Women’s Weight Gain Analysis Using the Neural Network Method in Medan, Indonesia

Abstract

BACKGROUND: Obesity has become a global problem and has even been declared a global epidemic by the WHO. The high percentage of non-infectious diseases in Indonesia is 69.911%. This is experienced by people aged over 18 years. Central obesity is experienced by 26.6% of Indonesia's population (44.3 million people). Non-infectious diseases are the biggest cause of death and disability in Indonesia, 80% of non-infectious diseases are caused by an unhealthy lifestyle. The impact caused by this case can affect various aspects in the health, economic, socio-cultural, and psychological fields of the sufferer. Therefore, it is very important to prevent and control it. Many health promotion efforts have been carried out to overcome them, from conventional to modern health promotion activities that are considered not optimal to overcome them. The use and utilization of technology is one of the best solutions for solving public service problems. At least its utilization can overcome various geographical, time, and socio-economic problems.
 AIM: Assessment is required to determine the primary cause of weight gain.
 METHODOLOGY: This type of research is a survey with an explanatory type, to analyze the causal relationship between research variables and body mass index. It was conducted in the city of Medan in 21 (twenty-one) districts. Through sampling with two-stage cluster sampling, as many as 210 women aged 35–50 years were included in the research sample.
 RESULTS: The results of the Random Forest Algorithm calculation test and the Neural Network method with MLP (Multilayer Perceptron) showed that the history of being overweight, contraceptive use, and diet were the dominant factors influencing Body Mass Index.
 CONCLUSION: The history of weight gain with age is the dominant factor influencing changes in body mass index.