Prediction of Occurrence of Discrete Events - Geomathematics: Theoretical Foundations, Applications and Future Developments - page 173

Fig. 5.26 Diagrams
showing the training areas
(spatially coherent blocks)
used in ( a ), Experiment
3, ( b ) Experiment 4, and
( c ) Experiment 5 (Source:
Agterberg and Bonham-
Carter 2005 , Fig. 3)

small (100 m 100 m) cells. Only 139,643 (25 % of total area) of these were used

for training. The inset is an enlargement of part of the complete pattern of Fig. 5.25 .
In total, 23 of the 90 Au occurrences belong to the training set for Experiment
1. This experiment was repeated: the training set for Experiment 2 (not shown as a
figure) consists of 145,701 cells (26 % of total area) containing 15 Au occurrences.
Figure 5.26 shows the three geographically coherent subareas or 'blocks' used for
training in Experiments 3-5.