GitLab

* add feature

* add label idx move func to .cpp file

* fix

* move func to predictor

* restore files

* fix by commits

* move the non-used features to the end of the feature vector

* fix by commits

* fix bug by commits

* Explain individual predictions using SHAP value feature attributions

* Address code review

* fix multi-threading.

* fix name style.

* support in CLI version.

* remove warnings.

* Not default parameters.

* fix if...else... .

* fix bug.

* fix warning.

* refine c_api.

* fix R-package.

* fix R's warning.

* fix tests.

* fix pep8 .

* Add early stopping for prediction

* Fix GBDT if-else prediction with early stopping

* Small C++ embelishments to early stopping API and functions

* Fix early stopping efficiency issue by creating a singleton for no early stopping

* Python improvements to early stopping API

* Add assertion check for binary and multiclass prediction score length

* Update vcxproj and vcxproj.filters with new early stopping files

* Remove inline from PredictRaw(), the linker was not able to find it otherwise

* refine prediction logic.

* fix test.

* fix out_len in training score of Dart.

* improve predict speed for high dimension data.

* try use unordered_map for sparse prediction.

* avoid using unordered_map.

* clean code.

* fix test.

* move predict buffer to Predictor.

* refine prediction logic.

* fix test.

* fix out_len in training score of Dart.

* improve predict speed for high dimension data.

* Added possibility to compile without openmp (useful for clang on Mac)

* Added CMake option to create dylib on MacOS

Categorical feature support (#108)

* RAII for utils, application and c_api(partical)

* raii for class in include folder

* raii for application and boosting

* raii for dataset and dataset loader

* raii for dense bin and parser

* RAII refactor for almost all classes

* RAII for c_api

* clean code

* refine repeated code

* Decouple the "sigmoid" between objective and boosting.

* change std::vector<bool> back to std::vector<char> due to concurrence problem

* slight reduce some memory cost

Packages that parse LightGBM’s logs will require minor changes to
parsing logic to work correctly.

support init_score for multiclass classification (#62)

Support multiclass classification (#53)

* a python regression example

* add early-stopping

* fix bugs

* remove not needed files

* update easy-stopping; fix warnings; add alias

* add comments; remove useless line

* change code order

* change Print to PrintAndGetLoss: return loss

* change comment; add bracket

* change return type of PredictLeafIndex in boosting.h

* move string cast into lambda function

* move if-then outside predict_fun