/cascoKNN

CASCO anomaly detection model for operational indicator data

Primary LanguagePythonBSD 2-Clause "Simplified" LicenseBSD-2-Clause

Anomaly Detection on Subway Operational Data

  • This is mini project during my internship in CASCO Shanghai, China, Summer 2018.
  • This project is implemented and modified based on the python anomaly detection tooklit pyod.
  • I created a new model called cascoKNN to process CASCO subway operational indicator data.

Installation

For more information, check out pyod.

git clone https://github.com/paragon520/Pyod_CASCO 
cd Pyod_CASCO
python setup.py install

Required Dependency

  • Python 2.7, 3.4, 3.5 or 3.6
  • numpy>=1.13
  • scipy>=0.19.1
  • scikit_learn>=0.19.1
  • matplotlib
  • nose

Test and run

Simply run:

cd example
python cascoKNN_example.py

View your result

  • ROC value and & rank will be printed on the cmd.
  • You can also save the scatter plots to png file.