From 6bba7aa3bac619dd99ef9ec973c0fde9f2fb3b96 Mon Sep 17 00:00:00 2001 From: "Anthony D. Blaom" Date: Fri, 19 Jan 2024 08:57:57 +1300 Subject: [PATCH] raw_glm_model -> glm_model --- src/MLJGLMInterface.jl | 26 +++++++++++++------------- test/runtests.jl | 10 +++++----- 2 files changed, 18 insertions(+), 18 deletions(-) diff --git a/src/MLJGLMInterface.jl b/src/MLJGLMInterface.jl index 8a67a49..7a682e3 100644 --- a/src/MLJGLMInterface.jl +++ b/src/MLJGLMInterface.jl @@ -51,10 +51,10 @@ const VALID_KEYS = [ :stderror, :vcov, :coef_table, - :raw_glm_model, + :glm_model, ] const VALID_KEYS_LIST = join(map(k-> "`:$k`", VALID_KEYS), ", ", " and ") -const DEFAULT_KEYS = setdiff(VALID_KEYS, [:raw_glm_model,]) +const DEFAULT_KEYS = setdiff(VALID_KEYS, [:glm_model,]) const KEYS_TYPE = Union{Nothing, AbstractVector{Symbol}} @mlj_model mutable struct LinearRegressor <: MMI.Probabilistic @@ -295,8 +295,8 @@ function glm_report(glm_model, features, reportkeys) end report_dict[:coef_table] = coef_table end - if :raw_glm_model in reportkeys - report_dict[:raw_glm_model] = glm_model + if :glm_model in reportkeys + report_dict[:glm_model] = glm_model end return NamedTuple{Tuple(keys(report_dict))}(values(report_dict)) @@ -602,7 +602,7 @@ Here An offset is a variable which is known to have a coefficient of 1. - `report_keys`: `Vector` of keys for the report. Possible keys are: $VALID_KEYS_LIST. By - default only `:raw_glm_model` is excluded. + default only `:glm_model` is excluded. Train the machine using `fit!(mach, rows=...)`. @@ -646,8 +646,8 @@ When all keys are enabled in `report_keys`, the following fields are available i - `coef_table`: Table which displays coefficients and summarizes their significance and confidence intervals. -- `raw_glm_model`: The raw fitted model returned by `GLM.lm`. Note this points to training - data. +- `glm_model`: The raw fitted model returned by `GLM.lm`. Note this points to training + data. Refer to the GLM.jl documentation for usage. # Examples @@ -729,7 +729,7 @@ Train the machine using `fit!(mach, rows=...)`. the factor used to update the linear fit. - `report_keys`: `Vector` of keys for the report. Possible keys are: $VALID_KEYS_LIST. By - default only `:raw_glm_model` is excluded. + default only `:glm_model` is excluded. # Operations @@ -765,8 +765,8 @@ The fields of `report(mach)` are: - `coef_table`: Table which displays coefficients and summarizes their significance and confidence intervals. -- `raw_glm_model`: The raw fitted model returned by `GLM.glm`. Note this points to training - data. +- `glm_model`: The raw fitted model returned by `GLM.lm`. Note this points to training + data. Refer to the GLM.jl documentation for usage. # Examples @@ -861,7 +861,7 @@ Train the machine using `fit!(mach, rows=...)`. the factor used to update the linear fit. - `report_keys`: `Vector` of keys for the report. Possible keys are: $VALID_KEYS_LIST. By - default only `:raw_glm_model` is excluded. + default only `:glm_model` is excluded. # Operations @@ -898,8 +898,8 @@ The fields of `report(mach)` are: - `coef_table`: Table which displays coefficients and summarizes their significance and confidence intervals. -- `raw_glm_model`: The raw fitted model returned by `GLM.glm`. Note this points to training - data. +- `glm_model`: The raw fitted model returned by `GLM.lm`. Note this points to training + data. Refer to the GLM.jl documentation for usage. # Examples diff --git a/test/runtests.jl b/test/runtests.jl index b63b363..a266fc6 100644 --- a/test/runtests.jl +++ b/test/runtests.jl @@ -326,7 +326,7 @@ end X = (a=[1, 4, 3, 1], b=[2, 0, 1, 4], c=[7, 1, 7, 3]) y = categorical([true, false, true, false]) # check that by default all possible keys are added in the report, - # except raw_glm_model: + # except glm_model: lr = LinearBinaryClassifier() _, _, report = fit(lr, 1, X, y) @test :deviance in keys(report) @@ -334,16 +334,16 @@ end @test :stderror in keys(report) @test :vcov in keys(report) @test :coef_table in keys(report) - @test :raw_glm_model ∉ keys(report) + @test :glm_model ∉ keys(report) # check that report is valid if only some keys are specified - lr = LinearBinaryClassifier(report_keys = [:stderror, :raw_glm_model]) + lr = LinearBinaryClassifier(report_keys = [:stderror, :glm_model]) _, _, report = fit(lr, 1, X, y) @test :deviance ∉ keys(report) @test :stderror in keys(report) @test :dof_residual ∉ keys(report) - @test :raw_glm_model in keys(report) - @test report.raw_glm_model isa GLM.GeneralizedLinearModel + @test :glm_model in keys(report) + @test report.glm_model isa GLM.GeneralizedLinearModel # check that an empty `NamedTuple` is outputed for # `report_params === nothing`