Table 15 The outcomes of t-test for GCN/GAT-models.

From: A comprehensive analysis of digital inclusive finance’s influence on high quality enterprise development through fixed effects and deep learning frameworks

Models

T-statistic

P-value

Significant (p < 0.05)

Mean difference

Cohen’s d

Improvement (%)

GCN-BiLSTM

Original-BiLSTM

5.3124

0.0060

True

0.0645

4.4786

29.8049

GAT-BiLSTM

Original-BiLSTM

4.9100

0.0080

True

0.1668

3.8254

77.0824

GAT-BiLSTM

GCN-BiLSTM

4.5961

0.0101

True

0.1023

2.4095

67.3515

GCN-LSTM

Original-LSTM

10.5228

0.0005

True

0.0627

6.4280

28.5884

GAT-LSTM

Original-LSTM

10.6047

0.0004

True

0.0690

7.1069

31.4901

GAT-LSTM

GCN-LSTM

7.1438

0.0020

True

0.0064

2.3017

4.0634

GCN-GRU

Original-GRU

7.2396

0.0019

True

0.0625

5.2862

28.8072

GAT-GRU

Original-GRU

10.6855

0.0004

True

0.0689

7.1554

31.7386

GAT-GRU

GCN-GRU

1.5485

0.1964

False

0.0064

0.8617

4.1176

GCN-Transformer

Original-Transformer

18.7487

0.0000

True

0.1452

11.2909

77.8923

GAT-Transformer

Original-Transformer

14.0360

0.0001

True

0.1508

10.6058

80.9401

GAT-Transformer

GCN-Transformer

1.6396

0.1764

False

0.0057

0.7886

13.7864