(ML Spec)

A performance plot for a learning algorithm on a data set(s).

  • x-axis: the size of the dataset
  • y-axis: performance metric (e.g. error, accuracy)

Computational cost

It’s useful to plot learning curves for different algorithms (and their cost function) on different subset of the data, to diagnose Bias-Variance as a performance metric. However, it’s quite computationally expensive