Model Performance Evaluation Metrics for Classification Problems

references

• Luque et al., 2019Luque A., Carrasco A., Martín A., de las Heras A., The impact of class imbalance in classification performance metrics based on the binary confusion matrix, Pattern Recognition 91 (2019) 216–231, 10.1016/j.patcog.2019.02.023.
  • https://doi.org/10.1016/j.patcog.2019.02.023
  • The list of metrics and formulas in the paper

highlights of paper

"GM (Geometric Mean) and BM (Bookmaker Informedness) are the best performance metrics if their focus on successes (dismissing the errors) presents no limitation for the specific application where they are used. However, if classification errors must also be considered, then MCC (Matthews Correlation Coefficient) arises as the best choice."

metrics and formulas

• Confusion matrix for binary classification.

• Definition of classification performance metrics.

• Behaviour of performance metrics with imbalanced datasets.

download formulas: word | pdf