How to improve Classification/Regression generalization?
2 views (last 30 days)
Show older comments
Hi,
I have a dataset 6x20000 as inputs and 100 outpus for a class problem.
The problem is that, the dataset is changing every second (not so much, but is different than the initially trained one).
Is there a way (a class/regress problem), to solve this kind of scenario, without counting with the "generalization" characteristic ?
Example:
Original trained input and output: 10 15 5 2 3 41 => 1
Original test input (not on the trained dataset) and output: 9 15 4 2 3 40 => 1.32
** "Changed" trained input: 9-11 14-16 4-6 1-3 2-4 2-5 40-42 => 1
Thank you.
André
0 Comments
Answers (0)
See Also
Categories
Find more on Pattern Recognition and Classification in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!