PREDICTING A STUNTING PREVALENCE USING SEMI-SUPERVISED LEARNING MODELS IN EAST NUSA TENGGARA

. Stunting is a complicated problem to solve. The impact of stunting is on the long-term development of children where they may never reach their full high potential and have poor cognitive development which leads to less than optimal educational performance and decreased intellectual capacity, motor, and socioeconomic development. In the case of Indonesia, WHO includes Indonesia in countries with a high risk of stunting (30%-39%). This study aims to predict the risk of stunting using a semi-supervised learning model. However, it is necessary to explore the dominant determinants of stunting first. Unsupervised learning is superior for finding attributes that have a high correlation, while supervised learning is used to map attributes that correlate with stunting risk targets. Stunting data is recorded in each community health center for each district. There are some public health centers names such as "KAPAN", "PANTJE", "KUANFATU", "OEEKAM", and "OIMLASI". We found that the community health centers "KAPAN" and "PANTJE" form a stunting prevalence cluster with the highest values at 50% and above. "KUANFATU", "OEEKAM", and "OIMLASI" followed at 30% to 50%.