Micro-Recall Metric

A Micro-Recall Metric is a micro-averaged performance measure that is a recall metric computed by aggregating true positives and false negatives globally across all classes.

• AKA: Micro-Recall, Global Recall Metric, Pooled Recall Measure, Aggregate Sensitivity Metric.
• Context:
  • It can typically calculate Micro-Recall Score Values by summing all micro-recall true positives and dividing by the sum of all micro-recall true positives and micro-recall false negatives across all classes.
  • It can typically weight Micro-Recall Class Contributions implicitly by their micro-recall actual frequency, giving more influence to micro-recall frequent classes.
  • It can typically provide Micro-Recall Global Sensitivity that measures overall micro-recall detection rate across all micro-recall positive instances.
  • It can typically equal Micro-Recall Overall Accuracy in certain micro-recall multi-class settings when micro-recall class distributions are considered.
  • It can typically serve as a component for computing Micro-F1 Measures when combined with micro-precision metrics.
  • ...
  • It can often yield Micro-Recall Different Values than macro-recall metrics depending on micro-recall class imbalance.
  • It can often be preferred when Micro-Recall Overall Detection Rate matters more than micro-recall per-class sensitivity.
  • It can often be computed efficiently from Micro-Recall Global Counters without micro-recall per-class calculations.
  • It can often be interpreted as Micro-Recall Instance-Level Sensitivity rather than micro-recall class-level sensitivity.
  • ...
  • It can range from being a Low Micro-Recall Metric to being a High Micro-Recall Metric, depending on its micro-recall detection quality.
  • It can range from being a Conservative Micro-Recall Metric to being a Liberal Micro-Recall Metric, depending on its micro-recall prediction threshold.
  • It can range from being a Binary-Derived Micro-Recall Metric to being a Native Multi-Class Micro-Recall Metric, depending on its micro-recall computation method.
  • It can range from being a Stable Micro-Recall Metric to being a Volatile Micro-Recall Metric, depending on its micro-recall temporal variance.
  • It can range from being a Balanced Micro-Recall Metric to being an Imbalanced Micro-Recall Metric, depending on its micro-recall class distribution.
  • ...
  • It can be calculated using Micro-Recall Formula: ΣTP / (ΣTP + ΣFN) across all classes.
  • It can be visualized through Micro-Recall Performance Curves showing micro-recall threshold sensitivity.
  • It can be monitored through Micro-Recall Tracking Dashboards for micro-recall model evaluation.
  • It can be optimized using Micro-Recall Enhancement Techniques targeting micro-recall false negative reduction.
  • It can be compared with Micro-Recall Baseline Scores to assess micro-recall improvement.
  • ...
• Example(s):
  • Medical Diagnosis Micro-Recall Metrics, such as:
    • Cancer Detection Micro-Recall Metric for micro-recall disease screening tasks.
    • Rare Disease Micro-Recall Metric for micro-recall condition identification tasks.
    • Drug Response Micro-Recall Metric for micro-recall treatment prediction tasks.
  • Security System Micro-Recall Metrics, such as:
    • Intrusion Detection Micro-Recall Metric for micro-recall threat identification tasks.
    • Fraud Detection Micro-Recall Metric for micro-recall anomaly detection tasks.
    • Malware Detection Micro-Recall Metric for micro-recall security scanning tasks.
  • Information Retrieval Micro-Recall Metrics, such as:
    • Document Retrieval Micro-Recall Metric for micro-recall completeness assessment tasks.
    • Search Engine Micro-Recall Metric for micro-recall result coverage tasks.
  • Quality Control Micro-Recall Metrics, such as:
    • Defect Detection Micro-Recall Metric for micro-recall manufacturing inspection tasks.
    • Error Detection Micro-Recall Metric for micro-recall quality assurance tasks.
  • ...
• Counter-Example(s):
  • Macro-Recall Metric, which calculates recall for each class individually and then averages without considering class frequency.
  • Weighted Recall Metric, which applies explicit class weights rather than implicit frequency-based weighting.
  • Micro-Precision Metric, which measures positive predictive value rather than true positive rate.
  • Micro-F1 Measure, which combines micro-recall with micro-precision rather than measuring recall alone.
  • Class-Specific Recall Metric, which evaluates individual class recall rather than global aggregation.
• See: Recall Metric, Micro-Averaged Performance Measure, Macro-Recall Metric, Micro-Precision Metric, Micro-F1 Measure, Sensitivity Measure, True Positive Rate, Multi-Class Classification Task, True Positive, False Negative, Confusion Matrix.

---