient Descent is only allowed with a separable penalty: " *
-                "$(penalty)"
+                "$(penalty)",
             )
         end
 
         params = CGDParams(outerloop)
 
         return new{T,M,typeof(invvar),P}(
-            X,
-            zeros(T, m),
-            zeros(T, m),
-            Σ,
-            penalty,
-            invvar,
-            params,
-            zeros(T, m),
+            X, zeros(T, m), zeros(T, m), Σ, penalty, invvar, params, zeros(T, m)
         )
     end
 end
 
 function CGDRegPred(
-    X::M, penalty::P, wts::V, outerloop::Bool=true,
+    X::M, penalty::P, wts::V, outerloop::Bool=true
 ) where {M<:AbstractMatrix{T},V<:AbstractVector,P<:PenaltyFunction} where {T<:BlasReal}
     n, m = size(X)
     ll = size(wts, 1)
@@ -525,7 +518,7 @@ function update_βμ!(
     outerM = p.params.outerloop ? m : 1
     update_all = p.params.update_all
 
-#    copyto!(newβ, β)
+    #    copyto!(newβ, β)
     # outer loop
     @inbounds for _ in 1:outerM
 
@@ -539,19 +532,19 @@ function update_βμ!(
 
             # remove component due to index j in μ
             # µ -= X[:, j] * β[j]
-#            if !iszero(βj)
-#                @inbounds @simd for i in eachindex(μ, Xj)
-#                    μ[i] -= Xj[i] * β[j]
-#                end
-##                broadcast!(muladd, μ, -β[j], Xj, μ)
-#            end
+            #            if !iszero(βj)
+            #                @inbounds @simd for i in eachindex(μ, Xj)
+            #                    μ[i] -= Xj[i] * β[j]
+            #                end
+            ##                broadcast!(muladd, μ, -β[j], Xj, μ)
+            #            end
 
             newβ[j] = βj
             if isempty(wts)
                 @inbounds @simd for i in eachindex(axes(X, 1), res, Xj)
                     newβ[j] += Xj[i] * res[i]
-#                @inbounds @simd for i in eachindex(axes(X, 1), y, μ, Xj)
-#                    newβ[j] += Xj[i] * (y[i] - μ[i] + Xj[i] * βj)
+                    #                @inbounds @simd for i in eachindex(axes(X, 1), y, μ, Xj)
+                    #                    newβ[j] += Xj[i] * (y[i] - μ[i] + Xj[i] * βj)
                 end
             else
                 @inbounds @simd for i in eachindex(axes(X, 1), res, Xj, wts)
@@ -564,18 +557,18 @@ function update_βμ!(
             #            verbose && println("∇βj before prox: $(gradβ)")
             # Proximal operator on a single coordinate
             proximal!(pen, newβ, j, newβ, invvar[j] * σ2)
-#            proximal!(pen, β, j, newβ, invvar[j] * σ2)
+            #            proximal!(pen, β, j, newβ, invvar[j] * σ2)
             #            verbose && println("β after prox: $(p.β)")
 
             # re-add component due to index j in μ
             #            μj = p.X[:, j] * p.β[j]
             #            p.µ .+= μj
-##            if !iszero(β[j])
-##                @inbounds @simd for i in eachindex(μ, Xj)
-##                    μ[i] += Xj[i] * β[j]
-##                end
-###                broadcast!(muladd, μ, β[j], view(X, :, j), μ)
-##            end
+            ##            if !iszero(β[j])
+            ##                @inbounds @simd for i in eachindex(μ, Xj)
+            ##                    μ[i] += Xj[i] * β[j]
+            ##                end
+            ###                broadcast!(muladd, μ, β[j], view(X, :, j), μ)
+            ##            end
             Δβ = newβ[j] - βj
             if !iszero(Δβ)
                 β[j] = newβ[j]
@@ -630,9 +623,7 @@ mutable struct FISTAParams{T<:AbstractFloat}
     function FISTAParams{T}(
         restart::Bool, use_backtracking::Bool, bt_maxiter::Integer=20, bt_delta::Real=0.5
     ) where {T<:AbstractFloat}
-        return new{T}(
-            restart, use_backtracking, bt_maxiter, bt_delta, one(T), one(T), 1,
-        )
+        return new{T}(restart, use_backtracking, bt_maxiter, bt_delta, one(T), one(T), 1)
     end
 end
 
