📏 Micro Recall

Micro Recall#

Recall computed from globally pooled true positives and false negatives.

Important

✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.

What it is#

Recall answers: of all the items that are actually positive, how many did the model catch?

\[\text{Recall} = \frac{TP}{TP + FN},\]

so it is sensitive to false negatives (missed positives). Micro recall extends this to \(K\) classes by pooling counts across classes rather than averaging per-class recall (which is macro recall).

How it’s computed#

Sum true positives and false negatives over every class, then divide:

\[\text{Recall}_{\text{micro}} = \frac{\sum_{i=1}^{K} TP_i}{\sum_{i=1}^{K} (TP_i + FN_i)}.\]

This treats the whole multi-class problem as one pooled “caught vs missed” question, so majority classes dominate the result.

The micro identity#

In single-label problems, micro recall equals micro precision equals micro F1 (and equals accuracy): once everything is pooled, the denominators coincide. The three diverge only in the multi-label setting.

Worked example#

Three classes with TP = (40, 30, 10), FN = (10, 20, 30):

Recall(A)=0.80, Recall(B)=0.60, Recall(C)=0.25 → Macro recall = 0.55.
Micro recall \(= 80/140 \approx 0.571\).

Macro weights every class equally (exposing class C’s weak 0.25); micro is a global, sample-weighted figure.

When recall matters most#

Favour recall when a missed positive is costly — disease screening, fraud detection, safety alerts — where you would rather tolerate false alarms than let a true case slip through.

In code#

from sklearn.metrics import recall_score

micro = recall_score(y_true, y_pred, average="micro")
macro = recall_score(y_true, y_pred, average="macro")

Theme: Classification & Averaging Metrics · All terminology

Hint

Mind map — connected ideas

Micro Precision · Micro F1 · Macro Recall · Micro AUROC

Hint

More in Classification & Averaging Metrics

Accuracy · AUC (Area Under the Curve) · Average Precision (AP) · Binary Classification · Classification Probability · Discriminatory Power · F1-score · Gini Coefficient · Harmonic Mean · Log Loss (also called Logarithmic Loss or Cross-Entropy Loss) · Macro AUC · Macro AUROC (Macro-Averaged AUROC) · Macro Averaging · Macro F1