Table 4 Comparison of the results of different BP neural network training functions
From: A credit risk assessment model based on SVM for small and medium enterprises in supply chain finance
Function | Algorithm | Iteration number | MSE | Gradient |
|---|---|---|---|---|
traingdx | Variable learning rate momentum gradient descent algorithm | 6064 | 0.000100 | 0.00350 |
trainrp | RPROP (BP) algorithm, The minimum memory requirements | 496 | 0.000101 | 0.00182 |
traincgp | Conjugate gradient algorithm | 342 | 0.000099 | 0.00108 |
trainscg | Scaled conjugate gradient algorithm | 275 | 0.000099 | 0.00183 |