Clustering analysis and machine learning algorithms in the prediction of dietary patterns: Cross‐sectional results of the Brazilian Longitudinal Study of Adult Health (ELSA‐Brasil)

Abstract

Background Machine learning investigates how computers can automatically learn. The present study aimed to predict dietary patterns and compare algorithm performance in making predictions of dietary patterns. Methods We analysed the data of public employees (n = 12,667) participating in the Brazilian Longitudinal Study of Adult Health (ELSA‐Brasil). The K‐means clustering algorithm and six other classifiers (support vector machines, naïve Bayes, K‐nearest neighbours, decision tree, random forest and xgboost) were used to predict the dietary patterns. Results K‐means clustering identified two dietary patterns. Cluster 1, labelled the Western pattern, was characterised by a higher energy intake and consumption of refined cereals, beans and other legumes, tubers, pasta, processed and red meats, high‐fat milk and dairy products, and sugary beverages; Cluster 2, labelled the Prudent pattern, was characterised by higher intakes of fruit, vegetables, whole cereals, white meats, and milk and reduced‐fat milk derivatives. The most important predictors were age, sex, per capita income, education level and physical activity. The accuracy of the models varied from moderate to good (69%–72%). Conclusions The performance of the algorithms in dietary pattern prediction was similar, and the models presented may provide support in screener tasks and guide health professionals in the analysis of dietary data.

Crash Risk Predictors in Older Drivers: A Cross-Sectional Study Based on a Driving Simulator and Machine Learning Algorithms

“…The group's interpretability was examined to confirm the final number of clusters and if a group was sufficiently large for adequate statistical power, that is, at least 10% of the total sample. The clustering distance measurements were carried out using Euclidean distances [14].…”

“…Machine Learning (ML) is an expanding field because of its high predictive power. Algorithms based on ML can help in tasks that involve data analysis and identify the characteristics of elderly drivers associated with performance when driving a vehicle [14]. We tested the hypothesis that driver's clusters can be derived by an unsupervised learning algorithm (K-Means) and that the number of road crashes can be predicted by a supervised learning algorithm (random forest) from a data set of motor, visual, and cognitive tests.…”

“…The correlation between predicted and observed crashes was 0.98 and R 2 was 0.81. Values ≤ 0.5 corresponded to the performance of a random model, values > 0.5 and < 0.6 indicate moderate predictive performance, and values > than 0.7 indicate good predictive performance[14].…”

A changing of the guard: reflecting on the past and looking to the future at JHND

“…We look to critically appraise submissions in a manner that supports learning and maintains scientific excellence, but at the same time embracing research which uses novel and innovative research designs and methodology. 13,14 Together, perhaps we will attract new authors who had previously targeted other nutrition journals, further securing JHND's reputation as a highquality publication in its own right. JHND will be a flagship example of Wiley's mission of unlocking human potential through education.…”

“…Our team of editors will continue to operate with a vibrant, positive culture, while also interacting with nutrition researchers from across the world. We look to critically appraise submissions in a manner that supports learning and maintains scientific excellence, but at the same time embracing research which uses novel and innovative research designs and methodology 13,14 . Together, perhaps we will attract new authors who had previously targeted other nutrition journals, further securing JHND's reputation as a high‐quality publication in its own right.…”