When using the classification method called KShape by tslearn, it is described as eResumed cause of empty cluster と
In this case, is it a problem with the data set you are using, or is it possible to categorize it better by improving the number of times you learn?
I'd like some advice.
Thank you for your cooperation.
from tslearn.clustering import KShape
from tslearn.preprocessing import TimeSeriesScalerMeanVariance
seed = 0
np.random.seed(seed)
sz=train_data.shape[1]
train_data.shape# (400, 1,300)
ks=Kshape(n_clusters=2, n_init=30, verbose=True, random_state=seed)
y_pred=ks.fit_predict(train_data)
# Here, it is called Resumed cause of empty cluster.
If you wait even if the display appears, the result will be output.
I don't know if it's a factor, but
I use data shape as (number of data, length of time series, 1)
© 2024 OneMinuteCode. All rights reserved.