iDT-diet: Toward Personalized Health Forecasting-An Intelligent Digital Twin Model for Diet-Influenced Biomarker...

Area of Science:

• Computational health science
• Biomedical informatics
• Personalized nutrition

Background:

• Diet quality significantly impacts health biomarkers and chronic disease development.
• Modeling long-term dietary effects requires advanced analytical approaches.
• Current methods often lack personalized, dynamic visualization and interpretation.

Purpose of the Study:

• To introduce iDT-diet, an intelligent digital twin prototype.
• To model the long-term influence of diet quality on health.
• To integrate machine learning, natural language, and 3D visualization for health trajectory analysis.

Main Methods:

• Utilized a random forest model with Choquet LASSO feature selection for temporal health data.
• Developed a translation module for natural language output of health states.
• Integrated a generative 3D engine for dynamic, personalized digital twin visualization.

Main Results:

• Successfully created a prototype linking machine learning, interpretable communication, and visualization.
• Demonstrated the capability for retrospective digital twin generation.
• Established an architecture supporting real-time data integration and simulation.

Conclusions:

• The iDT-diet framework offers a novel approach to understanding diet's long-term health impact.
• The system facilitates personalized health monitoring and predictive simulation.
• Future applications include real-time diet and lifestyle management recommendations.