Using neural networks to predict weight loss

Yazdan Bakhsh, Anooshe; Poor Ebrahimi, Alireza; Norouzi, Abdolreza

doi:JABM.3.2.15564.3556.3274

Using neural networks to predict weight loss

Document Type : Excerpt from doctoral thesis

Authors

Anooshe Yazdan Bakhsh ¹

Alireza Poor Ebrahimi ²

Abdolreza Norouzi ³

¹ PhD in Information Technology Management, Smart Business Branch, Department of Management, Faculty of Management, Islamic Azad University, Kish Branch, Iran.

² Assistant Professor, Department of Industrial Management, Faculty of Management, Islamic Azad University, Karaj Branch, Iran.

³ Associate Professor, Department of Nutrition Sciences, Mashhad University of Medical Sciences, Iran.

JABM.3.2.15564.3556.3274

Abstract

Obesity is a major public health challenge that can lead to heart disease, type 2 diabetes, and high blood pressure. Predicting weight loss plays an important role in obesity management and optimizing diet programs. In this study, three deep learning models, including recurrent neural networks (RNN), convolutional neural networks (CNN), and a combination of these models with an attention mechanism, were investigated. The results showed that the CNN model performed better than the RNN alone, and adding an attention mechanism to the RNN did not significantly improve its performance. However, adding an attention mechanism to the CNN increased the accuracy of the model and had the best performance among the models studied. These findings confirm the importance of choosing the appropriate model and utilizing the attention mechanism in improving the accuracy of weight loss prediction.

Keywords

Recurrent neural network

convolutional neural network

attention mechanism

weight loss prediction

Subjects