Goals:
1.) Continue work on representing/predicting network attack data with HMMs in Matlab
2.) Be able to use HMMs and our programs on multiple data sets, including DDoS, Kyoto, IoTPOT, and others
3.) Begin to use HMM Train (Baum-Welch algorithm) more with our data to more realistically create an HMM without using/knowing any state sequence
4.) Increase focus on IoT in research/presentations
5.) Get features for every type of attack we want to represent in our model
6.) Start getting ideas for implementing intent into a 2-D HMM along with the attack types that we've been focusing on
Monday: Gave our midterm presentation at the weekly discussion and provided feedback on others' midterm presentations. Spent time updating website to contain all of the resources we've collected over the course of the 5 weeks. Worked some more with feature pruning in Matlab with the Kyoto data set and read up some more on HMMs.
Tuesday: Gained access to the IoTPOT data set and the data set used by the researchers who wrote the paper on Sybil Attack Detection. Began sorting through the data in the IoTPOT data set and looking into ways to sort the data into a form that could be usable for creating an HMM with wider usability due to the variety of attacks in the IoTPOT data set.
Wednesday:
Thursday:
Friday: