2
$\begingroup$

I need to create model which will find suspicious entries or anomalies in a network, whose characteristics or features are the asset_id, user_id, IP accessed from and time_stamp.

Which unsupervised anomaly detection algorithms or models should I use to solve this task?

$\endgroup$
0

2 Answers 2

1
$\begingroup$

If you are OK to use python, thy novelty-detection with sklearn:

https://scikit-learn.org/stable/modules/outlier_detection.html

$\endgroup$
1
  • $\begingroup$ I tried this but the results are not that convincing $\endgroup$
    – Abishek
    Commented Mar 18, 2019 at 10:20
1
$\begingroup$

Hierarchical Temporal Memory is a model well suited for anomaly detection. It is also pretty interesting and different from currently typical Deep Learning models.

$\endgroup$
1
  • $\begingroup$ I would add the detail that HTM-based algorithms are better suited for cases where you expect a stream of data (i.e. you want to continuously learn and detect anomalies online). $\endgroup$
    – nbro
    Commented Mar 14, 2019 at 10:44

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .