Skip to main content

Table 3 Paired samples test of the LDMGNN model with 11 baseline models

From: Long-distance dependency combined multi-hop graph neural networks for protein–protein interactions prediction

     

95% confidence interval of the difference

   
  

Mean

SD

Std. error mean

Lower

Upper

t

df

Sig.(2-tailed)

Pair 1

LDMGNN-SVM

22.09500

7.58174

3.09523

14.13845

30.05155

7.138

5

0.001

Pair 2

LDMGNN-RF

29.37667

14.80078

6.04239

13.84420

44.90913

4.862

5

0.005

Pair 3

LDMGNN-LR

27.26500

5.42631

2.21528

21.57044

32.95956

12.308

5

0.000

Pair 4

LDMGNN-HIN2Vec

20.42333

5.50138

2.24593

14.64998

26.19668

9.093

5

0.000

Pair 5

LDMGNN-SDNE

15.59667

9.21296

3.76118

5.92826

25.26507

4.147

5

0.009

Pair 6

LDMGNN-LPIDLDN

19.15333

7.98559

3.26010

10.77297

27.53369

5.875

5

0.002

Pair 7

LDMGNN-LPIdeepGBDT

21.65833

7.89829

3.22446

13.36958

29.94708

6.717

5

0.001

Pair 8

LDMGNN-DTICDF

16.48333

6.47322

2.64268

9.69010

23.27656

6.237

5

0.001

Pair 9

LDMGNN-PIPR

15.13333

10.28307

4.19805

4.34191

25.92476

3.605

5

0.015

Pair 10

LDMGNN-GAT

12.71167

7.96402

3.25130

4.35394

21.06940

3.910

5

0.011

Pair 11

LDMGNN-GNNPPI

3.25167

3.85639

1.57436

-0.79537

7.29870

2.065

5

0.094

  1. Bold number represents the p-value of the corresponding paired sample