Application of a New Feature Generation Algorithm in Intrusion Detection System

Algorithm 1 mgfg method for continuous feature columns.

(i)	Input: //Continuous Feature Col
(ii)	m: the number of feature cols
(iii)	ε: 100
(iv)	cut: Converts the f column attribute to the Q class
(v)	Catboost_encoders: Converts the discrete attributes of the Q class to continuous values
(vi)	for i = 1 to m do
(vii)	temp = value_counts((fi))
(viii)	temp = (max(fi))//2
(ix)	If (temp > ε)
(x)	for q = 10 do
(xi)	temp_col = cut(fi, q)
(xii)	fi_q = Catboost_encoders(temp_col)
(xiii)	F.append(fi_q)
(xiv)
(xv)	end
(xvi)	end
	Output: //Feature columns with different granularity.