Portal Vein Thrombosis (PVT) Risk Prediction System

Intelligent Prediction Model Based on Stacking Ensemble Learning

Patient Information Entry

Prediction Results

Model predicted PVT risk is
---%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
← Low Risk Medium Risk High Risk →
--
Model Information
  • Model Type: Stacking Ensemble Learning
  • Base Models: RF, XGBoost, GBM, AdaBoost, CatBoost
  • Meta Models: MLP, LR, LightGBM
  • Feature Count: 13 clinical features