Goals:
1.) * Prepare for midterm presentation.
2.) Feature selection via systematic reduction.
3.) Create test data subset for testing accuracy.
4.) Build multiple HMMs for different features
5.) Use both Kyoto and DDoS set.
6.) Figure out how to detect intention and control.
7.) Keep searching for more data.
Monday: Presented our weekly update at the group discussion. Met with Coach, Aziz, and Ali to discuss what to do this week, as well as what sort of things to include on the Midterm Presentation. Worked more with the Kyoto and DDoS data sets in Matlab. Began writing code to select the best features from a data set.
Tuesday: Worked on the code to select best features from a data set and functions necessary to implement it. Worked on selecting features from the Kyoto data set, fixing labels, and finding which features were useful for implementation in an HMM.
Wednesday: Finished the program to select the best features from a data set for use in an HMM. Made contact with the researchers in charge of IoTPOT to get their IoT device-based honeypot data. Attempted training the HMM for the Kyoto data set, found another data set: the KDD Cup 99 data set. Discussed our thoughts/ideas on how to identify intent behind attacks and attacks on control.
Thursday: Began work on the midterm presentation and met with Coach, Aziz, and Ali to discuss what should/shouldn't be on the presentation, as well as what visual aids would work best. Worked on debugging the feature reduction code that was finished on Wednesday.
Friday: Code for feature reduction completely finished and ready to be implemented in whatever way necessary. Worked on and finished up the midterm presentation, including expanding on any explanations, adding visual aids, and giving examples of different concepts of HMMs and attacks on IoT/networks.