v0.3.4

RandomFeatures v0.3.4

Diff since v0.3.3

Merged pull requests:

• Update examples with CES.jl loss functions and other cleanup (#57) (@odunbar)
• update EKP compat to v2 (#59) (@odunbar)
• New loss functions, and bugfixes for CI pipeline (#61) (@odunbar)
• Update Project.toml (#62) (@odunbar)

Closed issues:

• TODO list as of Aug 30 (#1)
• Update examples with the robust training framework developed in CalibrateEmulateSample.jl (#52)
• Plots/FFMPEG CI issue, required adding Plots into Project.toml - find another workaround (#60)

v0.3.3

RandomFeatures v0.3.3

Diff since v0.3.2

Merged pull requests:

• Update [compat] for new requirements (#53) (@odunbar)
• updated workflows for removal of bors (#54) (@odunbar)
• Changing how we deal with Regularization (#55) (@odunbar)
• project.toml v0.3.3 (#56) (@odunbar)

Closed issues:

• Default / Advise setting sqrt(train pts) as regularization value (#51)

v0.3.2

RandomFeatures v0.3.2

Diff since v0.3.1

Closed issues:

• update license (#45)
• Set threads = 10^9 not threads = false (#47)
• Use input-output dim here, over inferring them from distribution (#49)

Merged pull requests:

• CompatHelper: bump compat for StatsBase to 0.34, (keep existing compat) (#46) (@github-actions[bot])
• replace false with 10^9 (#48) (@odunbar)
• Change inferred sizing to concrete sizing (#50) (@odunbar)

v0.3.1

RandomFeatures v0.3.1

Diff since v0.3.0

Closed issues:

• multithreading control? (#31)

Merged pull requests:

• multithreading flags (#43) (@odunbar)
• Project to 0.3.1 (#44) (@odunbar)

v0.3.0

RandomFeatures v0.3.0

Diff since v0.2.5

Closed issues:

• Use SVD of Phi to build matrices? (#38)

Merged pull requests:

• Continued accelerations from profiling (#40) (@odunbar)

v0.2.5

RandomFeatures v0.2.5

Diff since v0.2.4

Merged pull requests:

• Accelerations learnt from profiling (#39) (@odunbar)

v0.2.4

RandomFeatures v0.2.4

Diff since v0.2.3

Merged pull requests:

• compat for EKP (#37) (@odunbar)

v0.2.3

RandomFeatures v0.2.3

Diff since v0.2.2

Closed issues:

• If cholesky fails, revert to a default method (#33)

Merged pull requests:

• More stable solves for non-diagonal regularizers (#36) (@odunbar)

v0.2.2

RandomFeatures v0.2.2

Diff since v0.2.1

Closed issues:

• We should not be doing block pseudo-inverses (#34)

Merged pull requests:

• bugfix pos-def matrix regularization (#35) (@odunbar)

v0.2.1

RandomFeatures v0.2.1

Diff since v0.2.0

Closed issues:

• remove copy(rng) (#27)

Merged pull requests:

• Orad/noise regularization (#25) (@odunbar)
• Docs vector features (#26) (@odunbar)
• small bugfixes (#28) (@odunbar)
• add code and example for pos def Tikhonov parameter (#29) (@odunbar)
• CompatHelper: bump compat for EnsembleKalmanProcesses to 0.14, (keep existing compat) (#30) (@github-actions[bot])