This research is related to time-series data mining and/or sequential data mining. The goals are
1. designing a methodology to find out the behaviour patterns from either
a. the incoming network packets from remote machines or
b. the requested commands, for examples, chown and cp, made by remote users
2. clustering the remote machines or users into different alarm level groups based on its behaviour patterns
For the project-based students, this research does not only have to thinking about the methodology of clustering the incoming packets and/or the requested commands with time-series or sequential data mining theories/techniques, but also implement an experiment tools/system with some simulation/real data to demo the usefulness of the solution.
| 