Home › Key concepts › Overfitting and underfitting
Key concept · Crawl (beginner) · the same on every platform
Overfitting and underfitting
Memorising the training data versus not learning enough from it.
What it is
An overfit model learns the noise in its training data and does badly on new data. An underfit model is too simple to capture the pattern. You see overfitting when training scores are much better than validation scores.
Why it matters
It is the central balance in machine learning.
An example
A deep tree scores 99% on training and 70% on validation; limiting its depth gives 85% on both.
The common mistake
Adding complexity until the training score is perfect.
Where you meet it on each platform
The idea stays the same; only the product name changes.
| Job | AWS | Azure | Google Cloud | Databricks | Open source |
|---|---|---|---|---|---|
| Trainingfit a model on data, on a laptop or a cluster | SageMaker training jobs | Azure ML jobs | Vertex AI training | Databricks ML runtime | scikit-learn, PyTorch |
Related ideas
Where it shows up
Part of the Crawl level on these paths: ML Engineering (learning basics).