:: Volume 8, Issue 2 (3-2021) ::
2021, 8(2): 13-19 Back to browse issues page
Classifying people based on fat by a Neuro-Fuzzy System
Mohammadreza Valizadeh , Ali Karamshahi , Kurosh Djafarian , Akbar Azizifar
Department of Computer and Information Technology, Ilam University, Ilam, Iran , valizadehmr@gmail.com
Abstract:   (1410 Views)
Introduction: Using BIA for body fat calculation is a normal method. The body fat factor is one of the most useful measures for assessing the risk of obesity. In this research, people are classified based on body fat. This research does not use any device. Adaptive Network-based Fuzzy Inference System (ANFIS) which is widely used in medical sciences, has been used to predict the exact category of fat.
Materials and Methods: A nutrition clinic in Tehran has collected 610 samples from its patients. Each data has six attributes: age, height, weight, BMI, gender, and fat percentage. Based on percentage fat, people are divided into six fat classes from very low fat to very high fat. This research uses ANFIS system to estimate body fat class. Age, height, weight, BMI, and gender are used as inputs of the system and fat class as output. Furthermore, for evaluating the proposed method, precision method is used.
Results: This research used machine learning techniques (i.e., ANFIS) to predict the class of fat people without using costly tools. The data showed that our method has an accuracy of 90.83%.
Conclusion: The results of this research show that using ANFIS can estimate accurately the category of body fat without any device. Therefore, it reduces diagnosis price.
Keywords: Learning algorithm, Body fat category, Data mining, ANFIS
Full-Text [PDF 632 kb]   (534 Downloads)    
Type of Study: Research | Subject: Biostatistics
Received: 2020/02/14 | Accepted: 2020/05/13 | Published: 2021/03/2


XML     Print



Rights and permissions
Creative Commons License This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Volume 8, Issue 2 (3-2021) Back to browse issues page