@@ -913,7 +904,7 @@ struct AMARegPred{
     T<:BlasReal,
     M<:AbstractMatrix{T},
     V<:Vector{T},
-#    C,
+    #    C,
     P<:PenaltyFunction,
     M2<:AbstractMatrix{T},
     V2<:AbstractVector{T},
@@ -932,8 +923,8 @@ struct AMARegPred{
     scratchbeta::V
     "`penbeta0`: vector of length `p`, used in [`penalized_coef`](@ref) method"
     penbeta0::V
-#    "`chol`: cholesky factorization"
-#    chol::C
+    #    "`chol`: cholesky factorization"
+    #    chol::C
     "`wXt`: transpose of the (weighted) model matrix"
     wXt::M
     "`A`: matrix of the constraint equation `A . u  - b - v = 0`"
@@ -961,12 +952,12 @@ struct AMARegPred{
     ) where {M<:AbstractMatrix{T},P<:PenaltyFunction} where {T<:BlasReal}
         m = size(Σ, 1)
 
-#        chol = cholesky(Σ)
+        #        chol = cholesky(Σ)
         bhat = A * b
         params = AMAParams{T}(restart)
 
         beta0 = zeros(T, m)
-#        return new{T,M,typeof(beta0),typeof(chol),P,typeof(A),typeof(b)}(
+        #        return new{T,M,typeof(beta0),typeof(chol),P,typeof(A),typeof(b)}(
         return new{T,M,typeof(beta0),P,typeof(A),typeof(b)}(
             X,
             beta0,
@@ -975,7 +966,7 @@ struct AMARegPred{
             penalty,
             zeros(T, m),
             zeros(T, m),
-#            chol,
+            #            chol,
             wXt,
             A,
             bhat,
@@ -1055,7 +1046,7 @@ function update_beta!(
     wXt = p.wXt
     βkp1 = p.delbeta
     scratch = p.scratchbeta
-#    chol = p.chol
+    #    chol = p.chol
     A = p.A
     b = p.b
 
@@ -1085,7 +1076,7 @@ function update_beta!(
     # βkp1 = chol \ (wXt * wrky - ρ * σ2 * A' * whatk)
     scratch = mul!(scratch, A', whatk)
     scratch = mul!(scratch, wXt, y, 1, -ρ * σ2)
-#    βkp1 = chol \ scratch
+    #    βkp1 = chol \ scratch
     cg!(βkp1, p.Σ, scratch)
     # vkp1 = proximal(penalty(p), A * βkp1 - b + whatk, 1/ρ)
     scratch = mul!(copyto!(scratch, b), A, βkp1, 1, -1)
@@ -1217,7 +1208,7 @@ struct ADMMRegPred{
     T<:BlasReal,
     M<:AbstractMatrix{T},
     V<:Vector{T},
-#    C,
+    #    C,
     P<:PenaltyFunction,
     M2<:AbstractMatrix{T},
     V2<:AbstractVector{T},
@@ -1236,8 +1227,8 @@ struct ADMMRegPred{
     scratchbeta::V
     "`penbeta0`: vector of length `p`, used in [`penalized_coef`](@ref) method"
     penbeta0::V
-#    "`chol`: cholesky factorization"
-#    chol::C
+    #    "`chol`: cholesky factorization"
+    #    chol::C
     "`wXt`: transpose of the (weighted) model matrix"
     wXt::M
     "`A`: matrix of the constraint equation `A . u  - b - v = 0`"
@@ -1269,12 +1260,12 @@ struct ADMMRegPred{
     ) where {M<:AbstractMatrix{T},P<:PenaltyFunction} where {T<:BlasReal}
         m = size(Σ, 1)
 
