Privacy Implications Of Keystroke Dynamics

The popularity of Internet and the rapid development of information technology facilitate people’s life significantly; at the same time, new challenges are rising as well. When users use various kinds of information technologies on computers, the leakage of personal information and privacy becomes a major concern for users. Some companies make a great deal of profit by selling users’personal information to advertisers.There exist several approaches retrieving users’personal information at present, such as by IP addresses, HTTP Cookies, User-Agent Header and so on. However, the problems of privacy leakage aroused by these techniques are mainly due to the technical designs and implementations. Since the user’s typing pattern has been evaluated as a stable and unique characteristic to some extent previously, we want to investigate the privacy leakage issues when the users are typing on a computer.In order to accomplish our goal, we first examined the feasibility of techniques for eavesdroppers and look into the typing data collected by them. Next, we extracted typing features from the typing data and analyze the user’s typing patterns based on typing features. We propose a novel feature ranking approach to select such features and evaluate this approach in the two types of matching circumstances:many test samples vs. many training samples and few test samples vs. many training samples. Moreover, we study whether a user’s typing pattern on different typing devices is consistent and introduce a clustering method to measure the similarity of typing patterns between two users. Last but not least, we extend a user detection model of revealing user’s identity by incorporating more types of features and demonstrate better model accuracy. Our aim is to raise awareness of the privacy implications to general computer users.