遇见数据集

Dataset_FOS_Supervised.xlsx

收藏
DataCite Commons2025-08-01 更新2025-09-08 收录
官方服务:

资源简介:

Using the security factor (FOS) calculated by TRIGRS as a physical label, design five sets of interactive features to enhance physical perception ability. Feature selection preserves the top 15 pieces of mutual information. The model is trained under weighted loss and pseudo labels, and the Stacking features are generated through 5-fold cross validation. The study compared the dependence and non dependence conditions of FOS and verified the applicability and interpretability of the method in complex geological environments, providing new support for landslide risk assessment.

提供机构:
figshare
创建时间:
2025-08-01
二维码
社区交流群
二维码
科研交流群
商业服务