(Not to be confused with pyfire
which is
completely different and is actual useful
software)
Fuzzy Intrusion Detection Engine written in Python. For a Masters assignment; ingests datasets similar to NSL-KDD dataset.
This is a reimplementation of the "Fuzzy Intrusion Recognition Engine"(FIRE) originally designed by Dickerson et. al.
Due to time constraints this is an offline system which must be trained first and then tested with static data files. It supports the NSL-KDD dataset and/or any datasets that are in the same format. NSL-KDD is a modified version of the original KDD-Cup '99 data that is popular within Network Intrusion Detection research. NSL-KDD claims to improve the false-positive rate across all machine learning algorithms tested, compared to the original KDD-Cup data. This code accepts the NSL-KDD data in its' ARFF version, which is more expressive than Comma-Separated Values (CSV).