This project contains a notebook for detecting anomalies in a real-world temperature sensor data from an internal component of a large industrial machine, published by Numenta Anomaly Benchmark. https://github.com/numenta/NAB
Following algorithms have been implemented for ambient_temperature_system_failure.csv data file:
- Markov Chains (Used for ordered sequential anomalies)
- Isolation Forest (used for unordered collective anomalies)
- One class SVM (used for unordered collective anomalies)
- RNN (used for ordered sequential anomalies)