Table 3 Classification between individuals with schizophrenia, and ASD.

From: Machine-learning classification using neuroimaging data in schizophrenia, autism, ultra-high risk and first-episode psychosis

 

ASD and schizophrenia

(Subcortical)

(Surface area)

(Cortical thickness)

(All features)

Classifier

Score (%)

Schi

ASD

All

Schi

ASD

All

Schi

ASD

All

Schi

ASD

All

Logistic regression

Mean accuracy

  

75.0

  

70.0

  

80.0

  

70.0

Recall score

72.7

77.7

 

72.7

66.6

 

90.9

66.6

 

81.8

55.5

 

Precision score

80.0

70.0

 

72.7

66.6

 

76.9

85.7

 

69.2

71.4

 

F1/F2 scores

76.1/74.0

73.6/76.0

 

72.7

66.6

 

83.3/87.7

75.0/69.7

 

75.0/78.9

62.5/58.1

 

Support vector machine

Mean accuracy

  

80.0

     

75.0

  

75.0

Recall score

81.8

77.7

    

90.9

55.5

 

90.9

55.5

 

Precision score

81.8

77.7

    

71.4

83.3

 

71.4

83.3

 

F1/F2 scores

81.8

77.7

       

80.0/86.2

66.6/59.5

 

Random Forest

Mean accuracy

  

75.0

         

Recall score

90.9

55.5

          

Precision score

71.4

83.3

          

F1/F2 scores

80.0/86.2

66.6/59.5

          

Adaboost

Mean accuracy

  

75.0

     

85.0

   

Recall score

81.8

66.6

    

100.0

66.6

    

Precision score

75.0

75.0

    

78.0

100.0

    

F1/F2 scores

78.2/80.3

70.5/68.1

    

88.0/94.8

80.0/71.4

    

k-nearest neighbor

Mean accuracy

  

85.0

        

75.0

Recall score

90.9

77.7

       

90.9

55.5

 

Precision score

83.3

87.5

       

71.4

83.3

 

F1/F2 scores

86.9/89.2

82.3/79.5

       

80.0/86.2

66.6/59.5

 
  1. ASD autism spectrum disorder, TD typically developing, Schi schizophrenia, All whole brain or all features combined.