Classification Report Explained

The classification report is part of the scikit-learn module in python. It is report containing the key metrics in a classification problem and showing the quality of the predictions.

There are 4 different outcomes to a prediction:

True Negative — A negative result correctly classified as negative.

True Positive — A positive result correctly classified as positive.

False Negative — A negative result incorrectly classified as positive.

False Positive — A positive result incorrectly classified as negative.

Example classification report output:

Using the classification report results in a table including precision, recall, f1-score, and support along the top.

Precision — The percentage of correctly classified results among that class.

Recall — The number of true positive cases found over the total number of positive cases found (true positives + false negatives).

F1-score — The harmonic mean of precision and recall. The F1-score will always be between 1.00 and 0.00 with 1.00 being the best score.