USER IDENTIFICATION THROUGH KEYSTROKE BIOMETRICS AT AN INTERNET SCALE
Loading...
Authors
Veazey, Mark W.
Subjects
artificial intelligence
machine learning
neural networks
keystroke biometrics
keystroke dynamics
authentication
identification
cyber security
fingerprinting
machine learning
neural networks
keystroke biometrics
keystroke dynamics
authentication
identification
cyber security
fingerprinting
Advisors
Monaco, John
Date of Issue
2019-12
Date
Publisher
Monterey, CA; Naval Postgraduate School
Language
Abstract
Identification of users on the internet has broad-reaching implications in the computer science discipline regarding cyber security and privacy. Keystroke biometrics leverages the unique dynamics of how a user types to perform identification; however, current methods of authentication and identification using keystroke dynamics do not scale well beyond a few hundred users. This thesis investigates the feasibility of using conventional machine learning and deep learning techniques to identify users at an internet scale. By analyzing free-text keystroke information from a collection of over 100,000 users, several methods to perform user identification and profiling are identified, with a focus on determining how the size of the dataset affects identification accuracy. This thesis includes a novel method of representing keystroke data in a two-dimensional format suitable for a convolutional neural network, and it examines to what extent keystroke biometrics has implications for privacy on the internet.
Type
Thesis
Description
Series/Report No
Department
Computer Science (CS)
Organization
Identifiers
NPS Report Number
Sponsors
Funder
Format
Citation
Distribution Statement
Approved for public release; distribution is unlimited.
Rights
This publication is a work of the U.S. Government as defined in Title 17, United States Code, Section 101. Copyright protection is not available for this work in the United States.