Hi for a network traffic analysis task i need a probabilistic model to analysis sequences of network data. Euch observation is here an event consisting of structured Information (e.g. ip adresses, ports, protocol type). Im interested in the dependencies between these observations using a generative model. Any ideas?
I would have thought this would be difficult becasue, for example, TCP/IP chunks it's data up based on byte boundaries. I know increasingly network implementations use huge chunks, but it means you can still end up with 'half words'.
But as a suggestion you could juts do a count on a stream of all the patterns you see - youo could build it up using a Lempel-Ziv algorithm - that would at least give you a frequency distribution of patterns you see in a stream.
I may be missinterpreting your question though, and unless you can separate the packets out into indifidual streams I am not sure you will see much.
André Kummerow
Here, you might find a good starting point:
Article A Statistical Model for Event Sequence Data
- Badges
- Science topic