User identification method based on multi-kernel learning fused with mouse and keyboard behavior characteristics

Abstract

The invention discloses a user identification method based on multi-kernel learning fused with mouse and keyboard behavior characteristics. In an uncontrolled environment, human-computer interaction behaviors of a user are monitored through a program to obtain real data generated by keyboard and mouse operations in daily work, two characteristics of a mouse and a keyboard are extracted, characteristic fusion and efficient classification are carried out through MKL, and finally efficient identity authentication of the user is achieved.

Description

technical field [0001] The invention belongs to the technical field of identification, and in particular relates to a user identification method based on multi-kernel learning (MKL) fusion of mouse and keyboard behavior characteristics. Background technique [0002] With the rapid development of the Internet era, people's lives are closely related to the Internet. Information systems can bring convenience to people's lives, but security has become a concern for the majority of people, and how to protect users' privacy has become a problem. An important research topic. Whether it is for an enterprise or an individual, device identity authentication is unavoidable. Identity authentication is to identify the user's operating authority to the system, for example, log in to the computer system, enter the mobile phone system, and log in to the bank. For business, etc., the simplest and most traditional authentication method is user name and password authentication. After the user ...

AI Technical Summary

Problems solved by technology

[0003] Identity authentication has always been a very popular research topic in the academic field. As a new identity authentication technology, biometric authentication has been widely used in various fields. Compared with traditional technologies, biometric authentication has higher reliability. Features include physiological features and behavioral features. Physiological features become the focus of authentication due to their unique characteristics, mainly including fingerprint features, iris features and face features, etc. However, the use of physiological features not only requires additional hardware support, but also requires verification and authentication. Individual feature collection cannot achieve continuous authentication without interference, while the behavioral feature biometric system uses commonly available computer interface devices to collect data, such as keyboards and mice, and the behaviors taken during human-computer interaction do not require additional assistance. equipment, so it has the advantages of low cost

[0004] So far, user behavior characteristics have been extensively studied, mainly including keyboard dynamics and mouse dynamics. A large number of studies on keystroke dynamics and mouse dynamics have mostly adopted machine learning algorithms and statistical learning methods. With the continuous improvement of this method, the existing research has achieved remarkable results. However, due to the very large differences in various data acquisition and data processing schemes, there are also great differences in recognition effects, and the combination of mouse dynamics and keyboard dynamics The research is only a small part. In reality, there are great differences in the behavior of users operating the mouse and keyboard within a certain period of time.

Images