byrdofafeather/narc

Senior Product- A Canvas LMS anomaly detection algorithm

NARC

NARC 1.0.0* is designed to be an OS-Independent service, using a web-based model. Mainly it is an experiment designed to explore further developments of artificially intelligent cheat-detection solutions.

_{*1.0.0 is designated as this version is to be incompatible and unrecognizable from previous versions of NARC. It does not indicate a finished version or complete version of NARC.}

Built With

K-Means for Javascript (clusterfck): https://github.com/harthur/clusterfck/blob/master/lib/kmeans.js

Tensorflow for Javascript: https://github.com/tensorflow/tfjs

Tensorflow: https://github.com/tensorflow/tensorflow

Django Rest Framework: https://github.com/encode/django-rest-framework

Django: https://github.com/django/django

(Inspired from) Harvard/django-canvas-oauth: https://github.com/Harvard-University-iCommons/django-canvas-oauth

Bulma: https://bulma.io

Scikit-learn: https://github.com/scikit-learn/scikit-learn

fcm-django: https://github.com/xtrinch/fcm-django

pandas: https://github.com/pandas-dev/pandas

numpy: https://github.com/numpy/numpy

whitenoise: https://github.com/evansd/whitenoise

narc's Issues

This line of code breaks everything.

I thought that it should be documented somewhere that this line:

NARC/DataProcessing/collectors.py

Line 301 in 4c6c61f

if items['event_type'] == 'page_blurred':

breaks this entire application. Page leaves are the MOST important feature in describing cheaters, this is a faulty piece of code as it uses "==" to compare strings in python when instead x in y should be used.