The Effect of Temporal Characteristics on Developing a Practical Rainfall-Induced Landslide Potential Evaluation Model Using Random Forest Method

With the unique rainfall patterns of typhoons, plum rains, and short-term heavy rainfalls, frequent landslide debris flow disasters have caused severe loss to people in Taiwan. In studies susceptibility, information factors used for analysis was usually annual-based content, it assumed that same elements from different years were independent between each year. However, occurrence landslides not simply due changes within a Instead, triggered because affected potential reached critical conditions after cumulative change with time. Therefore, this study had well evaluated influence temporal characteristics ratios antecedent areas past five evaluation model. The conducted through random forest (RF) algorithm. Additional events 2017 test proposed model’s performance understand its practicality. results show area, RF model considerably acceptable performance. also demonstrated essential describe significance time when conducting evaluation.