Skip to main content

Table 9 RF-Shapley regression in global set (Sorted by the size of share)

From: An innovative machine learning workflow to research China’s systemic financial crisis with SHAP value and Shapley regression

Feature

Sign

Share

(SSCvalue)

Coefficients

P value

Exponentiation

Change of \(\frac{{y}_{risk}}{1-{y}_{risk}}\)

CreditCostGrowth

 + 

0.1379***

0.0401

0.000

1.0409

Significantly increased by 4.09%

CPI-Dif

 + 

0.1365***

0.0373

0.000

1.0380

Significantly increased by 3.80%

YieldCurveSlope

 + 

0.1222***

0.0469

0.000

1.0480

Significantly increased by 4.80%

Credit-Dif

 + 

0.1151***

0.0547

0.000

1.0562

Significantly increased by 5.62%

GlobalNetExport

 − 

0.1013***

0.0323

0.000

1.0328

Significantly decreased by 3.28%

NarrowMoney

 − 

0.0970

-0.0455

0.000

0.9555

Insignificant increase

FinancialConcentration

 − 

0.0959

-0.0157

0.038

0.9844

Insignificant increase

Credit

 + 

0.0816***

0.0560

0.000

1.0576

Significantly increased by 5.76%

DidYield

 + 

0.0610

-0.0255

0.000

0.9748

Insignificant increase

NetExport

 − 

0.0515

-0.0393

0.000

0.9615

Insignificant increase

  1. ***, **, and * respectively means that the null hypothesis is rejected at the significance level of 1%, 5%, 10%