Abstract
The collapses of cut-off slope during heavy rain had been caused by slope evaluation factors that are topograpical, geological and earth-work related factors as well as rain-fall factors over a certain limit of value. Until now, there are many studies related to an evaluation of slope-collapse possibility based on only rain-fall factors. But, there are few studies based on the combinatoin of slope evaluation and rain-fall factors. In this research, we have constructed a system that can forecast the collapse or non-collapse of metamorphic rock-slope in real time during heavy rain using neural network based on slope evaluation together with rain-fall factors. Then, we present a rule for selecting efficient study data and the best combination of rain-fall factors to improve the reliability of the system.
