top of page
Search
Andrew DeJonge, Miguel Millan

Week 8

Week Goals: Initialize HMMs with clustering to provide a good guess at beginning states and probabilities. Focus into DDoS data set to create a tiered system of HMMs to predict different attacks and subattacks better

 

Wednesday: Presented our weekly presentation at the group discussion, began working on a plan for the tiered HMM system. Looked into using clustering to initialize HMMs

 

Thursday:Successfully used K medoids clustering to initialize our HMMs and coded a tiered HMM process in Matlab to detect attacks, with each attack having its own set of features used to detect it and separate HMMs for each possible attack.

 

Friday: Began testing with the new tiered system. Worked on our last weekly presentation for the group meeting next Tuesday. Met with Coach, Aziz, and Ali over the presentation, what changes to make to it, and what we can still accomplish with the last two weeks of the REU program.


2 views0 comments

Recent Posts

See All

Week 10 - Final Week

Week Goals: Finish gathering results for the tiered and 2-dimensional HMMs. Go over drafts of research papers with Coach. Finish final...

Week 9

Week Goals: Finish testing the new tiered HMM system and obtain results for accuracy of the new system. Begin writing our paper over the...

Week 7

Goals: 1.) Finish extracting features from the IoTPOT Data set 2.) Find equations/publications for Hierarchical HMMs 3.) Gather...

bottom of page