Artificial Intelligence & Machine Learning
Underfitting
Definition
Underfitting is a modeling error in machine learning that occurs when a model is too simple to capture the underlying patterns in the data. It performs poorly on both the training data and new, unseen data.
Why It Matters
An underfitted model is not a useful model because it has failed to learn the relationships in the data.
Contextual Example
Trying to model a complex, curved relationship in data using a simple straight line (linear regression) would be an example of underfitting. The model is not complex enough for the data.
Common Misunderstandings
- The opposite of underfitting is overfitting.
- Underfitting can often be fixed by using a more complex model, adding more features, or training for longer.