Precision and Recall

Synonyms

False negative rate; Positive predictive value; Sensitivity

Definition

Recall measures the ability of a search engine or retrieval system to locate relevant material in its index. Precision measures its ability to not rank nonrelevant material. With everything above rank cut-off n considered “retrieved” and everything below considered “not retrieved,” precision and recall can be stated mathematically as:

$$ \begin{array}{l}\mathrm{precision}=\frac{\left|\mathrm{retrieved}\kern0.5em \&\kern0.5em \mathrm{relevant}\kern0.5em \mathrm{at}\kern0.5em \mathrm{rank}\kern0.5em n\right|}{\left|\mathrm{retrieved}\kern0.5em \mathrm{at}\kern0.5em \mathrm{rank}\kern0.5em n\right|}\\[8pt] {}\mathrm{recall}=\frac{\left|\mathrm{retrieved}\kern0.5em \&\kern0.5em \mathrm{relevant}\kern0.5em \mathrm{at}\kern0.5em \mathrm{rank}\kern0.5em n\right|}{\left|\mathrm{relevant}\right|}\end{array} $$

Key Points

Precision and recall are the traditional metrics for retrieval system performance evaluation,...