Learning Curves

Learning Curves

Plot error of cross validation set and training set against number of samples (m) by artificially reducing m to small size (ie. 10, 20, 30 samples)

High bias: getting more training data will not help much, the error gap is small but the error is high

High variance: large gap between cross validation and training set error, getting more training data is likely to help reducing the error gap

Resources:
- https://www.coursera.org/learn/machine-learning/supplement/79woL/learning-curves