Help me understand the Classifier App

2 views (last 30 days)
I am looking to create a classifier in Matlab, but I'm having some issues in understanding how exactly it works. I have several data sets of n x 9 arrays with readings from 3 separate 3-axis analog sensors (outputting values in m/s^2), and I want to create a classifier to distinguish between a state described by each of these arrays (walking, standing, running, etc.). I am having significant issue understanding how the classifier app in Matlab can train a model to do such a thing, however, and each time I attempt to "train" it, I get exceptionally poor model performance. I am obviously missing something key here, and I'm not sure where to go. I do have a fairly elementary understanding of classifiers, but I'm finding the documentation available to be confusing.
  2 Comments
Image Analyst
Image Analyst on 20 Apr 2019
Are you talking about the training and classify()/predict() functions, or the Classification Learner app?

Sign in to comment.

Accepted Answer

Stephan
Stephan on 20 Apr 2019
Hi,
see example 1 from this webinar:
Also note the links to the example code on this website.
Best regards
Stephan

More Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!