v1.0.1

NearestCorrelationMatrix v1.0.1

Diff since v1.0.0

• Fixed autotune for AlternatingProjections
• Fixed signature of trait for Newton
• Fixed init for when an algorithm type is passed in (it was being overridden)

Merged pull requests:

• Fix signatures of specialized methods (#20) (@adknudson)
• fixed init to not override the supplied algtype. updated tests (#21) (@adknudson)
• Bumped version for patch release (#22) (@adknudson)

Closed issues:

• Autotune not working for AlternatingProjections (#17)
• init(prob, algtype) is overriding the algtype (#18)

v1.0.0

NearestCorrelationMatrix v1.0.0

Diff since v0.3.0

Stable API, stable methods! There will be no more changes to the API going forward, as the CommonSolve interface sufficiently allows for changes without breaking any method calls. All algorithms are tested for robustness.

Merged pull requests:

• added Downgrade CI to dev branch (#13) (@adknudson)
• Added JuliaFormatter (#14) (@adknudson)

v0.3.0

NearestCorrelationMatrix v0.3.0

Diff since v0.2.2

• The package now uses the CommonSolve.jl interface. This comes with a host of new ways to set up and solve the nearest correlation matrix problem. See the README for an overview of how to use the new interface.
• nearest_cor and nearest_cor! are still available and will not be removed
• The JuMPAlgorithm is now available as an extension method, and uses the JuMP interface for solving the NCM problem

Breaking changes:

• Dropped support for Julia versions prior to 1.9. They may be added back in the future with more testing
• AlternatingProjection has been renamed to AlternatingProjections
• Keyword parameter τ renamed tau in Newton and DirectProjection algorithms

Merged pull requests:

• CompatHelper: add new compat entry for UnPack at version 1, (keep existing compat) (#6) (@github-actions[bot])

Closed issues:

• Package fails for Julia 1.6 (#4)
• Switch to CommonSolve.jl interface? (#5)
• Silently convert Float16 matrices or fail fast? (#12)

v0.2.2

NearestCorrelationMatrix v0.2.2

Diff since v0.2.1

• Updated doc strings and examples
• Updated tests
• Default tolerance for AlternatingProjection changed to be generally applicable for both Float32 and Float64 matrices

v0.2.1

NearestCorrelationMatrix v0.2.1

Diff since v0.2.0

v0.2.0

NearestCorrelationMatrix v0.2.0

Diff since v0.1.1

v0.1.1

NearestCorrelationMatrix v0.1.1

Diff since v0.1.0

Closed issues:

• Type stability of eigen (#3)

v0.1.0

NearestCorrelationMatrix v0.1.0

Merged pull requests:

• CompatHelper: add new compat entry for SnoopPrecompile at version 1, (keep existing compat) (#1) (@github-actions[bot])