Prabhjyot's Status Report - #4 of 17

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

Hi all,

Started coding this week (yay!).

Accomplishments
* Implemented a basic random forest (using the opencv library) on the
entire nmap ipv6 fp dataset.
  i) This model performed way below expectation.
  ii) Trained a similar model on a simpler and smaller dataset (3 groups,
36 records). The basic random forest hence created manages to get a 100%
accuracy on the smaller dataset.
* Discovered and helped fix a small bug in opencv

Priorities:
* perform experiments with different set of parameters for the random
forest, to produce an accurate random forest classifier. I will try
different values for the termination criteria (number of trees in the
forest),  maximum depth of those trees, number of samples to use for each
tree et cetera.
* Increase the training size to the entire ipv6 fp dataset.
* Think about reducing the number of groups in the entire dataset (and if
it would make more sense compared to the current setting)

Cheers,
Prabhjyot

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