precision and recall

_{scroll ↓ to Resources}

Note

Precision

• High fraction of TP within what the model labels as positive class: $\frac{TP}{TP+FP}$
• *bank wants to give a loan ONLY to those who certainly will return it, even if this number is small and excludes some people who also will return the money *
• ==low precision = wasting resources (processing irrelevant content)==

Recall

• Powerful “coverage” of the positive class: $\frac{TP}{TP+FN}$
• same as true positive rate
• bank wants to find all clients who will return the loan, indifferent to those who won’t be able to return it
• ==low recall = missing important information==

---

• See also Precision@k and Recall@k at different cut-off points K
  • for object retrieval: Metrics
  • for retrieval evaluation
• ==In practice it is good to track evolution of metrics for various $k$: 1, 10, 50, 1000…==

Resources

---

Links to this File

table file.inlinks, file.outlinks from [[]] and !outgoing([[]])  AND -"Changelog"