-#        chol = cholesky(sparse(Σ))
+        #        chol = cholesky(sparse(Σ))
         bhat = A * b
         params = ADMMParams{T}(restart, adapt)
 
         beta0 = zeros(T, m)
-#        return new{T,M,typeof(beta0),typeof(chol),P,typeof(A),typeof(b)}(
+        #        return new{T,M,typeof(beta0),typeof(chol),P,typeof(A),typeof(b)}(
         return new{T,M,typeof(beta0),P,typeof(A),typeof(b)}(
             X,
             beta0,
@@ -1283,7 +1274,7 @@ struct ADMMRegPred{
             penalty,
             zeros(T, m),
             zeros(T, m),
-#            chol,
+            #            chol,
             wXt,
             A,
             bhat,
@@ -1408,7 +1399,7 @@ function update_beta!(
     broadcast!(+, βkp1, broadcast!(-, βkp1, vhatk, whatk), b)  # dumb βkp1
     mul!(scratch, A', βkp1)  # dumb βkp1
     mul!(scratch, wXt, y, 1, ρ * σ2)
-#    βkp1 = facΣρ \ scratch
+    #    βkp1 = facΣρ \ scratch
     cg!(βkp1, wrkΣ, scratch)
     # proximal!(penalty(p), vkp1, A * βkp1 - b + whatk, 1/ρ)
     scratch = mul!(copyto!(scratch, b), A, βkp1, 1, -1)
@@ -1457,8 +1448,8 @@ function updatepred!(p::ADMMRegPred, σ::Real; verbose::Bool=false, force::Bool=
 
         # Refactorize
         # TODO: should be improved
-#        cholesky!(p.chol, sparse(Σ + ρ * σ^2 * Hermitian(A'A)))
-#        p.wrkΣ = Hermitian(p.Σ + σ2 * ρ * A'A)
+        #        cholesky!(p.chol, sparse(Σ + ρ * σ^2 * Hermitian(A'A)))
+        #        p.wrkΣ = Hermitian(p.Σ + σ2 * ρ * A'A)
         copyto!(p.wrkΣ, p.Σ)
         mul!(p.wrkΣ, A', A, ρ * σ^2, 1)
         copyto!(p.wrkΣ, Hermitian(p.wrkΣ))

src/robustlinearmodel.jl

@@ -1665,7 +1665,13 @@ function StatsAPI.fit!(
     m::RobustLinearModel{T,R,P};
     correct_leverage::Bool=false,
     updatescale::Bool=false,
-    maxiter::Integer=(P<:CGDRegPred) ? 10_000 : (P<:FISTARegPred) ? 1_000 : 100,
+    maxiter::Integer=if (P <: CGDRegPred)
+        10_000
+    elseif (P <: FISTARegPred)
+        1_000
+    else
+        100
+    end,
     minstepfac::Real=1e-3,
     atol::Real=1e-8,
     rtol::Real=1e-7,

test/data/Animals2.jl

@@ -214,5 +214,4 @@ sX = SparseMatrixCSC(X)
 y = data.logBrain
 nt = (; logBrain=data.logBrain, logBody=data.logBody)
 
-data_tuples = ((form, data), (form, nt), (X, y), (sX, y))
-;
+data_tuples = ((form, data), (form, nt), (X, y), (sX, y));

test/data/starsCYG.jl

@@ -111,5 +111,4 @@ sX2 = SparseMatrixCSC(X2)
 y2 = data2.logLight
 nt2 = (; logTemp=logTemp, logLight=logLight)
 
