Support for joined multi-variables profile classification
gmaze opened this issue · comments
Guillaume Maze commented
It is necessary to be able to select more than one variable to be used to classify profiles.
Use case:
# Load a dataset with a collection of profiles with multiple variables
ds = xr.open_dataset('PROFILES.nc')
# Init a PCM
m = pcm(K=8, axis=np.arange(0,1000,5), var_list=['TEMP','PSAL'])
# Fit the PCM on the dataset
m = m.fit(ds)
This implies that the pcm package will consume xarray.dataset (see #2 )
Guillaume Maze commented
This is already in the design but not some proper work with regard to pre-processing in order to classify 2 variables in the appropriate reduced space.
Guillaume Maze commented
This is done since version 0.4: b57da1a