Description

1. In documentation body it says the default value of parameter percent10 is False. The real value is True (which is correctly documented in the one line description in the beginning). It should be fixed.

2. The documentation says

SF is obtained as in [2] by simply picking up the data whose attribute logged_in is positive, thus focusing on the intrusion attack, which gives a proportion of 0.3% of attack.

However, in [2] (A Geometric Framework for Unsupervised Anomaly Detection: Detecting Intrusions in Unlabeled Data (2002) by Eleazar Eskin, Andrew Arnold, Michael Prerau, Leonid Portnoy, Sal Stolfo) they only mention that

In order to make the data set more realistic we filtered many of the attacks so that the resulting data set consisted of 1 to 1.5% attack and 98.5 to 99% normal instances

They didn't write how they did it. Is it documented anywhere ? if it is, the proper reference should be given. Otherwise, the documentation should reference the paper only with regard to the filtering goal and not how filtering is done.