Abstract
A new pattern classification approach based on multi-user communication techniques is proposed for improving multi-class learning performance. The spreading gain of code division multiple access is used instead of the coding gain of error-correcting output codes to increase classification accuracy. Results show that spread spectrum codes give better classification accuracy than error-correcting output codes up to 14.25%.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have