Table 3.

Comparison of load disaggregation algorithms.

Learning Algorithm	Features Sta/Trb	Accuracy %	Training Sc/Ud	Online/Offline	Scalability	Appliance Types
SVM [11,17,33,54]	B e	75–98	S	Online	Yes	I, II, III & IV
Bayes [12,50,54]	St	80–99	S	B	No	I & II
HMM [49,59,60]	St	75–95	B	Offline	No	I & II
Neural Networks [17,37,61]	B	80–97	S	Online	Yes	I & II & III
KNN [6,9,62]	B	70–90	S	B	Yes	I & II
Optimization [7,18,20,35]	St	60–97	S	Offline	No	I & II

^aSteady-State

^bTransient

^cSupervised

^dUnsupervised

^eBoth.