Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[BUG] detect fake anomalies when the buckets are not completely full #1108

Open
MCdeamon7 opened this issue Dec 4, 2023 · 2 comments
Open
Labels
bug Something isn't working

Comments

@MCdeamon7
Copy link

MCdeamon7 commented Dec 4, 2023

I have generated fake data with a gauss function, the data are generated from ~8AM to ~6PM for 30 days with the related timestamp.
When the data starts or ends, I frequently get spikes that trigger the anomaly detection algorithm.
anomaly detection
anomalyDetectionSpike

Observing the data in "visualize" that depending on the bucket size the algorithm (avg) actually show random spikes:
spike

But zooming, the data are actually as expected.
zoom-visualize

To reproduce the bug, you can try to import data generated with my python script:
https://github.com/MCdeamon7/data-generator
import the data and then try to make an anomaly detection job on it

I imagine that is something about bucket size because by adjusting it from 5 to 10 or to 15 the spikes disappear in some places and reappear in other places.

I'm using it on a docker container on Linux recompiled by me from version 2.11.0 (i modified just the max features from 5 to 10)

Sorry for my English and thanks to everyone that can help me.

@MCdeamon7 MCdeamon7 added bug Something isn't working untriaged labels Dec 4, 2023
@minalsha
Copy link

@kaituo thoughts?

@kaituo
Copy link
Collaborator

kaituo commented Mar 11, 2024

@MCdeamon7 Can you list steps on how to generate data using https://github.com/MCdeamon7/data-generator ?

Also, what do you mean by bucket size/ What's your detector configuration?

@kaituo kaituo removed the untriaged label Mar 11, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

No branches or pull requests

3 participants