A State-Based Sensitivity Analysis for Distinguishing the Global Importance of Predictor Variables in Artificial Neural Networks

Table 1

Hypothetical database for an artificial neural network. Note that data are not normalized, with vectors sorted based on descending values of .

−1.80	−0.41	−0.71	1.01	0.36	−2.03
−1.77	−0.78	0.76	−1.79	0.02	−1.14
−1.53	−0.64	0.67	−0.24	0.69	−1.43
−1.50	0.62	1.36	−0.02	1.35	−0.99
−0.71	−1.90	−2.05	−1.23	−0.22	−1.66
−0.70	−2.37	−1.19	−1.08	−0.58	−1.78
−0.50	−0.81	0.56	−0.64	−1.42	−0.10
−0.38	−0.04	−0.24	−0.40	−1.54	−0.18
−0.31	0.46	−0.36	1.14	0.84	−0.51
−0.16	0.84	−0.58	−0.06	0.60	−0.56
−0.13	−0.36	−0.21	−0.35	1.19	−0.36