Skip to main content

Table 4 The cross-validation results using the feature set with the lowest error rate for the main dataset

From: Unraveling city-specific signature and identifying sample origin locations for the data from CAMDA MetaSUB challenge

True

AKL

BER

BOG

HAM

HGK

ILR

LON

MAR

NYC

OFA

PXO

SAC

SAO

SOF

STO

TOK

Predict

RF

 AKL

11

  

1

 

1

      

1

   

 BER

 

19

            

1

 

 BOG

  

14

       

1

     

 HAM

   

14

            

 HGK

    

15

           

 ILR

1

    

23

  

1

1

      

 LON

      

18

   

1

 

2

 

2

 

 MAR

       

9

        

 NYC

    

1

   

25

1

  

1

1

 

1

 OFA

         

18

      

 PXO

      

1

   

16

 

1

   

 SAC

 

1

         

18

    

 SAO

 

1

1

1

  

1

   

1

 

18

2

  

 SOF

             

7

  

 STO

1

     

1

1

    

1

 

16

1

 TOK

1

     

1

   

1

    

23

 Error rate (%)

21.43

9.52

6.67

12.50

6.25

4.17

18.18

10.00

3.85

10.00

20.00

0.00

25.00

30.00

15.79

8.00

SVM

 AKL

11

  

1

 

1

   

1

  

1

   

 BER

 

17

     

1

  

2

1

  

1

 

 BOG

  

14

             

 HAM

1

  

13

            

 HGK

    

16

           

 ILR

 

1

   

23

  

1

       

 LON

      

22

   

1

 

2

 

2

 

 MAR

       

8

        

 NYC

   

1

    

25

   

1

  

1

 OFA

         

19

   

1

  

 PXO

  

1

    

1

  

16

 

1

   

 SAC

           

17

    

 SAO

 

1

 

1

      

1

 

19

4

  

 SOF

 

1

           

5

  

 STO

1

      

1

    

1

 

16

 

 TOK

1

1

             

24

 Error rate (%)

21.43

19.05

6.67

18.75

0.00

4.17

0.00

20.00

3.85

5.00

20.00

5.56

20.83

50.00

15.79

4.00

LDA

 AKL

9

  

1

        

1

   

 BER

 

19

 

1

      

1

 

1

 

1

 

 BOG

  

13

             

 HAM

   

13

            

 HGK

  

1

 

16

    

1

      

 ILR

     

22

  

1

    

1

  

 LON

1

     

22

   

2

 

4

 

2

 

 MAR

       

9

1

       

 NYC

        

24

      

1

 OFA

     

2

   

18

      

 PXO

1

  

1

   

1

  

17

     

 SAC

1

          

17

    

 SAO

1

 

1

        

1

17

1

  

 SOF

         

1

  

1

8

  

 STO

              

16

1

 TOK

1

2

             

23

 Error rate (%)

35.71

9.52

13.33

18.75

0.00

8.33

0.00

10.00

7.69

10.00

15.00

5.56

29.17

20.00

15.79

8.00

Three methods

City

AKL

BER

BOG

HAM

HGK

ILR

LON

MAR

NYC

OFA

PXO

SAC

SAO

SOF

STO

TOK

 Average error rate (%)

26.19

12.69

8.89

16.67

2.08

5.56

6.06

13.33

5.13

8.33

18.33

3.70

25.00

33.33

15.79

6.67