Modeling users keyboard handwriting using a neural network
Article Sidebar
Main Article Content
Abstract
User identification is one of the main tasks of modern cybersecurity today. Secure login allows users to access sensitive data: bank accounts, tax authorities, social networks. Losing passwords or obtaining them by attackers can cause irreparable harm not only to individual citizens, but also to private and public organizations. In some cases, this damage can be irreparable. The most popular method is identification by login and password. But with the development of social engineering, an increase in the number of hacks and password leaks, this method ceases to be safe. Every year the number of cyberattacks aimed at stealing passwords only grows. Therefore, biometric methods have become popular in recent years. Such methods allow for fairly accurate identification of a person by fingerprints, facial biometrics. But such methods require specific sensors. As a rule, some of such sensors are already built into modern mobile devices, but are completely absent in standard desktop computers. The share of such computers still makes up more than half of all devices. Therefore, it is advisable to use behavioral biometrics ‒ keyboard handwriting, which reflects the individual characteristics of each person. It is important that to obtain the technical characteristics of keyboard handwriting, no additional sensors and devices are required. The article discusses the main characteristics of keyboard handwriting that will allow in the future to build a neural network for recognizing users by keyboard handwriting. Obtaining several technical characteristics will allow using more parameters for the neural network.
Article Details
References
2. Shaffer L.H. Reading and Typing ‒ https://www.researchgate.net/publication/233266615_Reading_and_Typing
3. Chalaya L.E. User identification model based on keyboard handwriting. «Artificial Intelligence», No. 4. 2004, Р. 811-817.
4. Saket Maheshwary, Soumyajit Ganguly, Vikram Pudi. Deep Secure: A Fast and Simple Neural Network based approach for User Authentication and Identification via Keystroke Dynamics. https://www.researchgate.net/publication/322952671_Deep_Secure_A_Fast_and_Simple_Neural_Network_based_approach_for_User_Authentication_and_Identification_via_Keystroke_Dynamics
5. Daus, Y., Samoilov, S., Daus, M., & Larin, D. (2025). User identification by keyboard handwriting using neural networks. Bulletin of the Odessa National Maritime University, (75), P. 212-227. https://doi.org/10.47049/2226-1893-2025-1-212-227.
6. Bishop C.M., Pattern Recognition and Machine Learning. – Springer New York, 2016 – 778 p.
7. The influence of the sigmoid function parameters on the speed of back- propagation learning / Han J., Moraga C. // Springer Berlin Heidelberg: materials Conference Computational Models of Neurons and Neural Nets. ‒ Berlin, 1995. ‒ Р. 195-201.