Error in predicting Land Use classes for large area using Random Forest in R

المشرف العام

Administrator
طاقم الإدارة
I have been trying to predict Land Use Classes in R using Random Forest and Caret packages. My area of interest is 4300km2 (raster dimension X 4109:Y2800). I am working with 4 multispectral Landsat bands.

I was able to do the whole process and classify the target classes but some areas within my map were not properly classified. When I plot the results from the function prediction() I have like "gaps" within my final map. The whole process runs without any warning or error.

I would like to know which is the best way to predict Land Use classes for a a big dataset.

I have created my model

Model using the training dataset (caret package)


trControl = trainControl(method="cv",number=10, repeats =30, allowParallel=TRUE) rfGrid
 
أعلى