Correlation-Based Anomaly Detection Method for Multi-sensor System

Algorithm 1

Framework of data augmentation in the dataset.

Algorithm 1	: Data augmentation in the dataset.
	Input: The moving steps of data augmentation window lena; the size of data augmentation window ; the number of attributes of time-series data m; the data augmentation sequence array S; the original time-series data array D;
	Output: The time-series data array Da after data augmentation;
(1):	function DATAAUGMENTATION (D, S, m, wa, lena)
(2):	rm←0
(3):	i← 0
(4):	length←sizeof (D)
(5):	while i < length do
(6):	i←i + lena
(7):	For j = 0⟶ do
(8):	D[i + j][rm]←D[i + j][rm] + S[rm]
(9):	end for
(10):	if rm+ 1 <m then
(11):	rm←rm+ 1
(12):	else
(13):	rm←0
(14):	end if
(15):	end while
(16):	Da←D
(17):	eturn Da
(18):	end function