A Predictive Model for Guillain-Barré Syndrome Based on Single Learning Algorithms

Table 3

Four GBS subtypes’ classification. The standard deviation of each metric is shown in normal font.

Classifier	Optimal parameters	Average accuracy	Multiclass AUC
SVMPoly	coef = 1	0.9235	0.8985
		0.0080	0.0199
C4.5		0.9211	0.8857
		0.0109	0.0242
SVMLap		0.9201	0.8712
		0.0072	0.0240
SVMGaus		0.9193	0.8897
		0.0067	0.0221
NN		0.9179	0.8783
		0.0041	0.0188
SVMLin		0.9175	0.8632
		0.0096	0.0232
JRip		0.8999	0.8729
		0.0143	0.0291
Naive Bayes		0.8986	0.8632
		0.0079	0.0244
MLP		0.8974	0.8514
		0.0122	0.0257
SLP		0.8972	0.8452
		0.0147	0.0230
MLR		0.8926	0.8405
		0.0082	0.0279
LDA		0.8806	0.8256
		0.0083	0.0223
RBF-DDA		0.8797	0.8249
		0.0079	0.0287
OneR		0.7744	0.7528
		0.0164	0.0249