-
Notifications
You must be signed in to change notification settings - Fork 35
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Add methods for calling multi-response gaussian family (#66)
Fixes #65
- Loading branch information
1 parent
47da7a0
commit 4167f0c
Showing
19 changed files
with
1,111 additions
and
10 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,188 @@ | ||
| import Distributions.MvNormal | ||
|
|
||
| MvNormal() = MvNormal([0, 0], [1 0; 0 1]) # MvNormal(0, ([1;])) | ||
| modeltype(::MvNormal) = "MvNormal" | ||
|
|
||
|
|
||
| function predict(path::GLMNetPath{<:MvNormal}, X::AbstractMatrix, | ||
| model::Union{Int, AbstractVector{Int}}=1:length(path.lambda); | ||
| outtype = :link, offsets = zeros(size(X, 1), size(path.betas, 2))) | ||
|
|
||
| nresp = size(path.betas, 2); | ||
| out = zeros(Float64, size(X, 1), nresp, length(model)); | ||
| for i = 1:length(model) | ||
| out[:, :, i] = repeat(path.a0[:,model[i]]', size(X, 1)) + X * path.betas[:, :, model[i]] + offsets | ||
| end | ||
| if outtype != :link | ||
| for i = 1:size(X, 1), j = 1:length(model) | ||
| out = exp.(out) | ||
| end | ||
| end | ||
| if length(model) == 1 | ||
| return out[:, :, 1] | ||
| else | ||
| return out | ||
| end | ||
| end | ||
|
|
||
| nactive(g::GLMNetPath{<:MvNormal}, b::AbstractVector{Int}=1:size(g.betas, 3)) = | ||
| [nactive(g.betas, j, dims=2) for j in b] | ||
|
|
||
| function show(io::IO, g::GLMNetPath{<:MvNormal}) | ||
| println(io, "$(modeltype(g.family)) GLMNet Solution Path ($(size(g.betas, 3)) solutions for $(size(g.betas, 1)) predictors in $(g.npasses) passes):") | ||
| print(io, DataFrame(df=nactive(g), pct_dev=g.dev_ratio, λ=g.lambda)) | ||
| end | ||
|
|
||
| struct MultiMSE <: Loss | ||
| y::Matrix{Float64} | ||
| end | ||
| loss(l::MultiMSE, i, mu) = sum(abs2.(l.y[i,:] .- mu)) | ||
|
|
||
| devloss(::MvNormal, y) = MultiMSE(y) | ||
|
|
||
| function loss(path::GLMNetPath{<:MvNormal}, X::AbstractMatrix{Float64}, | ||
| y::Union{AbstractVector{Float64}, AbstractMatrix{Float64}}, | ||
| weights::AbstractVector{Float64}=ones(size(y,1)), | ||
| lossfun::Loss=devloss(path.family, y), | ||
| model::Union{Int, AbstractVector{Int}}=1:length(path.lambda); | ||
| offsets = zeros(size(X, 1), size(path.betas, 2))) | ||
|
|
||
| validate_x_y_weights(X, y, weights) | ||
| mu = predict(path, X; offsets = offsets) | ||
| devs = zeros(size(mu, 3)) | ||
| for j = 1:size(mu, 3), i = 1:size(mu, 1) | ||
| devs[j] += loss(lossfun, i, vec(mu[i, :, j]))*weights[i] | ||
| end | ||
| devs/sum(weights) | ||
| end | ||
|
|
||
| loss(path::GLMNetPath{<:MvNormal}, X::AbstractMatrix, y::Union{AbstractVector, AbstractMatrix}, | ||
| weights::AbstractVector=ones(size(y,1)), va...; kw...) = | ||
| loss(path, convert(Matrix{Float64}, X), | ||
| convert(Array{Float64}, y), | ||
| convert(Vector{Float64}, weights), va...; kw...) | ||
|
|
||
|
|
||
|
|
||
| macro check_and_return_mvnormal() | ||
| esc(quote | ||
| check_jerr(jerr[], maxit,pmax) | ||
| lmu = lmu_ref[] | ||
| # first lambda is infinity; changed to entry point | ||
