Case Study

Choose a warehouse sensor under a hard false-alarm limit.

Case context: A safety team is evaluating several sensors that detect smoke in a storage warehouse. Any sensor that raises more than 2 false alarms per day is too disruptive to use. The remaining sensors differ in how often they fail to detect a real fire.

Question: How should the team filter the sensors, and which metric should decide the final selection?

Sample answer: First remove every sensor that exceeds the 2-false-alarm-per-day limit, because those options do not satisfy the required operating constraint. Then compare the false negative rates of the sensors that remain and choose the one with the smallest false negative rate, since that is the metric to optimize among acceptable options.

Key points:

  • Use the false-alarm limit as a requirement for eligibility.
  • Discard any sensor that is too noisy.
  • Among the sensors that qualify, compare false negative rates.
  • Pick the qualifying sensor with the lowest false negative rate.

Rubric: The response should treat the false-alarm limit as a gatekeeper constraint and then select the eligible sensor with the lowest false negative rate.

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