Prabhjyot's Status Report - #5 of 17

[Prabhjyot Singh Sodhi <prabhjyotsingh95 () gmail com>]

Hey list,

This week had a bunch of twists and turns. As abhishek, I too kind of
deviated from my objective. We have been thinking about alternative methods
for data representation in the prints.

Accomplishments
* performed experiments and found the set of parameters that work well with
the dataset.
* Tried the classifier implementation in the python lib sklearn. Got a
295/301 accuracy on the training sample.
* Thought about the current representation of the dataset and discussed an
alternate representation. (this is related to the reducing of groups I had
mentioned last week)
* Pushed gsoc_trial (git repo that I use) to nmap-private-dev/fp/ipv6 to
allow other nmap commiters to view my progress.

Priorities
* Identify problems with opencv (accuracy is just 115/301, on similar
parameters)
* Try out the new dataset representation and the resulting accuracy.
* Try out a training testing ratio of 80:20, to better gauge how accurate
the models are.

Cheers,
Prabhjyot

_______________________________________________
Sent through the dev mailing list
https://nmap.org/mailman/listinfo/dev
Archived at http://seclists.org/nmap-dev/