v0.10.13

DecisionTree v0.10.13

Diff since v0.10.12

• (bug fixes) Ensure nfoldsCV passes the RNG on to AdaBoostClassifier models. Fix non-threadsafe use of RNGs in random forests (#174) @rikhuijzer
• (testing) Replace all RNGs in tests with StableRNGs to ensure consistency across Julia versions and introduce tests over multiple seeds (#174) @rikhuijzer
• (enhancement) Improve display returned by print_tree (making it consistent with the AbstractTrees.jl fallback) and fix docstring (#171, #172, #173) @rikhuijzer

Merged pull requests:

• Fix typo in print_tree description (#171) (@rikhuijzer)
• Replace L by < and R by ≥ (#172) (@rikhuijzer)
• Round digits in print_tree (#173) (@rikhuijzer)
• Test multiple seeds (#174) (@rikhuijzer)
• For a 0.10.13 release (#178) (@ablaom)

v0.10.12

DecisionTree v0.10.12

Diff since v0.10.11

• (enhancement) Add method wrap which allows one to wrap a decision tree (Leaf or Node object) in a new tree structure implementing the AbstractTrees.jl interface. Unlike raw DecisionTree.jl decision trees, the nodes of the wrapped objects can include the names of splitting-features at nodes, and any other node metadata. See the wrap docstring for details. (#158) @roland-KA

Closed issues:

• Avoid "regressor" terminology (#49)
• Support sample weights? (#141)
• TagBot trigger issue (#145)
• Splitting criterion in regression tree (#148)
• Transfer of ownership? (#152)
• Update General.jl to reflect new ownership (#153)
• Issue to trigger new releases (#162)

Merged pull requests:

• Add docstring to print_tree (#151) (@KronosTheLate)
• Update readme to reflect new ownership (#154) (@ablaom)
• Add MLJ usage instructions (#156) (@ablaom)
• Add implementation of AbstractTrees-interface (#158) (@roland-KA)
• For a 0.10.12 release (#161) (@ablaom)
• Relax some test tolerances (#163) (@ablaom)
• Add compat AbstractTrees = "0.3" (#165) (@ablaom)

v0.10.11

What's Changed

• Add GitHub actions CI by @IanButterworth in https://github.com/bensadeghi/DecisionTree.jl/pull/143
• Add option to print feature names in print_tree by @IanButterworth in https://github.com/bensadeghi/DecisionTree.jl/pull/142
• Move check_input to utils to avoid duplicate code by @barucden in https://github.com/bensadeghi/DecisionTree.jl/pull/137
• Fix feature_names kwarg structure for julia 1.0 by @IanButterworth in https://github.com/bensadeghi/DecisionTree.jl/pull/144

New Contributors

• @IanButterworth made their first contribution in https://github.com/bensadeghi/DecisionTree.jl/pull/143

Full Changelog: bensadeghi/DecisionTree.jl@v0.10.10...v0.10.11

v0.10.10

DecisionTree v0.10.10

Diff since v0.10.9

Merged pull requests:

• calibrate using training data for cross validation (#133) (@salbert83)

v0.10.9

DecisionTree v0.10.9

Diff since v0.10.8

Merged pull requests:

• Fitting using AbstractMatrix instead of Matrix (#132) (@barucden)

v0.10.8

DecisionTree v0.10.8

Diff since v0.10.7

Closed issues:

• README update (#130)

Merged pull requests:

• Allow fitting using AbstractMatrix instead of Matrix (#131) (@barucden)

v0.10.7

DecisionTree v0.10.7

Diff since v0.10.6

Merged pull requests:

• use non-abstract types for structs (#129) (@bensadeghi)

v0.10.6

DecisionTree v0.10.6

Diff since v0.10.5

Closed issues:

• Features of type AbstractVector in apply_tree? (#124)
• Forest building on parallel threads is not a good idea (#125)
• Request: option to suppress printing in nfoldCV_tree (#127)

Merged pull requests:

• seeding for reproducible multi-threaded RFs (#126) (@bensadeghi)
• use AbstractVector and AbstractMatrix (#128) (@bensadeghi)

v0.10.5

DecisionTree v0.10.5

Diff since v0.10.4

v0.10.4

DecisionTree v0.10.4

Diff since v0.10.3

Closed issues:

• build_adaboost_stumps produces unexpected convert errors (#121)

Merged pull requests:

• Add convert: Leaf to Node (#122) (@bensadeghi)