Model Bias

Underfitting happens when a machine learning model is too simple to capture the underlying structure of data, leading to poor performance on both training and test sets.