You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Dear Team,
I am using luminol for the larger dataset which has almost 25k rows and the issue is by using algorithm as "bitmap_detector" ,the anomaly score for the initial 200 values and the last 200 values remains as "0.0" even if there is some difference within the input values .
If we try to pass only those 200 values we receive some anomaly scores. please respond asap.
Thanks,
Prabhat
The text was updated successfully, but these errors were encountered:
Dear Team,
I am using luminol for the larger dataset which has almost 25k rows and the issue is by using algorithm as "bitmap_detector" ,the anomaly score for the initial 200 values and the last 200 values remains as "0.0" even if there is some difference within the input values .
If we try to pass only those 200 values we receive some anomaly scores. please respond asap.
Thanks,
Prabhat
The text was updated successfully, but these errors were encountered: