A Globally Generalized Emotion Recognition System Involving Different Physiological Signals

. 2018 Jun 11;18(6):1905. doi: 10.3390/s18061905

Algorithm 1 G-means algorithm.

1:
Input:X the set of data and $α$ confidence level = $90 %$ (Gaussian)
2:
Output: The new set of centers C (using K-means++ [40])
3:
Given C the initial set of centers
4:
$C \leftarrow k - m e a n s (C, X)$
5:
Let ${x_{i} | c l a s s (x_{i}) = j}$ be the set of the data assigned to center $c_{j}$
6:
Check if each ${x_{i} | c l a s s (x_{i}) = j}$ follow a Gaussian distribution (at confidence level $α$ )
7:
if The data follow a Gaussian distribution then
8:
Keep $c_{j}$
9:
else
10:
replace $c_{j}$ with two centers (using K-means++ [40]).
11:
end if
12:
Repeat from step 5 until no more centers are added.