Macro-Recall Metric

A Macro-Recall Metric is a macro-averaged performance measure that is a recall metric computed as the unweighted arithmetic mean of per-class recall scores.

• AKA: Macro-Recall, Unweighted Mean Recall, Class-Averaged Sensitivity Measure.
• Context:
  • It can typically calculate Macro-Recall Score Values by computing individual macro-recall class-specific recalls and averaging them without weighting.
  • It can typically treat Macro-Recall Class Contributions equally regardless of macro-recall class frequency or macro-recall class support.
  • It can typically provide Macro-Recall Balanced Assessments of macro-recall detection capability across all macro-recall classes.
  • It can typically reveal Macro-Recall Detection Gaps in macro-recall minority classes that macro-recall aggregate metrics might hide.
  • It can typically serve as a component for computing Macro-F1 Measures when combined with macro-precision metrics.
  • ...
  • It can often produce Macro-Recall Different Values than micro-recall metrics especially in macro-recall imbalanced datasets.
  • It can often be preferred when Macro-Recall Per-Class Sensitivity matters more than macro-recall overall detection rate.
  • It can often be sensitive to Macro-Recall Poor Detection in any single macro-recall class.
  • It can often be reported with Macro-Recall Variance Measures to indicate macro-recall performance spread.
  • ...
  • It can range from being a Low Macro-Recall Metric to being a High Macro-Recall Metric, depending on its macro-recall detection quality.
  • It can range from being a Consistent Macro-Recall Metric to being an Inconsistent Macro-Recall Metric, depending on its macro-recall cross-class uniformity.
  • It can range from being a Conservative Macro-Recall Metric to being a Liberal Macro-Recall Metric, depending on its macro-recall threshold setting.
  • It can range from being a Stable Macro-Recall Metric to being a Volatile Macro-Recall Metric, depending on its macro-recall temporal variance.
  • ...
  • It can be calculated using Macro-Recall Formula: (1/n) × Σ(Recall_i) for n classes.
  • It can be visualized through Macro-Recall Heatmaps showing macro-recall class-wise performance.
  • It can be monitored through Macro-Recall Tracking Dashboards during macro-recall model development.
  • It can be improved using Macro-Recall Enhancement Strategys targeting macro-recall weak classes.
  • ...
• Example(s):
  • Medical Diagnosis Macro-Recall Metrics, such as:
    • Multi-Disease Detection Macro-Recall Metric for macro-recall screening tasks.
    • Symptom Recognition Macro-Recall Metric for macro-recall diagnostic tasks.
    • Treatment Response Macro-Recall Metric for macro-recall prediction tasks.
  • Security System Macro-Recall Metrics, such as:
    • Multi-Threat Detection Macro-Recall Metric for macro-recall security monitoring tasks.
    • Attack Type Recognition Macro-Recall Metric for macro-recall intrusion detection tasks.
  • Quality Control Macro-Recall Metrics, such as:
    • Multi-Defect Detection Macro-Recall Metric for macro-recall inspection tasks.
    • Error Type Detection Macro-Recall Metric for macro-recall quality assurance tasks.
  • ...
• Counter-Example(s):
  • Micro-Recall Metric, which aggregates true positives and false negatives globally rather than averaging per-class recall.
  • Weighted Recall Metric, which applies class weights based on sample frequency rather than equal weighting.
  • Macro-Precision Metric, which measures per-class precision rather than per-class recall.
  • Macro-F1 Measure, which combines macro-recall with macro-precision rather than measuring recall alone.
• See: Recall Metric, Macro-Averaged Performance Measure, Micro-Recall Metric, Macro-Precision Metric, Macro-F1 Measure, Sensitivity Measure, True Positive Rate, Multi-Class Classification Task.

---