-data2_tuples = ((form2, data2), (form2, nt2), (X2, y2), (sX2, y2))
-;
+data2_tuples = ((form2, data2), (form2, nt2), (X2, y2), (sX2, y2));

test/ipod.jl

@@ -91,8 +91,13 @@ end
             @testset "data type: $(typeof(A))" for (A, b) in ((X, y), (sX, y))
                 name = "Θ-IPOD($(l), $(pen); method=$(method)),\t"
                 name *= (
-                    (A isa FormulaTerm) ? "formula" :
-                    (A isa SparseMatrixCSC) ? "sparse " : "dense  "
+                    if (A isa FormulaTerm)
+                        "formula"
+                    elseif (A isa SparseMatrixCSC)
+                        "sparse "
+                    else
+                        "dense  "
+                    end
                 )
 
                 m0 = fit(RobustLinearModel, A, b, est; method=:chol, initial_scale=σ)
@@ -130,7 +135,7 @@ end
     def_rtol = 1e-5
 
     kwargs = (; initial_scale=1, maxiter=10_000)
-    
+
     @testset "solver method $(method)" for method in pen_methods
         rtol = def_rtol
         if method === :cgd
@@ -145,8 +150,13 @@ end
         @testset "data type: $(typeof(A))" for (A, b) in data_tuples
             name = "Θ-IPOD(L2Loss, $(pen2); method=$(method)),\t"
             name *= (
-                (A isa FormulaTerm) ? "formula" :
-                (A isa SparseMatrixCSC) ? "sparse " : "dense  "
+                if (A isa FormulaTerm)
+                    "formula"
+                elseif (A isa SparseMatrixCSC)
+                    "sparse "
+                else
+                    "dense  "
+                end
             )
 
             m2 = ipod(A, b, loss1, pen2; method=method, kwargs...)
@@ -208,8 +218,13 @@ end
     @testset "Omit intercept, data type: $(typeof(A))" for (A, b) in data_tuples
         name = "Θ-IPOD(L2Loss, $(pen); method=:auto),\t"
         name *= (
-            (A isa FormulaTerm) ? "formula" :
-            (A isa SparseMatrixCSC) ? "sparse " : "dense  "
+            if (A isa FormulaTerm)
+                "formula"
+            elseif (A isa SparseMatrixCSC)
+                "sparse "
+            else
+                "dense  "
+            end
         )
 
         m1 = fit(IPODRegression, A, b, loss1, pen; kwargs...)
@@ -255,8 +270,13 @@ end
         @testset "data type: $(typeof(A))" for (A, b) in data_tuples
             name = "Θ-IPOD(L2Loss, $(pen); method=$(method)),\t"
             name *= (
-                (A isa FormulaTerm) ? "formula" :
-                (A isa SparseMatrixCSC) ? "sparse " : "dense  "
+                if (A isa FormulaTerm)
+                    "formula"
+                elseif (A isa SparseMatrixCSC)
+                    "sparse "
+                else
+                    "dense  "
+                end
             )
 
             m1 = fit(IPODRegression, A, b, loss1, pen; method=method, kwargs...)

test/penalties.jl

@@ -135,8 +135,13 @@ t = 10 * randn(rng, 30)
             @testset "data type: $(typeof(A))" for (A, b) in data2_tuples
                 name = "rlm($(pen); method=$(method)),\t"
                 name *= (
-                    (A isa FormulaTerm) ? "formula" :
-                    (A isa SparseMatrixCSC) ? "sparse " : "dense  "
+                    if (A isa FormulaTerm)
+                        "formula"
+                    elseif (A isa SparseMatrixCSC)
+                        "sparse "
+                    else
+                        "dense  "
+                    end
                 )
 
                 m2 = rlm(A, b, pen; method=method, kwargs...)
@@ -166,8 +171,8 @@ t = 10 * randn(rng, 30)
                         @test all(abs.(var2) .>= abs.(var1))
                     elseif f in (leverage_weights,)
                         @test all(abs.(var2) .<= abs.(var1))
-                    # elseif f in (confint,)
-                    #     @test isapprox(var1, var2; rtol=1e-1)
+                        # elseif f in (confint,)
+                        #     @test isapprox(var1, var2; rtol=1e-1)
                     elseif f in (confint, projectionmatrix)
                         continue
                     else