Typing Pattern Based User Authentication Using Python

Read full text Summarise with VIRAT
Authors: Harsh Mankar, Mayur Kolage, Aditya Nikam, Tauqeer Ahmed, Amit Chakrawarti
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Over the past decade, the number of people relying on digital platforms for banking, communication, education, and commerce has grown at an extraordinary pace. This shift has made the security of user accounts more important than ever. Passwords, which remain the most widely used method of authentication, have proven to be consistently unreliable — not because of flaws in cryptography, but because of the very human behaviors surrounding their use. People tend to choose passwords that are easy to remember, which also makes them easy to guess, and they often reuse the same password across different platforms, creating cascading vulnerabilities. This paper introduces a user authentication system built around the concept of keystroke dynamics — the idea that everyone types in a subtly distinct way, shaped by years of motor habit and individual cognitive rhythm. By observing how long a person holds each key and how quickly they move from one key to the next, the system constructs a behavioral fingerprint that is specific to each user. Two machine learning algorithms, namely Random Forest and Support Vector Machine, are employed to learn these patterns and determine whether a login attempt is made by the legitimate account holder or by an unauthorized party. Developed entirely in Python, the system features an intuitive graphical interface for user registration and login, a SQLite database for secure profile storage, and a real-time classification engine. Testing conducted with twenty volunteers yielded an authentication accuracy of 92%, a precision of 90%, and a recall of 91%, confirming that the system performs reliably without requiring any additional hardware. Beyond one-time login verification, the system also supports continuous authentication during an active session, making it significantly more resistant to session hijacking compared to conventional approaches.

Fields of Study & Subjects

User Authentication and Security Systems Biometric Identification and Security Advanced Authentication Protocols Security

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
30 March 2026
Language
EN