Accuracy: Proportion of correctly classified cases among the total number of cases. - Overall correctness of predictions.
- Formula: (TP+TN)/Total number of cases.
Precision: Measures the accuracy of positive predictions. Proportion of true positives among all cases predicted as positive. AKA Positive Predictive Value (PPV). - Accuracy of positive predictions.
- Formula: TP/(TP+FP).
Recall: Defines how many positive cases that the model predicted are actually predicted right. Proportion of true positives among all actual positive cases. AKA True Positive Rate (TPR). - Ability to find all positive instances.
- Formula: TP/(TP+FN).
F1 Score: Harmonic mean of precision and recall, balancing between them.
- Formula: 2TP/(2TP+FP+FN).
Negative Predictive Value(NPV): Measures the proportion of actual negatives that are correctly identified by the model. Proportion of true negatives among all cases predicted as negative. - Accuracy of negative predictions.
- Formula: TN/(FN+TN).
Specificity: measures the proportion of actual negatives that are correctly identified by the model. Proportion of true negatives among all actual negative cases. AKA True Negative Rate (TNR). - Ability to identify true negatives.
- Formula: TN/(TN+FP).
