Intrusion Detection for Cyber-Physical Security System Using Long Short-Term Memory Model

Algorithm 1

Feature selection algorithm (RFP).

	Input: data frame
	Output: selected feature
	Procedure:
(1)	Creating a data frame with only numeric attributes of binary class and multiclass
(2)	Encoded label attribute
(3)	Finding the attributes which have more than 0.5 correlation with the encoded attack label attribute
(4)	Selecting attributes found by using the Pearson correlation coefficient
(5)	Joining the selected attribute with the one-hot encoded categorical data frame
(6)	Joining one-hot encoded, and original attack label attribute
(7)	Obtain the selected feature