Suppose I train a machine learning algorithm with instances and labels from a set of attribute rasters and known points (say 1600 points):
label: forest fire location with lat-long
attributes: rainfall, temperature, slope..... etc. with lat-long
after training and testing the ML algorithm, I get a satisfactory result. Now, how am I supposed to translate the result of ML to all the cells/ lat-log points of the output raster as prediction model?
One possibility is to spatially interpolate several known points and create the output raster. Is there any other way?
