Source: Deep Learning on Medium
What Precision, Recall, F1 Score and Accuracy Can Tell You ?
There is no doubt that to evaluate your classification algorithm you have to take accuracy into consideration. but, is accuracy everything ?
Firstly, What Is Accuracy ?
Accuracy is the number of correct predictions divided by the total number of predictions.
Accuracy = correct_preds / all_preds .
Imagine testing a model on a dataset consists of 90% dog images and only 10% cat images, what if all of the predictions it gives you are dogs ? you would easily get a 90% accuracy, Even if the accuracy is high, this still would be a poor model.
Accuracy tells you how good your model is performing generally. However, it does not give detailed information.
True Positive vs True Negative
A true positive is an outcome where the model correctly predicts the positive class.
a true negative is an outcome where the model correctly predicts the negative class.
False Positive vs False Negative
A false positive is an outcome where the model incorrectly predicts the positive class.
A false negative is an outcome where the model incorrectly predicts the negative class.
Precision is how often the model is accurate when it predicts positive.
Precision = true positive / ( true positive + false positive)
low precision tells you that there’s a high false positives rate.
Recall is what proportion of actual positives was classified correctly?
Recall = true positive / ( true positive + false negative)
low recall tells you that there’s a high false negatives rate.
F1 is an overall measure of a model’s accuracy that combines precision and recall.
F1 = 2 * (precision * recall)/(precision + recall)
High F1 score means that you have low false positives and low false negatives.
1 – Accuracy is suitable with balanced dataset when there are an equal number of observations in each class which isn’t common in real-life problems.
2 – Precision is important when the cost of false positives is high.
3 – Recall is important when the cost of false negatives is high.
4 – F1 score considers both the precision and recall.