Free Information Center
Underfitting is a phenomenon in machine learning where a model is too simple to capture the underlying patterns in the data, resulting in poor performance on both training and unseen data. It contrasts with overfitting, where a model captures noise as if it were signal.
The bias–variance tradeoff is a fundamental concept in machine learning and statistics describing the balance between the error introduced by the model’s assumptions and the error introduced by sensitivity to fluctuations in the training data. It explains how models with high bias oversimplify the data, while models with high variance overfit the data.
Ablation in neural network interpretability involves systematically removing components to assess their impact on model performance, aiding in understanding model decisions.
AUC-ROC is a performance measurement for classification models at various threshold settings. It represents the area under the Receiver Operating Characteristic curve, summarizing the trade-off between true positive and false positive rates.