Precision nutrition: A systematic literature review

Area of Science:

• Nutritional Science
• Computer Science
• Artificial Intelligence

Background:

• Precision Nutrition aims to personalize dietary advice using individual data.
• Machine learning (ML) offers potential for developing predictive models in Precision Nutrition.
• A comprehensive overview of ML applications in Precision Nutrition is currently lacking.

Purpose of the Study:

• To systematically review the current state-of-the-art of ML in Precision Nutrition.
• To identify how and where ML is applied, input features, data availability, and model evaluation methods.
• To answer nine defined research questions regarding ML in Precision Nutrition.

Main Methods:

• A Systematic Literature Review (SLR) was conducted.
• 4930 papers were retrieved from electronic databases.
• 60 primary studies were selected and analyzed.

Main Results:

• Fifteen distinct problems were identified across seven nutrition and health domains.
• Common ML tasks include regression, classification, recommendation, and clustering, predominantly supervised.
• Thirty ML algorithms were utilized, with 19 appearing multiple times; models were evaluated using 23 metrics across four approach groups.

Conclusions:

• ML integration into Precision Nutrition research is feasible and demonstrates high performance.
• Personalized approaches driven by ML can effectively address current challenges in nutrition.
• Researchers are encouraged to incorporate ML to develop advanced Precision Nutrition strategies.