Recall is a metric that measures the proportion of all possible true positives that the model was able

to identify for a given concept1. A true positive is a case where the model correctly predicts the

presence of a concept in the data. Recall is calculated as the ratio of true positives to the sum of true

positives and false negatives, where a false negative is a case where the model fails to predict the

presence of a concept in the data.🔍Recall can be interpreted as the sensitivity or completeness of the

model for a given concept2. For example, if there are 100 verbatims that should have been labelled

as ‘Request for information’, and the model detects 80 of them, then the recall for this concept is

80% (80 / (80 + 20)). A high recall means that the model is good at finding all the relevant cases for a

concept, while a low recall means that the model misses many of them.

Reference:🔍1:🔍Recall🔍2:🔍Precision and Recall