Computers and Industrial Engineering
| Hi-index | 0.00 |
In spatial interpolation domains, a popular method is Kriging. The contribution of this study is to prove mathematically that Kriging and RBF Networks methods produce identical results if properly configured, and that RBF networks are much faster. Complexity was calculated for both methods to show the relative speed of RBF networks. It is shown that both methods share a common structure and, as a consequence, all improvements in one method can be applied to the other. Finally, two experiments were made to show in practice the theoretical results obtained. The RBF networks were 200 times faster than Kriging in one particular experiment, and this difference increases as the data set gets larger.