Using Explainable Artificial Intelligence to Discover Interactions in an Ecological Model for Obesity

Allen, Ben; Lane, Morgan; Steeves, Elizabeth Anderson; Raynor, Hollie A.

doi:10.3390/ijerph19159447

Cited by 4 publications

(2 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the current literature, we identified some studies using ML to predict obesity in North American [24][25][26][27][28], South Korean [29], and Turkish adolescents [30,30]. Although, none these authors considered direct physical fitness levels as potential deterministic variables for their analysis.…”

Section: Introductionmentioning

confidence: 99%

A Deep Learning Neural Network to Classify Obesity Risk in Portuguese Adolescents Based on Physical Fitness Levels and Body Mass Index Percentiles: Insights for National Health Policies

Forte,

Encarnação,

Monteiro

et al. 2023

Behavioral Sciences

View full text Add to dashboard Cite

The increasing prevalence of overweight and obesity among adults is a risk factor for many chronic diseases and death. In addition, obesity among children and adolescents has reached unprecedented levels and studies show that obese children and adolescents are more likely to become obese adults. Therefore, both the prevention and treatment of obesity in adolescents are critical. This study aimed to develop an artificial intelligence (AI) neural network (NNET) model that identifies the risk of obesity in Portuguese adolescents based on their body mass index (BMI) percentiles and levels of physical fitness. Using datasets from the FITescola® project, 654 adolescents aged between 10–19 years old, male: 334 (51%), female: n = 320 (49%), age 13.8 ± 2 years old, were selected to participate in a cross-sectional observational study. Physical fitness variables, age, and sex were used to identify the risk of obesity. The NNET had good accuracy (75%) and performance validation through the Receiver Operating Characteristic using the Area Under the Curve (ROC AUC = 64%) in identifying the risk of obesity in Portuguese adolescents based on the BMI percentiles. Correlations of moderate effect size were perceived for aerobic fitness (AF), upper limbs strength (ULS), and sprint time (ST), showing that some physical fitness variables contributed to the obesity risk of the adolescents. Our NNET presented a good accuracy (75%) and was validated with the K-Folds Cross-Validation (K-Folds CV) with good accuracy (71%) and ROC AUC (66%). According to the NNET, there was an increased risk of obesity linked to low physical fitness in Portuguese teenagers.

show abstract

Section: Introductionmentioning

confidence: 99%

A Deep Learning Neural Network to Classify Obesity Risk in Portuguese Adolescents Based on Physical Fitness Levels and Body Mass Index Percentiles: Insights for National Health Policies

Forte,

Encarnação,

Monteiro

et al. 2023

Behavioral Sciences

View full text Add to dashboard Cite

show abstract

“…Machine learning may be the most powerful approach to modeling variation in obesity prevalence across the United States, but machine learning models are often opaque and difficult to interpret [8]. To open the black box of machine learning models, the field of explainable artificial intelligence has emerged with the goal of extracting domain knowledge about the outcomes being predicted [4,[9][10][11][12]. This paper shows an application of explainable artificial intelligence methods to machine learning models of geographic variation in obesity prevalence.…”

Section: Introductionmentioning

confidence: 99%

An interpretable machine learning model of cross-sectional U.S. county-level obesity prevalence using explainable artificial intelligence

Allen

2023

PLoS ONE

Self Cite

View full text Add to dashboard Cite

Background There is considerable geographic heterogeneity in obesity prevalence across counties in the United States. Machine learning algorithms accurately predict geographic variation in obesity prevalence, but the models are often uninterpretable and viewed as a black-box. Objective The goal of this study is to extract knowledge from machine learning models for county-level variation in obesity prevalence. Methods This study shows the application of explainable artificial intelligence methods to machine learning models of cross-sectional obesity prevalence data collected from 3,142 counties in the United States. County-level features from 7 broad categories: health outcomes, health behaviors, clinical care, social and economic factors, physical environment, demographics, and severe housing conditions. Explainable methods applied to random forest prediction models include feature importance, accumulated local effects, global surrogate decision tree, and local interpretable model-agnostic explanations. Results The results show that machine learning models explained 79% of the variance in obesity prevalence, with physical inactivity, diabetes, and smoking prevalence being the most important factors in predicting obesity prevalence. Conclusions Interpretable machine learning models of health behaviors and outcomes provide substantial insight into obesity prevalence variation across counties in the United States.

show abstract

Effects of high-intensity and progressive volume resistance training on functional, mental states, and quality of life of people with spinal cord injury

Rodrigues,

Pereira,

Lopes

et al. 2024

Sport Sci Health

View full text Add to dashboard Cite

show abstract

Using Explainable Artificial Intelligence to Discover Interactions in an Ecological Model for Obesity

Cited by 4 publications

References 49 publications

A Deep Learning Neural Network to Classify Obesity Risk in Portuguese Adolescents Based on Physical Fitness Levels and Body Mass Index Percentiles: Insights for National Health Policies

A Deep Learning Neural Network to Classify Obesity Risk in Portuguese Adolescents Based on Physical Fitness Levels and Body Mass Index Percentiles: Insights for National Health Policies

An interpretable machine learning model of cross-sectional U.S. county-level obesity prevalence using explainable artificial intelligence

Effects of high-intensity and progressive volume resistance training on functional, mental states, and quality of life of people with spinal cord injury

Contact Info

Product

Resources

About