Concept icon
Concept

Lower Error Rates Require Larger Review Sets

When a classifier makes fewer mistakes, a larger development sample is needed to collect enough wrong predictions for careful inspection. For example, if the error rate is 4%, then about 2,500 examples are needed to expect roughly 100 misclassified cases.

0

1

Concept icon
Updated 2026-08-12

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Machine Learning Strategy

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Yearning @ DeepLearning.AI