A novel landslide susceptibility prediction framework based on contrastive loss

Recently, the positive unlabeled (PU) learning algorithms have proven highly effective in generating accurate landslide susceptibility maps. The algorithms categorize samples exclusively into positive samples (landslides) and unlabeled samples for training, eliminating random or subjective selection...

ver descrição completa

Detalhes bibliográficos

Principais autores:	Shubing Ouyang, Weitao Chen, Hangyuan Liu, Yuanyao Li, Zhanya Xu
Formato:	Artigo
Idioma:	Inglês
Publicado em:	Taylor & Francis Group 2024-12-01
coleção:	GIScience & Remote Sensing
Assuntos:	Landslide susceptibility prediction contrastive loss positive unlabeled learning remote sensing
Acesso em linha:	https://www.tandfonline.com/doi/10.1080/15481603.2024.2306740
Tags:	Adicionar Tag Sem tags, seja o primeiro a adicionar uma tag!