CHANGELOG.md

CHANGELOG

v.1.1.1 - 5 July 2022

• DownscaledLST() should now be used with scikit-learn v.1.1.1.
• requirements.txt was updated and environment.yml was added.
• Removed unused imports. Fixed some typos.

v.1.1.0 - 5 March 2020

• The ensemble model is now derived using a stacked regressor instead of a voting regressor. The estimator used for the stacking is an elasticNET regressor with built-in cross-validation.
• The ensemble model is now built using a random forest, a ridge regressor and a SVM. The adaboost and elasticNET regressors used in v.1.0.0 have been dropped.
• The class now uses random search with cross validation to fine-tune the hyperparameters of the 3 regressors and thus does not require the user to estimate them beforehand. The default number of hyperparameter searches is 60 but it can be changed using the setter set_num_searches.
• The class now transforms the input predictors to follow a normal distribution using sklearn's QuantileTransformer. The number of bins is uses are equal to 50% of the length of y.
• The setters SetAdaBoostParams, SetRandomForestParams, SetElasticNetParams and SetRidgeRegrParams have been deleted.
• The methods set_num_searches for setting how many hyperparameter sets that will be tested and set_num_jobs for setting the maximum number of parallel jobs (default is 1) have been addded.