The Journal of nutrition · 2026
Feature Selection and Intelligent Classification of Dietary Nature Based on Tibetan Medicine Pharmacological Theory and Nutritional Components.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Caidan Duojie, Cuomao Sanzhi, Cairang Nima, Rangzhong Qieyang, Duojie Geri, Jia Renzeng, Dongzhi Luosang
Аннотация
BACKGROUND: The cold, hot, and neutral natures are core concepts in Tibetan dietary therapy and guide traditional food classification and dietary use. However, their identification has largely depended on empirical judgment, and quantitative standards based on modern nutritional features remain limited.
OBJECTIVES: This study aimed to develop an interpretable machine learning framework integrating feature selection, multimodel classification, and SHapley Additive exPlanations (SHAP) analysis to identify nutritional drivers of Tibetan dietary nature and support prediction for currently unlabeled food items.
METHODS: A structured database of 786 Tibetan dietary items was constructed by integrating traditional Tibetan medical attributes with modern nutritional data. Key features were selected using least absolute shrinkage and selection operator regression and recursive feature elimination with cross-validation. Nine machine learning algorithms were compared, and SHAP analysis was used to interpret nonlinear feature contributions to the cold, hot, and neutral classifications.
RESULTS: Ten core features were identified, including trace elements, vitamins, macronutrients, and sensory attributes, represented by selenium, manganese, niacin, protein, and bitter taste. eXtreme Gradient Boosting achieved the best overall performance, with an area under the receiver operating characteristic curve of 0.843 and a Brier score of 0.145. SHAP analysis showed that selenium and niacin were dominant predictors with nonlinear interactions. Cold foods were characterized by a bitter-taste-related, low nutrient-density pattern; hot foods showed a high-energy-driven pattern associated with niacin, protein, and sweet taste; and neutral foods reflected a dynamic balance of multiple driving factors.
CONCLUSIONS: This study established an interpretable intelligent classification system for the 3 natures of Tibetan dietary therapy and proposed a selenium-niacin-metabolism axis hypothesis. These findings provide quantitative evidence for the Tibetan dietary theory and support standardized dietary classification and individualized dietary guidance.
OBJECTIVES: This study aimed to develop an interpretable machine learning framework integrating feature selection, multimodel classification, and SHapley Additive exPlanations (SHAP) analysis to identify nutritional drivers of Tibetan dietary nature and support prediction for currently unlabeled food items.
METHODS: A structured database of 786 Tibetan dietary items was constructed by integrating traditional Tibetan medical attributes with modern nutritional data. Key features were selected using least absolute shrinkage and selection operator regression and recursive feature elimination with cross-validation. Nine machine learning algorithms were compared, and SHAP analysis was used to interpret nonlinear feature contributions to the cold, hot, and neutral classifications.
RESULTS: Ten core features were identified, including trace elements, vitamins, macronutrients, and sensory attributes, represented by selenium, manganese, niacin, protein, and bitter taste. eXtreme Gradient Boosting achieved the best overall performance, with an area under the receiver operating characteristic curve of 0.843 and a Brier score of 0.145. SHAP analysis showed that selenium and niacin were dominant predictors with nonlinear interactions. Cold foods were characterized by a bitter-taste-related, low nutrient-density pattern; hot foods showed a high-energy-driven pattern associated with niacin, protein, and sweet taste; and neutral foods reflected a dynamic balance of multiple driving factors.
CONCLUSIONS: This study established an interpretable intelligent classification system for the 3 natures of Tibetan dietary therapy and proposed a selenium-niacin-metabolism axis hypothesis. These findings provide quantitative evidence for the Tibetan dietary theory and support standardized dietary classification and individualized dietary guidance.