26 lines (20 loc) · 679 Bytes

Sequential Anomaly detecion

Anomaly detection for Sequential dataset

Model

ML
- Isolation Forest
- One-Class SVM
Autoencoder
- LSTM Autoencoder
- NN Autoencoder
- Unsupervised log message anomaly detection (2 Autoencode + Isolation)

Will do

Prediction
- NN
- LSTM
- Bidirectional LSTM
- Residual Bidirectional LSTM

Reference

Dataset, LSTM Autoencoder : https://github.com/BLarzalere/LSTM-Autoencoder-for-Anomaly-Detection
Gas Sensor Dataset : https://archive.ics.uci.edu/ml/datasets/Gas+Sensor+Array+Drift+Dataset
Unsupervised log message anomaly detection : https://www.sciencedirect.com/science/article/pii/S2405959520300643#b19