Please advice me on the meta- classifier to apply for IoT intrusion detection using machine/deep learning. Thank you.
Long Short-Term Memory (LSTM), Logistic Regression have a try. Don't forget Gaussian Model Method.
Dear Dahiru Abubakar,
see "Chapter 6: IoT Cybersecurity Objectives, Risks, and Threats" and "Chapter 7: Cybersecurity Areas and IoT" in the document
„NISTIR 8200: Interagency Report on the Status of International Cybersecurity Standardization for the Internet of Things (IoT)“
https://nvlpubs.nist.gov/nistpubs/ir/2018/NIST.IR.8200.pdf
The document NISTIR 8200 is one of the most important sources of information on the subject of "IoT security".
See also the NTIA publication: „Catalog of Existing IoT Security Standards“
https://www.ntia.doc.gov/files/ntia/publications/iotsecuritystandardscatalog_draft_09.12.17.pdf
Best regards
Anatol Badach
Dear Dahiru,
Just to complement previous answers, you may check conventional ML algorithms already implemented by scikit-learn library on this topic at:
https://scikit-learn.org/stable/modules/outlier_detection.html
For DL, you can also try autoencoders, an example with Keras is here:
https://keras.io/examples/timeseries/timeseries_anomaly_detection/
Best regards,
Xavier
- Badges
- Science topic