From: BertSRC: transformer-based semantic relation classification
\(Accuracy = \frac{{{\text{correct}}\;{\text{predictions}}}}{{{\text{total}}\;{\text{predictions}}}}\) | Useful when target classes are well balanced |
\(Recall = \frac{True\;positives}{{{\text{True}}\;{\text{positives}} + {\text{False}}\;{\text{negatives}}}}\) | The ability of a model to find all relevant cases within a dataset |
\(Precision = \frac{True\;positives}{{{\text{True}}\;{\text{positives}} + {\text{False}}\;{\text{positives}}}}\) | The ability of a model to identify only the relevant data points |
\(F1 - score = \frac{{2\left( {precision \times recall} \right)}}{precision + recall}\) | Combination between Precision and Recall Used to punish extreme values |