Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
xgboost/NEWS.md
Philip Cho 8d35c09c55 Tag version 0.7 (#2975) 
* Tag version 0.7

* Document all changes made in year 2016
2017-12-30 20:16:41 +08:00

8.4 KiB
Raw Blame History

XGBoost Change Log

This file records the changes in xgboost library in reverse chronological order.

v0.7 (2017.12.26)

  • Updated Sklearn API
    • Add compatibility layer for scikit-learn v0.18
    • Updated to allow use of all XGBoost parameters via **kwargs.
    • Updated nthread to n_jobs and seed to random_state (as per Sklearn convention).
  • Refactored gbm to allow more friendly cache strategy
    • Specialized some prediction routine
  • Robust DMatrix construction from a sparse matrix
  • Elide copies when building DMatrix from 2D NumPy matrices
  • Automatically remove nan from input data when it is sparse.
    • This can solve some of user reported problem of istart != hist.size
  • Minor fixes
    • Thread local variable is upgraded so it is automatically freed at thread exit.
    • Fix saving and loading count::poisson models
    • Fix CalcDCG to use base-2 logarithm
    • Messages are now written to stderr instead of stdout
    • Keep built-in evaluations while using customized evaluation functions
    • Use bst_float consistently to minimize type conversion
  • Migrate to C++11
    • The current master version now requires C++11 enabled compiled(g++4.8 or higher)
  • Predictor interface was factored out (in a manner similar to the updater interface).
  • Makefile support for Solaris
  • Test code coverage using Codecov
  • Add CPP tests
  • New functionality
    • Ability to adjust tree model's statistics to a new dataset without changing tree structures.
    • Extracting feature contributions to individual predictions.
    • Faster, histogram-based tree algorithm (tree_method='hist') .
    • GPU/CUDA accelerated tree algorithms (tree_method='gpu_hist' or 'gpu_exact'), including the GPU-based predictor.
    • Monotonic constraints: when other features are fixed, force the prediction to be monotonic increasing with respect to a certain specified feature.
    • Faster gradient caculation using AVX SIMD
    • Ability to export models in JSON format
    • Support for Tweedie regression
    • Ability to update an existing model in-place: this is useful for many applications, such as determining feature importance
  • Python package:
    • New parameters:
      • learning_rates in cv()
      • shuffle in mknfold()
    • Support binary wheel builds
    • Fix MultiIndex detection to support Pandas 0.21.0 and higher
    • Fix early stopping for evaluation sets whose names contain -
    • Support feature maps when plotting trees
  • R package:
    • New parameters:
      • silent in xgb.DMatrix()
      • use_int_id in xgb.model.dt.tree()
      • predcontrib in predict()
      • monotone_constraints in xgb.train()
    • Default value of the save_period parameter in xgboost() changed to NULL (consistent with xgb.train()).
    • It's possible to custom-build the R package with GPU acceleration support.
    • Integration with AppVeyor CI
    • Improved safety for garbage collection
    • Updated CRAN submission
    • Store numeric attributes with higher precision
    • Easier installation for devel version
  • JVM packages
    • Fix data persistence: loss evaluation on test data had wrongly used caches for training data.
    • Make IEvaluation serializable
    • Enable training of multiple models by distinguishing stage IDs
    • Better Spark integration: support RDD / dataframe / dataset, integrate with Spark ML package
    • Support training with missing data
    • Refactor JVM package to separate regression and classification models to be consistent with other machine learning libraries
    • Support XGBoost4j compilation on Windows
    • Parameter tuning tool
    • Publish source code for XGBoost4j to maven local repo
    • Scala implementation of the Rabit tracker (drop-in replacement for the Java implementation)
  • Documentation
    • Better math notation for gradient boosting
    • Updated installation instructions for Mac OS X
    • Template for GitHub issues
    • Add CITATION file for citing XGBoost in scientific writing
    • Fix dropdown menu in xgboost.readthedocs.io
    • Document updater_seq parameter
    • Style fixes for Python documentation
  • Backward compatiblity
    • XGBoost-spark no longer contains APIs for DMatrix (#1519); use the public booster interface instead.

v0.6 (2016.07.29)

  • Version 0.5 is skipped due to major improvements in the core
  • Major refactor of core library.
    • Goal: more flexible and modular code as a portable library.
    • Switch to use of c++11 standard code.
    • Random number generator defaults to std::mt19937.
    • Share the data loading pipeline and logging module from dmlc-core.
    • Enable registry pattern to allow optionally plugin of objective, metric, tree constructor, data loader.
      • Future plugin modules can be put into xgboost/plugin and register back to the library.
    • Remove most of the raw pointers to smart ptrs, for RAII safety.
  • Add official option to approximate algorithm tree_method to parameter.
    • Change default behavior to switch to prefer faster algorithm.
    • User will get a message when approximate algorithm is chosen.
  • Change library name to libxgboost.so
  • Backward compatiblity
    • The binary buffer file is not backward compatible with previous version.
    • The model file is backward compatible on 64 bit platforms.
  • The model file is compatible between 64/32 bit platforms(not yet tested).
  • External memory version and other advanced features will be exposed to R library as well on linux.
    • Previously some of the features are blocked due to C++11 and threading limits.
    • The windows version is still blocked due to Rtools do not support std::thread.
  • rabit and dmlc-core are maintained through git submodule
    • Anyone can open PR to update these dependencies now.
  • Improvements
    • Rabit and xgboost libs are not thread-safe and use thread local PRNGs
    • This could fix some of the previous problem which runs xgboost on multiple threads.
  • JVM Package
    • Enable xgboost4j for java and scala
    • XGBoost distributed now runs on Flink and Spark.
  • Support model attributes listing for meta data.
  • Support callback API
  • Support new booster DART(dropout in tree boosting)
  • Add CMake build system

v0.47 (2016.01.14)

  • Changes in R library
    • fixed possible problem of poisson regression.
    • switched from 0 to NA for missing values.
    • exposed access to additional model parameters.
  • Changes in Python library
    • throws exception instead of crash terminal when a parameter error happens.
    • has importance plot and tree plot functions.
    • accepts different learning rates for each boosting round.
    • allows model training continuation from previously saved model.
    • allows early stopping in CV.
    • allows feval to return a list of tuples.
    • allows eval_metric to handle additional format.
    • improved compatibility in sklearn module.
    • additional parameters added for sklearn wrapper.
    • added pip installation functionality.
    • supports more Pandas DataFrame dtypes.
    • added best_ntree_limit attribute, in addition to best_score and best_iteration.
  • Java api is ready for use
  • Added more test cases and continuous integration to make each build more robust.

v0.4 (2015.05.11)

  • Distributed version of xgboost that runs on YARN, scales to billions of examples
  • Direct save/load data and model from/to S3 and HDFS
  • Feature importance visualization in R module, by Michael Benesty
  • Predict leaf index
  • Poisson regression for counts data
  • Early stopping option in training
  • Native save load support in R and python
    • xgboost models now can be saved using save/load in R
    • xgboost python model is now pickable
  • sklearn wrapper is supported in python module
  • Experimental External memory version

v0.3 (2014.09.07)

  • Faster tree construction module
    • Allows subsample columns during tree construction via bst:col_samplebytree=ratio
  • Support for boosting from initial predictions
  • Experimental version of LambdaRank
  • Linear booster is now parallelized, using parallel coordinated descent.
  • Add Code Guide for customizing objective function and evaluation
  • Add R module

v0.2x (2014.05.20)

  • Python module
  • Weighted samples instances
  • Initial version of pairwise rank

v0.1 (2014.03.26)

  • Initial release