Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
January 20, 2026International Journal of Coal Science & TechnologyOpen Access

Performance evaluation and interpretation of hybrid models based on light gradient boosting machine to predict coal mining subsidence

View Full Paper
Ask AI
Bookmark
Share

Authors

XCXuan CuiSYShengli YangJWJiachen Wang

Discussion

Loading...

Member takes

Overview

Evaluates hybrid prediction models for coal mining subsidence in mining areas, highlighting effective safety measures.

Key Points

  • The aim is to develop and evaluate a hybrid model for predicting coal mining subsidence and to interpret the factors affecting it.
  • Developed a hybrid prediction model using light gradient boosting machine (LightGBM).
  • Collected a dataset of 163 mining subsidence cases with 12 features.
  • Evaluated model performance using multiple indicators like R^2.
  • Applied Shapley Additive Explanations (SHAP) for feature contribution analysis.
  • Conducted targeted interaction analysis on key parameters affecting mining subsidence.
  • HGS-LightGBM achieved an R^2 of 0.9589, outperforming single LightGBM which had an R^2 of 0.925.
  • The thickness of the coal seam was identified as the most influential factor for coal mining subsidence.
  • The hybrid model demonstrated significant interpretability and transparency in its predictions.

Cite This Study

Cui et al. (2026) studied this question.

synapsesocial.com/papers/696f1a239e64f732b51ee6b1https://doi.org/10.1007/s40789-025-00843-9
View Full Paper
Ask AI
Bookmark
Share