classWeightLearn 1.0.0

Optuna Class Weight Cost-Sensitive Learning

Cost Sensitive Learning

This is an Extended research from CTGAN-ENN, on algorithm-level approach by fine tuning class-weight using optuna, the supported model are SVM, Random Forest and LightBGM (other models come soon), the classifier model will run with the best class_weight parameter

Instalation

pip install classWeightLearn

Example Usage

from classWeightLearn import find_best_class_weight

best_weight_svc = find_best_class_weight(X, y, model_name='svc', n_trials=1)
classifier = SVC(class_weight=best_weight_svc)

Variables

• X : all features of your dataset
• y : target label of your dataset
• model_name : supported models are (svc, lightgbm and random_forest)
• n_trials : the number of trial that optuna will run (default:1)