Skip to content

Latest commit

 

History

History
82 lines (59 loc) · 3.41 KB

ChangeLog.md

File metadata and controls

82 lines (59 loc) · 3.41 KB
Raw

Version 1.2.0

  • Rename ELM class to MultiLayerELM class. This new class can be used to define deep ELM network by layer_sizes parameter.
  • Add AutomatedMhaElmTuner class that can be used to perform hyperparameter tuning for MhaElm models using either GridSearchCV or RandomizedSearchCV. Provides an interface for fitting and predicting using the best found model.
  • Add AutomatedMhaElmComparator class that automatic compare different MhaElm models based on provided optimizer configurations. It provides methods for cross-validation and train-test split evaluation.
  • Update docs, examples, and tests.

Version 1.1.1

  • Update seed value in all 4 classes to ensure reproducibility of your results
  • Add mode, n_workers, and termination parameter in model.fit() of MhaElmRegressor and MhaElmClassifier classes
    • These parameters are derived from Mealpy library
    • With mode parameter, you can speed your training model
    • With n_workers, you can set the number of threads or CPUs to speed up the training process
    • With termination, you can set early stopping strategy for your model.
  • Update docs, examples, and tests.

Version 1.1.0

  • Update core modules to fit upgraded version of Mealpy>=3.0.1, PerMetrics>=2.0.0, Scikit-Learn>=1.2.1
  • IntelELM no longer support Python 3.7. Only support Python >= 3.8
  • Update docs and add examples

Version 1.0.3

  • Fix bug lb and ub in BaseMhaElm class
  • Update docs and add example

Version 1.0.2

  • Fix bug in DataTransformer class
  • Fix bug in LabelEncoder class
  • Add more activation functions
  • Update documents, examples

Version 1.0.1

  • Add "evaluate" function to all Estimators (ElmRegressor, ElmClassifier, MhaElmRegressor, MhaElmClassifier)
  • Add new module "scaler"
  • Our scaler can be utilized with multiple methods.
  • Add "save_loss_train" and "save_metrics" functions to all Estimators
  • Add "save_model" and "load_model" functions to all Estimators
  • Add "save_y_predicted" function to all Estimators
  • Update all examples and documents

Version 1.0.0

  • Add supported information for each classes.
  • Restructure intelelm module to based_elm module and model subpackage that includes mha_elm and standard_elm modules.
  • Add traditional/standard ELM models (ElmRegressor and ElmClassifier classes) to standard_elm module.
  • Add examples and tests for traditional models
  • Add score and scores functions to all classes.
  • Fix bug calculate metrics and objective in ELM-based models.
  • Add examples with real-world datasets and examples with GridsearchCV to tune hyper-parameters of ELM-based models.
  • Add documents

Version 0.1.0 (First version)

  • Add infors (CODE_OF_CONDUCT.md, MANIFEST.in, LICENSE, README.md, requirements.txt, CITATION.cff)
  • Add supported classification and regression datasets
  • Add util modules (data_loader, validator, evaluator, encoder, activation)
  • Add MhaElmRegressor and MhaElmClassifier classes
  • Add publish workflows
  • Add examples and tests folders