Short Answer

Comparing dataset characteristics after validation failure

Question: A spam filter does well on data drawn from the same source as its training set, but it performs poorly on a validation collection gathered from a different region. What should you examine first to diagnose the mismatch?

Sample answer: You should compare the training set and validation set to figure out which data characteristics are different between their distributions.

Key points:

  • Compare the two datasets directly.
  • Look for differences in data characteristics.
  • Focus on how the distributions are not the same.

Rubric: The response must state that you should examine which characteristics differ between the training and validation data distributions.

0

1

Updated 2026-08-12

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