An insurance product pricing method and system based on clinical prognosis digital twinning
By using Transformer neural networks and reinforcement learning agents to process clinical time-series data, the problem of insufficient data processing and adaptive capabilities in insurance pricing has been solved, enabling real-time, refined assessment of individual risk and stable pricing decisions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- RENJI HOSPITAL AFFILIATED TO SHANGHAI JIAO TONG UNIV SCHOOL OF MEDICINE
- Filing Date
- 2026-06-15
- Publication Date
- 2026-07-17
AI Technical Summary
Existing insurance pricing technologies cannot effectively handle non-linear time-series clinical data, lack adaptive capabilities, and are unstable in decision-making when faced with missing or abnormal clinical data, resulting in delayed risk assessment and inaccurate pricing.
Employing a Transformer neural network architecture and reinforcement learning agent, this system processes clinical time-series data through a multi-head attention mechanism to generate risk prognosis survival function trajectories. Furthermore, it introduces a confidence gating mechanism based on prediction variance quantification to achieve dynamic pricing and abnormal circuit breaking.
It enables real-time and refined characterization of individual health risks, enhances the system's adaptability and decision-making efficiency, ensures computational stability and compliance, and improves the accuracy and reliability of pricing.
Smart Images
Figure CN122415236A_ABST