A Novel Machine Learning Approach to Preventing Account Takeovers for Enhanced Cybersecurity
CSEF · 2023 Behavioral & Social Sciences Honorable_mention Award
Overview
Account takeovers (ATOs) are a common type of identity theft. A 2021 study estimated that 38% of U.S. consumers were affected by ATOs in the prior two years, with an average cost of $12,000 per case. Preventing these attacks is becoming increasingly important as more people fall victim to ATOs. A current method to help combat ATOs is Multi Factor Authentication (MFA). However, MFA has several drawbacks including high cost, difficulty of use, compromised privacy, and lack of user awareness. The objective of this project is to build an effective machine learning model that predicts whether the person logging in to a system is an attacker or a valid user. Machine learning is well equipped to handle this task as it can account for many parameters and make accurate decisions quickly. Logistic regression, K-nearest neighbors, and decision tree models were trained on 25,000 data points associated with login attempts collected by Los Alamos National Laboratory to detect the validity of login attempts using the following parameters: time of login, source computer, destination computer, and user ID. The most successful model was the logistic regression model, achieving a testing accuracy of 99.34%. This method could complement or replace MFA to prevent ATOs. Given the high accuracy, it can not only save time and money but also prevent psychological stress for victims. Furthermore, the model conserves resources that companies would otherwise expend handling ATOs. Given the high accuracy of the model, this technique could prove useful in other areas such as wifi networks protection.
Source coverage
This record comes from a published award list, not a complete project archive. Its abstract comes from CSEF's public project showcase as archived by the Internet Archive before judging (https://web.archive.org/web/20230401224130/https://ca-csef.zfairs.com/showcase/ShowcaseInfo?f=838e60b7-ea75-46e8-865c-fde4864244b3); the version presented may differ.
Awards (1)
- Category Award: HM
Competition history
- CSEF 2023
Resources
Related projects
AJAS · 2024
Building a Deep Neural Network to Automate Protection Against Online Cyber Attacks
ISEF · 2017
Behave: An Authentication System Based on Keyboard and Mouse Dynamics
ISEF · 2020
Cyber Smart Security: Artificial Intelligence for Network Intrusion Detection Systems
ISEF · 2016
Using Machine Learning to Detect Computer Network Security Threats
JSHS · 2022
Luster Regained in Bits and Bytes: A Novel Cyber Incident Risk Prediction Model for Middle and High Schoolers Using Machine Learning
CSEF · 2026
Hybrid CNN-LSTM Network Intrusion Detection System Utilizing Multi-Task Learning and Simulated Network Defense
CSEF · 2026
A Novel Mechanistic Framework for Identifying and Neutralizing Latent Deception Signatures in Large Language Models
ISEF · 2019
Developing a Twitter 'Bot' Identification Application for Public Use
Closest projects by meaning, across every fair and year in the corpus.
Browse more like this
Source: California Science & Engineering Fair public projects