Sample sizing

In Chapter 8Creating a Facial Expression Platform on IBM Cloud , we constructed an expression analysis model using facial images for the model to train on and use to detect human emotions: happiness, sadness, and anger. In that exercise, we chose a sample size for each emotion of just 11 images:

In that exercise, the sample size (11) was sufficient since the goal was to prove our concept. Obviously, the bigger the sample size the more accurate the model.

A biometric authentication solution works on the same premise; however, rather than detecting emotion from an attempt, the model will compare the scanned image to its cataloged ...

Get Hands-On Machine Learning with IBM Watson now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.