Learn to over-sample and under-sample your data, apply SMOTE, ensemble methods, and cost-sensitive learning.
InstructorSoledad Galli
Editorial course preview
These 3 complementary views highlight concrete, legible examples from the course presentation.
A course slide titled You'll learn displays a central diagram connecting imbalanced data with six surrounding topics such as re-sampling and metrics.
Presentation slide titled Requirements listing machine learning and Python programming skills with Numpy, Pandas, and Scikit-learn
Presentation slide titled Code implementation featuring the imbalanced-learn logo with a seesaw graphic and text identifying a Python open-source package.
The curriculum explores diverse methods including Random Under-Sampling, Tomek Links, NearMiss, and various SMOTE iterations like SMOTE-NC and ADASYN.
Instruction extends beyond accuracy to include Precision, Recall, F-measure, Balanced Accuracy, and ROC-AUC, alongside probability calibration.
Lessons include hands-on code examples and Jupyter notebooks that serve as useful resources for quick revision.
One signal from 2023 suggests a clunky delivery characterized by repeated steps and excessive discussion on certain topics; note that the 2024 update label does not guarantee these pacing issues were addressed.
A recent signal indicates distracting mouth noises and background noise during recordings.
The curriculum aligns well with the requirements of intermediate learners interested in imbalanced data.