Table 5 The average accuracies of 10-round identifications with three different kernel functions
From: Identifying the pulsed neuron networks’ structures by a nonlinear Granger causality method
2 nodes (%) | 3 nodes (%) | 4 nodes (%) | 5 nodes (%) | 6 nodes (%) | 20 nodes (%) | 100 nodes (%) | |
|---|---|---|---|---|---|---|---|
GF | 100 | 99.64 | 98.64 | 98.37 | 98.37 | 84.87 | 80.56 |
RSF | 100 | 99.63 | 98.34 | 98.23 | 98.31 | 84.85 | 80.49 |
IMQF | 100 | 99.67 | 98.76 | 98.53 | 98.47 | 85.11 | 80.88 |