Mini toolsDefinition
A situation where classes in a dataset are not represented equally potentially biasing model performance.
Detailed Explanation
Class imbalance occurs when the distribution of examples across known classes is not uniform. This can lead to biased models that perform well on majority classes but poorly on minority classes. The imbalance ratio can range from slight to severe requiring specific techniques like resampling class weights or specialized loss functions to address.
Use Cases
Medical diagnosis systems fraud detection rare event prediction
