CHEATAH An Automatic Cheating Detection System Using ML 

Overview

With the advent of technology, it introduces a new era unprecedented in human history - in how we go from place to place, how we develop modern medicine, to how we communicate. Having an unregulated and broad space to communicate with anyone without the constraints of distance is a double edge sword; with one prevalent issue of cheating being made easier than ever in academia. The objective of Cheatah, our automatic cheating-detection algorithm, aims to use machine learning to help report and issue cheating among students. This program, which uses an easy to use website, encompasses 5 distance algorithms that checks a dataset of solutions from uploaded csv files. The results are then graphed with the help of clustering methods such as k-means and hierarchical that can help identify groups of cheating among students. Other features include the use of ensemble methods that employs the mixture of the five distance algorithms for the most optimal solution and hyperparameters that ensures the machine to determine a similarities amongst student solutions with the utilization of various thresholds. 

Community Benefit 

This program will help decrease the amount of cheating among students and assist in identifying by groups on potential cheating through similarity comparisons using the exam files. With this inclusion, the Cheatah website ensures the school admins in analyzing the groups of students detected in cheating after exams and swiftly resolve the issues. 

Team Members

• Anafiel Marie Abad
• Colin Hrzich
• Rayan Rabbi
• Billy Revelo
• Cale Schad

Sponsored By

• Dr. KwangSoo Yang