Feature Selection and Overlapping Clustering-Based Multilabel Classification Model

Table 2

Hamming loss measure for the models based on -means clustering on emotions dataset. is the number of attributes, and is the amount of clusters. is the average value of Hamming loss.

Classification model			Emotions
-means + ML-KNN	2	72	0.2018
EM + ML-KNN	2	72	0.2019
DEP_max + -means + ML-KNN	2	%	0.2021
MDI + -means + ML-KNN	2	%	0.2016
-means + ML-KNN	3	72	0.2021
EM + ML-KNN	3	72	0.2013
DEP_max + -means + ML-KNN	3	%	0.2034
MDI + -means + ML-KNN	3	%	0.1986
-means + ML-KNN	4	72	0.2106
EM + ML-KNN	4	72	0.2030
DEP_max + -means + ML-KNN	4	%	0.2156
MDI + -means + ML-KNN	4	%	0.2024
-means + ML-KNN	5	72	0.2119
EM + ML-KNN	5	72	0.2027
DEP_max + -means + ML-KNN	5	%	0.2168
MDI + -means + ML-KNN	5	%	0.2005
-means + ML-KNN	6	72	0.2089
EM + ML-KNN	6	72	0.2030
DEP_max + -means + ML-KNN	6	%	0.2078
MDI + -means + ML-KNN	6	%	0.2056
-means + ML-KNN	7	72	0.2120
EM + ML-KNN	7	72	0.2064
DEP_max + -means + ML-KNN	7	%	0.2126
MDI + -means + ML-KNN	7	%	0.2030