Emerging technology demands reliable authentication mechanisms, particularly in interconnected systems. Current systems rely on a single moment of authentication, however continuous authentication systems assess a users identity utilizing a constant biometric analysis. Spy Hunter, a continuous authentication mechanism uses keystroke dynamics to validate users over blocks of data. This easily-incorporated periodic biometric authentication system validates genuine users and detects intruders quickly. Because it verifies users in the background, Spy Hunter is not constrained to a password box. Instead, it is flexible and can be layered with other mechanisms to provide high-level security. Where other continuous authentication techniques rely on scripted typing, Spy Hunter validates over free text in authentic environments. This is accomplished in two phases, one where the user is provided a prompt and another where the user is allowed free access to their computer. Additionally, Spy Hunter focuses on the timing of different keystrokes rather than the specific key being pressed. This allows for anonymous data to authenticate users and avoids holding personal data. Utilizing a couple K-fold cross-validation techniques, Spy Hunter is assessed based on how often the system falsely accepts an intruder, how often the system falsely rejects a genuine user, and the time it takes to validate a users identity. Spy Hunter maintains error rates below 6% and identifies users in minimal numbers of keystrokes. Continuous authentication provides higher level security than one-time verification processes and Spy Hunter expands on the possibilities for behavioral analysis based on keystroke dynamics.
Employee theft and dishonesty is a major contributor to loss in the retail industry. Retailers have reported the need for more automated analytic tools to assess the liability of their employees. In this work, we train and optimize several machine learning models for regression prediction and analysis on this data, which will help retailers identify and manage risky employees. Since the data we use is very high dimensional, we use feature selection techniques to identify the most contributing factors to an employee's assessed risk. We also use dimension reduction and data embedding techniques to present this dataset in a easy to interpret format.