Frontiers in global women's health · 2026
Predicting mood swings in women of reproductive age using machine learning on metabolic, menstrual, and lifestyle indicators.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
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.
AlSaad Rawan, El Rayess Farah, Thomas Rajat
Аннотация
BACKGROUND: Mood swings in reproductive-age women arise from interacting hormonal, metabolic, and lifestyle factors, yet scalable screening tools remain limited. Artificial intelligence (AI) and machine learning (ML) approaches offer the potential to integrate diverse predictors and enable early, data-driven risk stratification.
OBJECTIVE: To evaluate the performance of ML algorithms in predicting mood swings among reproductive-age women using menstrual, metabolic, and lifestyle survey data and to identify the most influential predictors.
METHODS: The study cohort included 465 reproductive-age women, with fifteen survey-derived features categorized into metabolic (e.g., BMI, recent weight gain, polycystic ovary syndrome), menstrual (regular periods, period length), lifestyle (fast-food consumption, daily exercise), symptom burden score, and demographic (age) categories. We compared five ML models: Random Forest, SVM, Gradient Boosting, LightGBM, and CatBoost, using precision, recall, F1, accuracy, and AUCPR metrics. Feature importance was assessed with permutation feature importance (PFI) and shapley additive explanations (SHAP).
RESULTS: Across models, the highest values achieved were precision 0.83, recall 0.91, accuracy 0.74, and AUCPR 0.87. PFI and SHAP converged on symptom burden as the dominant predictor, with additional signal from lifestyle indicators (higher fast-food consumption, lower daily exercise) and metabolic/dermatologic markers. Menstrual regularity/length contributed minimally; age showed a modest inverse association.
CONCLUSIONS: Low-cost, self-reported features can support ML prediction of mood swings in reproductive-age women with good performance. Findings motivate prospective validation, dynamic prediction with wearables, and evaluation of AI-based approaches for early detection of women's mental health concerns in community and primary care settings.
OBJECTIVE: To evaluate the performance of ML algorithms in predicting mood swings among reproductive-age women using menstrual, metabolic, and lifestyle survey data and to identify the most influential predictors.
METHODS: The study cohort included 465 reproductive-age women, with fifteen survey-derived features categorized into metabolic (e.g., BMI, recent weight gain, polycystic ovary syndrome), menstrual (regular periods, period length), lifestyle (fast-food consumption, daily exercise), symptom burden score, and demographic (age) categories. We compared five ML models: Random Forest, SVM, Gradient Boosting, LightGBM, and CatBoost, using precision, recall, F1, accuracy, and AUCPR metrics. Feature importance was assessed with permutation feature importance (PFI) and shapley additive explanations (SHAP).
RESULTS: Across models, the highest values achieved were precision 0.83, recall 0.91, accuracy 0.74, and AUCPR 0.87. PFI and SHAP converged on symptom burden as the dominant predictor, with additional signal from lifestyle indicators (higher fast-food consumption, lower daily exercise) and metabolic/dermatologic markers. Menstrual regularity/length contributed minimally; age showed a modest inverse association.
CONCLUSIONS: Low-cost, self-reported features can support ML prediction of mood swings in reproductive-age women with good performance. Findings motivate prospective validation, dynamic prediction with wearables, and evaluation of AI-based approaches for early detection of women's mental health concerns in community and primary care settings.