在崩塌風險模式的建構上，是以二元Logistic迴歸為基礎，輸出是以單

在崩塌風險模式的建構上，是以二元Logistic迴歸為基礎，輸出是以單變量(崩塌機率)為主。另提出了使用類神經網路模式，該模式可以提供多變量(崩塌機率及災害機率)輸出。因此，使用影響變量及記憶變量來建構崩塌風險模式，該模式可以計算因條件改變而導致之崩塌發生(出現)機率及災害發生(出現)機率，以期作為未來各項土地利用開發、土地管理、災害預防及土砂治理推動等管理策略擬定之參考依據。

Results (English) 1:

Building collapse risk model, is based on a binary Logistic regression, the output is a single variable (avalanche probability). Another proposes using a neural network model, this model provides a multivariable (avalanche probability and probability disasters) output. Therefore, using variable and memory variable to building collapse risk model, which can occur due to conditions as a result of the collapse (appearance) and disaster risk (a) probability, with a view to future development of land use, land management and disaster prevention management strategies such as sand and soil management to promote the development of the reference.

Results (English) 2:

On the construction of collapse risk model, based on the basis of binary Logistic regression output is a single variable (avalanche probability) based. Another proposed the use of neural network model that can provide multiple variables (probability of collapse and disaster probability) output. Therefore, the impact variables and memory variables to construct collapsed risk model that can calculate the change due to the conditions which led to the collapse occurred (appears) and the probability of a disaster (present) probability, with a view to the future development of the land use, land management, earth and sand disaster prevention and control and other management strategies to promote the preparation of reference.

Results (English) 3:

Based on the construction of collapse risk model, the output is based on the two Logistic regression, and the output is based on the single variable (probability of collapse). The model can provide multiple variables (collapse probability and hazard probability). Therefore, the model can be used to construct the risk model.