This project explores different methods for anomaly detection in time series data, specifically focusing on CPU utilization data from an AWS EC2 instance. The following techniques are implemented and compared:
- Mean Absolute Deviation (MAD)
- Isolation Forest
- Local Outlier Factor (LOF)
We use real-life data on CPU utilization of an AWS EC2 instance, recorded every 5 minutes starting from February 14th at 14:30 PM. The dataset contains 4032 data points and is sourced from the Numenta Anomaly Benchmark (NAB) repository, available under the AGPL-3.0 License.