Table 3 The summary of all performance evaluation measure metrics of Multiclass classification models
From: Prediction of diabetes disease using an ensemble of machine learning multi-classifier models
Measures | Definitions | Formula |
|---|---|---|
Average accuracy | The classifier's average per-class effectiveness | \(\frac{{\sum }_{i=1}^{k}\frac{{tp}_{i}+{tn}_{i}}{{tp}_{i}+{tn}_{i}+{fp}_{i}+{fn}_{i}}}{k}\)(8) |
Micro-averaging | ||
Precision | The genuine class labels' average per-class agreement with the classifier's labels | \(\frac{\sum_{i=1}^{k}{tp}_{i}}{\sum_{i=1}^{k}{tp}_{i}+{fp}_{i}}\)(9) |
Recall | A classifier's average per-class efficacy in identifying class labels | \(\frac{\sum_{i=1}^{k}{tp}_{i}}{\sum_{i=1}^{k}\left({tp}_{i}+{fn}_{i}\right)}\)(10) |
F1-score | The macro-average precision and recall's harmonic mean | \(\frac{2\times Precision\times Recall}{Precision+Recall}\)(11) |
ROC (AUC) | Receiver operating characteristics (ROC) with the area under the ROC curve (AUC also measured the ranking of predictions rather than their absolute values) | |