Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
July 26, 2026Remote SensingOpen Access

Task-Specific Negative Sample Selection for Multi-Hazard Susceptibility Mapping of Slope-Instability Hazards

View Full Paper
Ask AI
Bookmark
Share

Authors

XLXiaohe LaiGZGuoye ZhaoJJJun Jiang

Discussion

Loading...

Member takes

Overview

Randomized trial evaluates a new negative sample selection method in multi-hazard mapping, highlighting improved classification outcomes.

Key Points

  • This study aims to enhance multi-hazard susceptibility mapping by addressing label confusion and sample mismatch in negative sampling.
  • Developed a task-specific negative sample selection method based on hazard-specific positive samples.
  • Compared the proposed method against five benchmark sampling strategies using multiple machine learning models.
  • Conducted analyses in the Minjiang River Basin in southeastern coastal China.
  • Average classification performance improved by 14.47% in AUC and 12.25% in F1-score.
  • Top 10% susceptible areas captured 39.26% of historical hazard events.
  • Only 27.06% of slope units covered 80% of hazards.

Cite This Study

Lai et al. (2026) studied this question.

synapsesocial.com/papers/6a65a660d3aea3239cd77cfdhttps://doi.org/10.3390/rs18152437
View Full Paper
Ask AI
Bookmark
Share