Apply pattern recognition output to another dataset

1 次查看(过去 30 天)
I have a DEM dataset that I have worked up into values such as slope, aspect, curvature, &c. At some points I have feature observations of the ground cover (bare, grass, trees). I now have a training dataset with the elevation, slope, aspect, curvature, &c. and a binary output with each ground cover as a separate column: [0,0,1;1,0,0;0,0,1;0,1,0] Means {'trees';'bare';'trees';'grass'}. I have run the pattern recognition tool from the neural network toolbox and the confusion matrix looks promising (>80% correct). How do I now apply the model developed to the whole DEM dataset?
  3 个评论
Tristan Graham
Tristan Graham 2018-1-22
编辑:Tristan Graham 2018-1-22
Thanks KSSV. Once I had my other data in the right format it worked very smoothly. Is there a way to clearly see how it is manipulating the input data? I would like to do some tests using different input combinations / spatial selections &c. and see how it affects the output.
KSSV
KSSV 2018-1-22
If your input data changes i.e if you change the combination.....you should train again and get new net.

请先登录,再进行评论。

回答(0 个)

类别

Help CenterFile Exchange 中查找有关 Pattern Recognition and Classification 的更多信息

Community Treasure Hunt

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

Start Hunting!

Translated by