/machine-learning-imbalanced-data

Code repository for the online course Machine Learning with Imbalanced Data

Primary LanguageJupyter NotebookOtherNOASSERTION

Machine Learning with Imbalanced Data - Code Repository

Python 3.6 Python 3.7 Python 3.8 License

Published November, 2020

Links

Table of Contents

  1. Metrics

    1. Limitations of the Accuracy
    2. Precision, Recall, F-Measure
    3. Confusion Matrix
    4. False Positive Rate and False Negative Rate
    5. Geometric Mean
    6. Dominance
    7. Index of imbalanced accuracy
    8. ROC-AUC
    9. Precision-Recall Curves
    10. Probability Distribution and Calibration
    11. Which metric to optimise
  2. Udersampling Methods

    1. Random Undersampling
    2. Condensed Nearest Neighbour
    3. Tomek Links
    4. One Sided Selection
    5. Edited Nearest Neighbours
    6. Repeated Edited Nearest Neighbours
    7. All KNN
    8. Neighbourhood Cleaning Rule
    9. NearMiss
    10. Instance Hardness Threshold
  3. Oversampling methods

    1. Random Oversampling
    2. ADASYN
    3. SMOTE
    4. BorderlineSMOTE
    5. KMeansSMOTE
    6. SMOTENC
    7. SVMSMOTE
  4. Over and Undersampling Methods

    1. SMOTENN
    2. SMOTETomek
  5. Ensemble Methods

    1. Coming Soon
  6. Cost Sensitive Learning

    1. Types of cost
    2. Obtaining the Cost
    3. Missclassification Cost
    4. Bayes Risk
    5. MetaCost
  7. Probability Calibration

    1. Probability Calibration Curves
    2. Brier Score
    3. Effect of under and over sampling on Probability Calibration
    4. Cost Sensitive Learning and Probability Calibration
    5. Calibrating a Classifier