LLM in Personalized Health: Automating HEI Scoring

Area of Science:

• Computational nutrition
• Artificial intelligence in health
• Dietary assessment methodologies

Background:

• Automating dietary assessments is crucial for public health.
• The Healthy Eating Index (HEI) is a key metric for evaluating diet quality.
• Large Language Models (LLMs) offer potential for data processing and analysis.

Purpose of the Study:

• To investigate the efficacy of LLMs in automating HEI calculations.
• To assess the accuracy of LLM-generated HEI scores compared to traditional methods.
• To identify areas for improvement in LLM application for dietary analysis.

Main Methods:

• Utilized LLMs for processing dietary data to calculate HEI scores.
• Compared LLM-derived HEI scores against established calculation benchmarks.
• Evaluated the efficiency and accuracy of the automated approach.

Main Results:

• LLMs demonstrated high efficiency in processing dietary information for HEI calculation.
• LLM predictions consistently underestimated the actual HEI scores.
• The study identified a consistent bias in the automated HEI scoring.

Conclusions:

• LLMs show promise for streamlining dietary assessments and HEI calculations.
• Further research and model refinement, including fine-tuning and larger datasets, are necessary to enhance accuracy.
• Optimized LLMs could significantly support personalized nutrition and public health initiatives.