GENERATE_utils
- BO_workflow(self, Xy_data, csv_df, ML_model)
Load hyperparameter space and perform a Bayesian optimization
- PFI_workflow(self, csv_df, ML_model, Xy_data)
Filter off parameters with low PFI (not relevant in the model)
- calc_desc_keep(self, Xy_data, PFI_discard_cols)
Calculate number of descriptors to keep in the PFI model
- detect_best(folder)
Check which combination led to the best results
- heatmap_workflow(self, folder_hm)
Create matrix of ML models, training sizes and errors/precision
- save_pfi_csv(self, csv_df, name_csv_hyperopt, PFI_dict, Xy_data_PFI, ML_model)
Saves CSV files with PFI models and information
- set_sets(csv_df, Xy_data)
Set a new column for the sets, including test set (if any)