Essay

Explain the methodology for constructing and plotting a dev-set error learning curve by varying training dataset size.

Question: Describe the systematic steps required to construct a learning curve that tracks dev-set error as training set size varies. In your response, explain how the models are trained on different sample sizes and how the final visualization is plotted.

Sample answer: To construct the learning curve, you run the algorithm using different training-set sizes. This is done by training separate copies of the algorithm on subsets of varying sizes (for example, if you have 1,000 total examples, you might train copies on 100, 200, 300, and more examples up to 1,000). Once each copy is trained, you evaluate its performance to determine the dev-set error. Finally, you plot the dev-set error on the y-axis against the corresponding training-set sizes on the x-axis to visualize how the error varies with the amount of training data.

Key points:

  • Train separate copies of the algorithm on different training-set sizes.
  • Measure the dev-set error for each trained copy.
  • Plot dev-set error against the training-set size.

Rubric: The answer must specify: 1) training separate copies of the algorithm on different training-set sizes, 2) evaluating these copies to measure the dev-set error, and 3) plotting dev-set error on the y-axis against training-set size on the x-axis.

0

1

Updated 2026-05-27

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Strategy

Machine Learning Yearning @ DeepLearning.AI

Related