Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add relative and absolute tolerance for rank. #29926

Merged
merged 16 commits into from
Dec 6, 2018
Merged

Add relative and absolute tolerance for rank. #29926

Show file tree
Hide file tree
Changes from 5 commits
Commits
Show all changes
16 commits
Select commit Hold shift + click to select a range
4b66716
Add reltol and abstol in rank
Nov 4, 2018
1730b39
Add reltol and abstol in rank
Nov 4, 2018
0373eee
Add reltol and abstol
Nov 4, 2018
2a22d18
Add space
Nov 4, 2018
a134702
Add rtol,atol to rank
Nov 4, 2018
99b50c8
Add rel and abs tol
Nov 8, 2018
eaa5a65
Add documentation
Nov 8, 2018
40a0b6c
Add rtol, atol
Nov 8, 2018
17bc946
fix test
Nov 9, 2018
d88dd89
Add check for 0 dimension case
Nov 10, 2018
5623b9f
resolve comments
Nov 10, 2018
9bf234b
typo
Nov 10, 2018
d952739
typo
Nov 10, 2018
87cf3ce
deprecate tol in rank
Dec 2, 2018
c9ba74d
remove depwarn
simonbyrne Dec 5, 2018
c5e6a38
Add default values of tolerances and future deprecated method to docs
simonbyrne Dec 5, 2018
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 10 additions & 8 deletions stdlib/LinearAlgebra/src/generic.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -715,10 +715,9 @@ end
rank(A[, tol::Real])

Compute the rank of a matrix by counting how many singular
values of `A` have magnitude greater than `tol*σ₁` where `σ₁` is
`A`'s largest singular values. By default, the value of `tol` is the smallest
dimension of `A` multiplied by the [`eps`](@ref)
of the [`eltype`](@ref) of `A`.
values of `A` have magnitude greater than `rtol*σ₁ + atol` where `σ₁` is
`A`'s largest singular values, atol and rtol are the absolute and relative
sam0410 marked this conversation as resolved.
Show resolved Hide resolved
tolerance respectively.
sam0410 marked this conversation as resolved.
Show resolved Hide resolved

# Examples
```jldoctest
Expand All @@ -728,16 +727,19 @@ julia> rank(Matrix(I, 3, 3))
julia> rank(diagm(0 => [1, 0, 2]))
2

julia> rank(diagm(0 => [1, 0.001, 2]), 0.1)
julia> rank(diagm(0 => [1, 0.001, 2]), rtol=0.1)
2

julia> rank(diagm(0 => [1, 0.001, 2]), 0.00001)
julia> rank(diagm(0 => [1, 0.001, 2]), rtol=0.00001)
3

julia> rank(diagm(0 => [1, 0.001, 2]), atol=0.00001, rtol=0.00001)
sam0410 marked this conversation as resolved.
Show resolved Hide resolved
3
```
"""
function rank(A::AbstractMatrix, tol::Real = min(size(A)...)*eps(real(float(one(eltype(A))))))
function rank(A::AbstractMatrix; atol=0.0, rtol=0.0)
sam0410 marked this conversation as resolved.
Show resolved Hide resolved
s = svdvals(A)
count(x -> x > tol*s[1], s)
count(x -> x > (atol + rtol * s[1]), s)
sam0410 marked this conversation as resolved.
Show resolved Hide resolved
end
rank(x::Number) = x == 0 ? 0 : 1

Expand Down
4 changes: 3 additions & 1 deletion stdlib/LinearAlgebra/test/generic.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -194,7 +194,9 @@ end
end

@test rank(fill(0, 0, 0)) == 0
@test rank([1.0 0.0; 0.0 0.9],0.95) == 1
@test rank([1.0 0.0; 0.0 0.9],rtol=0.95) == 1
@test rank([1.0 0.0; 0.0 0.9],atol=0.95) == 1
@test rank([1.0 0.0; 0.0 0.9],atol=0.95,rtol=0.95)==0
@test qr(big.([0 1; 0 0])).R == [0 1; 0 0]

@test norm([2.4e-322, 4.4e-323]) ≈ 2.47e-322
Expand Down