Table 1.

Notation

variables/symbols	description
	Preprocessing
Vc	compensation field strength
RT	retention time
X (I×J)	matrix representation of a chromatogram
U (I×J)	matrix of baseline corrected and aligned data
D (I×(J–4))	matrix of Savitzky-Golay first derivative of U, horizontally
E ((I–4) J)	matrix of Savitzky-Golay first derivative of U, vertically
I	number of scans in a chromatogram
J	number of compensation field strength channels in a chromatogram
N	number of samples
C	number of variables in an unfolded data matrix
	Peak Detection
t i × a	threshold at RT i
a	peak noise factor
m	position of the peak start in Vc dimension
r	position of the peak maximum in Vc dimension
s	position of the peak end in Vc dimension
g	position of the peak start in RT dimension
h	position of the peak end in RT dimension
Y (P × Q)	peak detail table
P	number of peaks detected in the entire voltagram
Q	characteristics of peaks found (peak start, peak maximum and peak end from RT and Vc dimension)
	Peak Merging
ϕ	overlap ratio
p	overlap region (in data points)
q 1	length of the target peak
q 2	length of the candidate merging peak
K init	yij – ycj (where i = 2;c = 1 and j = 1)
W	matrix recording the one-dimensional features of a 2D peak
Z (R × S)	peak region table
β1, β2,.. ., β7	peak maximum of each voltagram for example in Figure S-3, Supporting Information
δ	mean difference of a 2D feature
ω	number of 1D peaks forming a 2D feature
	Peak Matching
L (T × G)	matrix containing peaks in dimensions detected in all samples
T	the number of unique peaks detected
G(G = 4)	characteristics of each peak (origin of peak (indicates by 1,.. ., n), RT, Vc, and peak area
H	peak table (rows corresponding to variables and columns to samples)
F	matrix recording the RT and Vc of each unique peaks
V 1	tolerance window of RT (data points)
V 2	tolerance window of Vc shift (data points)
	Simulations
ν	underlying intensity at the peak maximum
φimax,φjmax	positions of the peak maxima in each dimension
σi, σj	width of each peak in each dimension