Using soft_predict

[JPico6]

Greetings. Can you offer guidance on properly using the soft_predict method? I apologize if I missed something obvious. The output preds is always just the base rates, so in the case of 3 clusters, .333, .333, .333

X = np.array of dimension 80,000 x 140

from fcmeans import FCM

fcm = FCM(n_clusters=3, m=1.5, max_iter=100, random_state=1)
fcm.fit(X)
# note I've experimented with different params

pred = fcm.soft_predict(X)

print(pred)

Out[60]: array([[0.33333334, 0.33333333, 0.33333333],
[0.33333333, 0.33333333, 0.33333333],
[0.33333334, 0.33333333, 0.33333333],
...,
[0.33333334, 0.33333333, 0.33333332],
[0.33333333, 0.33333333, 0.33333333],
[0.33333334, 0.33333333, 0.33333333]])

[omadson]

Hi @JPico6. I hope it can still help you. Well I created a simple example with the soft_prediction method and it works fine. I really don't understand why it doesn't work with your data. Could you give more information about your data?