Definition
Cost-sensitive machine learning extends traditional supervised learning by assigning different penalties to different types of errors. In real-world scenarios, false positives and false negatives often have unequal consequences. This approach modifies loss functions or sampling strategies to minimize the total expected cost of predictions, making it essential for domains like fraud detection or medical diagnosis where error costs vary significantly.
Summary
A machine learning paradigm that incorporates misclassification costs into the training process to optimize for economic impact rather than just accuracy.
Key Concepts
- Loss Function Modification
- Class Imbalance
- Misclassification Cost
- Optimization Objective
Use Cases
- Fraud detection in banking
- Medical disease screening
- Spam filtering with high false-positive costs