NARX Input Target connection
1 次查看(过去 30 天)
显示 更早的评论
Hello together,
I want to use NARX to investigate the impact of an external force (wind) on a pillar. I have measurements of a deformation of the pillar at 3 different location over time (top, middle, bottom). I believe that NARX is a good tool, as the future deformation depends not only on the external force but also on the previous deformation.
My first try was to create a NARX with 1 Input and 3 Targets (data structure similar to the pollution mortality example data), but the results were acceptable but not perfect. I'm wondering right now if one target and its values are taken into account as "input" for the other two targets? For example, I guess that the deformation at top of the pillar (let's say target 1) has an impact on the deformation at the middle (target 2).
The question is if each target is treated independently from the others and just depends on the external force and its own previous values or not?
Or would it be better to create three NARXs, one for each measurement location, and then use the output of the "top location" NARX as Input for the other two? Does this make sense?
Does someone have similar experiences or ideas to clear the issue?
Best regards, Markus
0 个评论
回答(0 个)
另请参阅
类别
在 Help Center 和 File Exchange 中查找有关 Sequence and Numeric Feature Data Workflows 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!