changelog.rst

Change Log

Version 0.6.dev

• New Cost-Sensitive Churn dataset
• Introduction to Example-Dependent Cost-Sensitive Classification
• Slides costcla for fraud detection

Version 0.5 (2016-01-29)

• Support of Python 3.5

Version 0.4 (2015-05-26)

• Documentation
• Cost-sensitive Decision Trees algorithm
• Cost-sensitive Ensemble methods
• Tutorial CS Credit Scoring
• Slides CS Credit Scoring

Version 0.03 (2014-09-19)

• Initial public release
• Include the cost-sensitive algorithms:
  • Bayes minimum risk classifier
  • Cost-sensitive Logistic Regression
  • Cost-sensitive Thresholding
• Sampling techniques:
  • Cost-proportionate sampling methods
  • SMOTE
  • under-sampling
• ROCCH probability calibration
• Evaluation metrics
• Cost-sensitive datasets:
  • Direct Marketing
  • Kaggle11 Credit Scoring
  • PAKDD09 Credit Scoring