- Author: Sebastian Garcia (eldraco@gmail.com, @eldracote)
An anomaly detector for numbers. You can give a file with one number per line and it will print the anomalies. You can also give numbers in the STDIN and it will print the anomalies every X amount of numbers.
Please install the following dependencies:
- pyod: PyOD is a comprehensive and scalable Python toolkit for detecting outlying objects in multivariate data.
pip install pyod
$ ./number_anomaly_detector.py -f test-numbers.txt
Simple Number Anomaly Detector. Version: 0.1
Author: Sebastian Garcia (eldraco@gmail.com)
Top anomalies
values score
35345 3.129754
24562 1.766415
2 1.338806
$ cat test-numbers.txt | ./number_anomaly_detector.py
Simple Number Anomaly Detector. Version: 0.1
Author: Sebastian Garcia (eldraco@gmail.com)
Top anomalies
values score
35345 1.999886
Top anomalies
values score
35345 3.331550
2 1.284752
Top anomalies
values score
35345 3.129754
24562 1.766415
2 1.338806
Using the PCA model, is capable of training and testing 1 million numbers in 0.38 seconds
To run the tool using Docker use our Docker Hub image:
docker run stratosphereips/number_anomaly_detector:latest number_anomaly_detector.py -f test-numbers.txt