Capstone 2, Spring 2020
Brian Taylor, Jon Henry, Robert Stevenson
Company: GE Aviation
Mentor: Eric Gero
Currently GE uses a system in which employee events are categorized and given a heat score. For example an employee sending an email to a competitor with an attachment or Employees accessing systems they do not normally need access to. However the heat scores currently hard-coded and static. This is a problem because situations can change fast, or events can be related to significantly impact the risk.
We plan to use a bayesian neural network and obfuscated data to generate heat scores that are more indictive to the individual preseved risk.
Language: Python Tensorflow
Trello: https://trello.com/b/zyP2VdJG/capstone2
Bitbucket: https://bitbucket.org/cps491s20-team8/capstone-2/