Table 4.

Results of classification for the data of OCD.

Classifier	Accuracy (%)	Sensitivity (%)	Specificity (%)	AUC
SVM	93.01 ± 5.40	89.71 ± 9.22	95.08 ± 7.70	0.94 ± 0.06
LR-L1	89.81 ± 6.11	88.46 ± 10.23	91.47 ± 9.25	0.92 ± 0.07
LR-L2	90.58 ± 5.89	89.71 ± 9.22	91.29 ± 7.48	0.94 ± 0.06
GCN	91.41 ± 5.37	89.71 ± 9.22	92.72 ± 7.64	0.95 ± 0.06
MLP	90.64 ± 6.83	89.71 ± 9.22	91.29 ± 7.48	0.94 ± 0.06
XGBoost	85.77 ± 8.85	87.78 ± 11.19	84.84 ± 17.02	0.90 ± 0.12
GBDT	88.97 ± 7.23	86.71 ± 12.72	93.12 ± 9.49	0.94 ± 0.05

OCD, obsessive-compulsive disorder; SVM, support vector machine; LR-L1, sparse L1 for logistic regression; LR-L2, non-sparse L2 regularization for logistic regression; GCN, graph convolution network; MLP, multilayer perceptron; XGBoost, extreme gradient boosting; GBDT, gradient boosting decision tree; AUC, area under the receiver-operating characteristic curve.