| CompareClassification {greenbrown} | R Documentation |
This function computes an agreement map of two classifications (RasterLayers with classified values). Additionally, it computes a frequency table with user, producer and total accuracies as well as the Kappa coefficient.
CompareClassification(x, y, names = NULL, samplefrac = 1)
x |
First raster layer with classification. |
y |
Second raster layer with classification. |
names |
a list with names of the two classifications and class names. See example section for details. |
samplefrac |
fraction of grid cells to be sampled from both rasters in order to calculate the contingency table |
The function returns a list of class "CompareClassification" with the following components:
raster a raster layer indicating the agreement of the two classifications.
table a contingency table with user, producer and total accuracies. Rows in the table correpond to the classification x, columns to the classifcation y.
kappa Kappa coefficient.
Matthias Forkel <matthias.forkel@geo.tuwien.ac.at> [aut, cre]
plot.CompareClassification, AccuracyAssessment, TrendClassification
# Example: calculate NDVI trends from two methods and compare the significant trends
# load a multi-temporal raster dataset of Normalized Difference Vegetation Index
data(ndvimap)
# calculate trends with two different methods
AATmap <- TrendRaster(ndvimap, start=c(1982, 1), freq=12, method="AAT", breaks=0)
plot(AATmap)
STMmap <- TrendRaster(ndvimap, start=c(1982, 1), freq=12, method="STM", breaks=0)
plot(STMmap)
# classify the trend estimates from the two methods into significant
# positive, negative and no trend
AATmap.cl <- TrendClassification(AATmap)
plot(AATmap.cl, col=brgr.colors(3))
STMmap.cl <- TrendClassification(STMmap)
plot(STMmap.cl, col=brgr.colors(3))
# compare the two classifications
compare <- CompareClassification(x=AATmap.cl, y=STMmap.cl,
names=list('AAT'=c("Br", "No", "Gr"), 'STM'=c("Br", "No", "Gr")))
compare
# plot the comparison
plot(compare)