| if isempty(lambda) && length(alm) > 2 | ||
| alm[1] = exp(2*log(alm[2])-log(alm[3])) | ||
| end | ||
| GLMNetPath(family, a0[:, 1:lmu], ca[sortperm(ia), :, 1:lmu], | ||
| null_dev[], fdev[1:lmu], alm[1:lmu], Int(nlp[])) | ||
| end) | ||
| end | ||
|
|
||
| # change of parameters from elnet to multelnet | ||
| # ka,parm,no,ni,nr, x,y,w,jd, vp,cl,ne,nx, nlam,flmin,ulam,thr, isd, ,intr,maxit,lmu, a0,ca,ia,nin, rsq,alm,nlp,jerr | ||
| # parm,no,ni,nr, x,y,w,jd, vp,cl,ne,nx, nlam,flmin,ulam,thr, isd,jsd,intr,maxit,lmu, a0,ca,ia,nin, rsq,alm,nlp,jerr | ||
| # multi-response normal | ||
| function glmnet!(X::Matrix{Float64}, y::Matrix{Float64}, | ||
| family::MvNormal=MvNormal(); | ||
| weights::Vector{Float64}=ones(size(y,1)), | ||
| naivealgorithm::Bool=(size(X, 2) >= 500), alpha::Real=1.0, | ||
| penalty_factor::Vector{Float64}=ones(size(X, 2)), | ||
| constraints::Array{Float64, 2}=[x for x in (-Inf, Inf), y in 1:size(X, 2)], | ||
| dfmax::Int=size(X, 2), pmax::Int=min(dfmax*2+20, size(X, 2)), nlambda::Int=100, | ||
| lambda_min_ratio::Real=(size(y, 1) < size(X, 2) ? 1e-2 : 1e-4), | ||
| lambda::Vector{Float64}=Float64[], tol::Real=1e-7, standardize::Bool=true, | ||
| standardize_response::Bool=false, | ||
| intercept::Bool=true, maxit::Int=1000000) | ||
|
|
||
| @validate_and_init_multi | ||
| standardize_response = Int32(standardize_response) | ||
|
|
||
| # Compute null deviance | ||
| yw = y .* repeat(weights, 1, size(y, 2)) | ||
| mu = mean(y, dims=1) | ||
| if intercept == 0 | ||
| mu = fill(intercept, 1, size(y,2)) | ||
| end | ||
| # Sum of squared error (weighted by obervation weights) | ||
| null_dev[] = sum(weights) .* mean(abs2.(yw .- mu)) | ||
|
|
||
|
|
||
| ccall((:multelnet_, libglmnet), Cvoid, | ||
|
|
||
| (Ref{Float64}, Ref{Int32}, Ref{Int32}, Ref{Int32}, | ||
| Ref{Float64}, Ref{Float64}, Ref{Float64}, Ref{Int32}, | ||
| Ref{Float64}, Ref{Float64}, Ref{Int32}, Ref{Int32}, | ||
| Ref{Int32}, Ref{Float64}, Ref{Float64}, Ref{Float64}, | ||
| Ref{Int32}, Ref{Int32}, Ref{Int32}, Ref{Int32}, Ref{Int32}, | ||
| Ref{Float64}, Ref{Float64}, Ref{Int32}, Ref{Int32}, | ||
| Ref{Float64}, Ref{Float64}, Ref{Int32}, Ref{Int32}), | ||
|
|
||
| alpha, nobs, nvars, nresp, | ||
| X, y, weights, jd, | ||
| penalty_factor, constraints, dfmax, pmax, | ||
| nlambda, lambda_min_ratio, lambda, tol, | ||
| standardize, standardize_response, intercept, maxit, lmu_ref, | ||
| a0, ca, ia, nin, | ||
| fdev, alm, nlp, jerr | ||
| ) | ||
|
|
||
| @check_and_return_mvnormal | ||
| end | ||
|
|
||
|
|
||
|
|
||
| # multi-response sparse normal | ||
| function glmnet!(X::AbstractSparseMatrix{Float64}, y::Matrix{Float64}, | ||
| family::MvNormal=MvNormal(); | ||
| weights::Vector{Float64}=ones(size(y, 1)), | ||
| naivealgorithm::Bool=(size(X, 2) >= 500), alpha::Real=1.0, | ||
| penalty_factor::Vector{Float64}=ones(size(X, 2)), | ||
| constraints::Array{Float64, 2}=[x for x in (-Inf, Inf), y in 1:size(X, 2)], | ||
| dfmax::Int=size(X, 2), pmax::Int=min(dfmax*2+20, size(X, 2)), nlambda::Int=100, | ||
