Replication data for: Collection and statistical analysis of a fixed-text keystroke dynamics authentication data set
Data set for keystroke dynamics authentication benchmarking and research, containing 6 passwords typed by a wide set of people, containing a large set of "attackers" and a smaller set of "legitimate users". This data set was collected for the paper "Collection and statistical analysis of a fixed-text keystroke dynamics authentication data set" for the CSNet23 conference.
Article Abstract :
Keystroke dynamics authentication is a promising method of improving account security with minimal detriment for user convenience. While there is an abundance of research, there is a lack of available data sets. In this study, data sets for keystroke dynamics authentication were collected for a set of 6 passwords from a group of participants, and a correlation algorithm was developed to analyze and use these data sets for authentication. The experiments aim to produce data for keystroke dynamics authentication benchmarking, and to show the effect of typing speed and consistency, password length and entropy on prediction accuracy. Through simple correlation methods, the authors achieve an Equal Error Rate varying between a range of 2.57% and 29.7%. These result give insight into what may cause the accuracy to vary depending on the person and the password.
Indra Navia AS
ReadMe file: Copy and open the link,fill in the form and upload it with the rest of the datafiles: https://bit.ly/3JzW5EA