add more test for robust hat matrix based regression estimator

Author: jbytecode

CHANGELOG.md

@@ -1,8 +1,10 @@
 # v0.11.5 (Upcoming Release)
 
 - Initial implementation of the robust hat matrix regression estimator
+- Add more test to robust hat matrix regression estimator
 - Introduce `view` in LTS and LMS.
 
+
 # v0.11.4 
 
 - More explicit return types, drop `Dict` with `Dict{String, Any}` or `Dict{String, Vector}`

src/robhatreg.jl

@@ -25,13 +25,12 @@ function m(mat::AbstractMatrix{T}, u::AbstractVector{T})::AbstractMatrix where {
     for i in 1:L
         y[i, 1] = u[i]
     end
-    result = m(mat, y)
-    return result
+    return m(mat, y)
 end
 
 function m(m1::AbstractMatrix{T}, m2::AbstractMatrix{T})::AbstractMatrix where {T<:Real}
-    n1, _ = size(m1)
-    _, p2 = size(m2)
+    n1 = size(m1, 1)
+    p2 = size(m2, 2)
     newmat = zeros(Float64, n1, p2)
     for i in 1:n1
         for j in 1:p2

test/testrobhatreg.jl

@@ -1,34 +1,74 @@
 
+@testset "Robust Hat Matrix based Robust Regression" verbose = true begin
 
-@testset "Robust Hat Matrix based Robust Regression" begin
-    # Create simple data
-    rng = MersenneTwister(12345)
-    n = 50
-    x = collect(1:n)
-    e = randn(rng, n) .* 2.0
-    y = 5 .+ 5 .* x .+ e
+    @testset "Random data" begin
+        # Create simple data
+        rng = MersenneTwister(12345)
+        n = 50
+        x = collect(1:n)
+        e = randn(rng, n) .* 2.0
+        y = 5 .+ 5 .* x .+ e
 
-    # Contaminate some values 
-    y[n] = y[n] * 2.0
-    y[n-1] = y[n-1] * 2.0
-    y[n-2] = y[n-2] * 2.0
-    y[n-3] = y[n-3] * 2.0
-    y[n-4] = y[n-4] * 2.0
+        # Contaminate some values 
+        y[n] = y[n] * 2.0
+        y[n-1] = y[n-1] * 2.0
+        y[n-2] = y[n-2] * 2.0
+        y[n-3] = y[n-3] * 2.0
+        y[n-4] = y[n-4] * 2.0
 
-    df = DataFrame(x=x, y=y)
+        df = DataFrame(x=x, y=y)
 
-    reg = createRegressionSetting(@formula(y ~ x), df)
-    result = robhatreg(reg)
+        reg = createRegressionSetting(@formula(y ~ x), df)
+        result = robhatreg(reg)
 
-    betas = result["betas"]
+        betas = result["betas"]
 
-    atol = 1.0
+        atol = 1.0
 
-    @test isapprox(betas[1], 5.0, atol=atol)
-    @test isapprox(betas[2], 5.0, atol=atol)
-end
+        @test isapprox(betas[1], 5.0, atol=atol)
+        @test isapprox(betas[2], 5.0, atol=atol)
+    end
+
+    @testset "Phone data" begin
+        df = phones
+        reg = createRegressionSetting(@formula(calls ~ year), df)
+        result = robhatreg(reg)
+
+        betas = result["betas"]
+
+        atol = 0.001
 
+        @test isapprox(betas[1], -54.967349441923226, atol=atol)
+        @test isapprox(betas[2], 1.1406353489513064, atol=atol)
+    end
 
+    @testset "Large Data" begin
+        X = randn(10000, 10)
+        y = randn(10000)
 
+        result = robhatreg(X, y)
 
+        betas = result["betas"]
+
+        atol = 0.1
+
+        for i in 1:10
+            @test isapprox(betas[i], 0.0, atol=atol)
+        end
+    end
+
+    @testset "Single Y outlier" begin
+        @testset "LAD - Algorithm - Exact" begin
+            df2 = DataFrame(
+                x=[1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
+                y=[2, 4, 6, 8, 10, 12, 14, 16, 18, 1000],
+            )
+            reg2 = createRegressionSetting(@formula(y ~ x), df2)
+            result2 = lad(reg2)
+            betas2 = result2["betas"]
+            @test betas2[1] == 0.0
+            @test betas2[2] == 2.0
+        end
+    end
+end