Multi-class intrusion detection using machine learning algorithms

Implementation Details:

We have used KDD dataset with multiple class intrusion namely DDOS, U2R, R2L, Probe, Normal, Unknown

We have pre-processed dataset by using the following numbers instead of attack type (class value)
dos = [1]
u2r = [2]
r2l = [3]
probe = [4]
normal = [5]
unknown = [6]

We applied machine learning algorithm namely, Logistic regression, Decision Tree and Random Forest.
Out of which Logistic regression outperforms interms of accuracy

Python Demo