| lambda_min_ratio::Real=(size(y, 1) < size(X, 2) ? 1e-2 : 1e-4), | ||
| lambda::Vector{Float64}=Float64[], tol::Real=1e-7, standardize::Bool=true, | ||
| standardize_response::Bool=false, | ||
| intercept::Bool=true, maxit::Int=1000000) | ||
|
|
||
| @validate_and_init_multi | ||
| standardize_response = Int32(standardize_response) | ||
|
|
||
| # Compute null deviance | ||
| yw = y .* repeat(weights, 1, size(y, 2)) | ||
| mu = mean(y, dims=1) | ||
| if intercept == 0 | ||
| mu = fill(intercept, 1, size(y,2)) | ||
| end | ||
| # Sum of squared error (weighted by obervation weights) | ||
| null_dev[] = sum(weights) .* mean(abs2.(yw .- mu)) | ||
|
|
||
|
|
||
| ccall((:multspelnet_, libglmnet), Cvoid, | ||
|
|
||
| (Ref{Float64}, Ref{Int32}, Ref{Int32}, Ref{Int32}, | ||
| Ref{Float64}, Ref{Int32}, Ref{Int32}, Ref{Float64}, Ref{Float64}, Ref{Int32}, | ||
| Ref{Float64}, Ref{Float64}, Ref{Int32}, Ref{Int32}, | ||
| Ref{Int32}, Ref{Float64}, Ref{Float64}, Ref{Float64}, | ||
| Ref{Int32}, Ref{Int32}, Ref{Int32}, Ref{Int32}, Ref{Int32}, | ||
| Ref{Float64}, Ref{Float64}, Ref{Int32}, Ref{Int32}, | ||
| Ref{Float64}, Ref{Float64}, Ref{Int32}, Ref{Int32}), | ||
|
|
||
| alpha, nobs, nvars, nresp, | ||
| X.nzval, X.colptr, X.rowval, y, weights, jd, | ||
| penalty_factor, constraints, dfmax, pmax, | ||
| nlambda, lambda_min_ratio, lambda, tol, | ||
| standardize, standardize_response, intercept, maxit, lmu_ref, | ||
| a0, ca, ia, nin, | ||
| fdev, alm, nlp, jerr | ||
| ) | ||
|
|
||
| @check_and_return_mvnormal | ||
| end | ||
|
|
||
| glmnet(X::Matrix{Float64}, y::Matrix{Float64}, family::MvNormal; kw...) = | ||
| glmnet!(copy(X), copy(y), family; kw...) | ||
| glmnet(X::SparseMatrixCSC{Float64,Int32}, y::Matrix{Float64}, family::MvNormal; kw...) = | ||
| glmnet!(copy(X), copy(y), family; kw...) | ||
| glmnet(X::AbstractMatrix, y::AbstractMatrix{<:Number}, family::MvNormal; kw...) = | ||
| glmnet(convert(Matrix{Float64}, X), convert(Matrix{Float64}, y), family; kw...) | ||
| glmnet(X::SparseMatrixCSC, y::AbstractMatrix{<:Number}, family::MvNormal; kw...) = | ||
| glmnet(convert(SparseMatrixCSC{Float64,Int32}, X), convert(Matrix{Float64}, y), family; kw...) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
4167f0cThere was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@JuliaRegistrator register
4167f0cThere was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Registration pull request created: JuliaRegistries/General/114950
Tip: Release Notes
Did you know you can add release notes too? Just add markdown formatted text underneath the comment after the text
"Release notes:" and it will be added to the registry PR, and if TagBot is installed it will also be added to the
release that TagBot creates. i.e.
To add them here just re-invoke and the PR will be updated.
Tagging
After the above pull request is merged, it is recommended that a tag is created on this repository for the registered package version.
This will be done automatically if the Julia TagBot GitHub Action is installed, or can be done manually through the github interface, or via: