Table 2.

Classification performance of four different classifier models based on DMPs identified from the placenta tissue of children with autism by signal detection.

Group	Accuracy	Sensitivity	Specificity	Detection Rate	FDR	Classifier
G1	0.995	0.998	0.966	0.906	0.004	PCA-QDA
G2	1.000	1.000	1.000	0.994	0.000	PCA-QDA
G2 pred. G1	0.935	0.996	0.332	0.904	0.064	LDA
G1 pred. G2	1.000	1.000	1.000	0.994	0.000	LDA

G1: The classifier models built based on Group 1 individuals (60% of DMP were used as training set), and then used to classify the rest (40%) within Group1. G2: The classifier models built based on Group 2 individuals (60% of DMP were used as training set), and then used to classify the rest (40%) within Group 2. G2 pred. G1: The classifier models built based on Group 2 individuals were used to classify the individuals within Group 1. G1 pred. G2: The classifier models built based on Group 1 individuals were used to classify the individuals within Group 2. All four classifier models performed well.