Confusion Metrics, Accuracy, Precision, Recall, F1-Score
Table Of Contents:
- Confusion Metrics
- Accuracy
- Precision
- Recall
- F1-Score
- Specificity (True Negative Rate)
- Fall-Out : False Positive Rate (FPR)
- Miss Rate: False Negative Rate (FNR)
- Balanced Accuracy
- Example Of Confusion Metrics
- Confusion Metrics For Multiple Class
(1) Confusion Metrics
- A confusion matrix is a tabular representation of the performance of a classification model.
- It helps visualize and understand how well a model’s predictions align with the actual outcomes, especially in binary and multi-class classification.
(2) Structure Of Confusion Metrics
Definition:
(3) Accuracy
- For imbalanced dataset Accuracy will give the wrong impression about the model performance. We can get 95% Accuracy but we have misclassified other classes.
- Accuracy gives equal weightage to all the classes, because in the denominator we always divide the total value, whether it is a higher False Positive or higher False Negative the Accuracy does not takes into focus on them.
(4) Precision (Positive Predictive Value)
- Precision is used when we want to focus on one class of data only.
- In precision we focus on only positive classes.
(5) Recall (Sensitivity or True Positive Rate)
(6) Specificity (True Negative Rate)
(7) F1-Score
(8) Fall-Out : False Positive Rate (FPR)
(9) Miss Rate: False Negative Rate (FNR)
(10) Balanced Accuracy
(11) Example Of Confusion Metrics
(12) Confusion Metrics For Multiple Class
(13) Why Decreasing False Negative Increases False Positive?
