2001 Volume 54 Issue 1 Pages 21-29
This paper investigates the difference of topographical features between the mountain collapse of a shallow depth due to earthquake and that due to torrential rainfall by using neural network. Neural network learns through collapse data containing uncertainty obtained from the different pubulished papers. Results obtained by using neural network after learning are as follows
1) Collapse due to earthquake easily occurs at the mountain slope of the convex type in cross section and site on the summit.
2) Collapse due to rainfall easily occurs at the mountain foot and the slope with concave topographical feature.
3) Degree of certainty of the first conclusion is higher than the second.