Research Article

Predicting the Robustness of Large Real-World Social Networks Using a Machine Learning Model

Table 4

Fit coefficients and the evaluation result given by the MLR. RIB or RRD columns show the slope coefficient, and in parenthesis, the standard error values for the NSI. The bold character with an asterisk indicates a significant relationship between the NSI and the robustness R, with a confidence level of 95%.

NbNSIRegression coefficients
R IBR RD

1α0.113 (0.028)0.088 (0.026)
2α2−0.118 (0.028)−0.083 (0.026)
3α−0.029 (0.013)−0.025 (0.012)
4Q0.047 (0.02)0.038 (0.019)
5C−0.034 (0.016)−0.016 (0.015)
6Cl0.005 (0.014)0.007 (0.013)
7N−0.013 (0.013)−0.014 (0.012)
8<k>0.077 (0.017)0.056 (0.016)
9Intercept0.173 (0.01)0.148 (0.009)

MLR results
RMSE0.07090.0657
R258.04%51.76%