Der Mikro-Durchschnitt ist eine statistische Methode, die häufig bei Mehrklassenklassifikation tasks within the field of Künstliche Intelligenz and Maschinelles Lernen. It provides a way to assess the performance of a classification model by considering all classes together, rather than calculating metrics for each class separately. This approach is particularly useful when dealing with unausgewogene Datensätze eingesetzt wird, bei denen einige Klassen deutlich mehr Proben haben als andere.
In micro-averaging, the true positives, false positives, and false negatives are summed across all classes before calculating the overall precision, recall, or F1 score. This means that each instance contributes equally to the final metric, regardless of which class it belongs to. The formula for micro-average precision (P) and recall (R) can be expressed as:
- Micro-Durchschnittspräzision: P = (Summe der True Positives) / (Summe der True Positives + Summe der False Positives)
- Micro-Durchschnittsrückruf: R = (Summe der True Positives) / (Summe der True Positives + Summe der False Negatives)
Micro-average is particularly advantageous in scenarios where the focus is on overall model performance rather than on individual class performance. This can help in providing a more holistic view of a model’s capabilities, especially in applications such as Bildklassifikation, text categorization, and speech recognition.