v0.2.3

v0.2.3 (2019-12-26)

Diff since v0.2.2

Closed issues:

• Update to MLJBase 0.9 (#40)

Merged pull requests:

• fix default solver for robust regression, MLJ issue 401 (#44) (tlienart)

v0.2.2

v0.2.2 (2019-12-19)

Diff since v0.2.1

Closed issues:

• RidgeCV, can also use Hessenberg factorisation (#39)
• Bump [compat] MLJBase="^0.8.1" (#37)

Merged pull requests:

• updating MLJBase bound and tests (#42) (tlienart)

v0.2.1

v0.2.1 (2019-11-19)

Diff since v0.2.0

v0.2.0

v0.2.0 (2019-10-29)

Diff since v0.1.0

Merged pull requests:

• Interface with MLJ (#34) (tlienart)

v0.1.0

v0.1.0 (2019-09-28)

Diff since 7681b664c2783d7df31a1b9e24dec282f5b3a7fa

Closed issues:

• Add option to toggle penalty of intercept term (#27)
• Add precise tests for all d* functions (#19)
• Reduce allocations more by keeping a cache (#18)
• Tests LAD/Quantile against R package (#12)
• Either remove or test (F)ADMM on simpler problem (#11)
• refs on accelerated ADMM (#8)
• Quantile loss (#5)
• Quantile reg (#4)
• MAD/LAD regression (#3)
• Investigate old "robustleastsquares" package (#1)

Merged pull requests:

• Penalizeintercept (#31) (tlienart)
• Benchmarking etc (#29) (tlienart)
• Malloc (#25) (tlienart)
• FISTA with robust (#22) (tlienart)
• cleaning up tests + comp to quantreg (#13) (tlienart)
• Quantile regression (#9) (tlienart)