Klasifikasi Status Stunting Pada Balita Menggunakan K-Nearest Neighbor Dengan Feature Selection Backward Elimination

Abstract

The main problem regarding nutrition faced by Indonesia is stunting, where Indonesia is ranked fifth in the world with the largest stunting prevalence rate in 2017, which is 29.6% of all Indonesian children. Stunting is a child under five years who has a z-score value of less than -3 standard deviations (SD). Stunting has a negative impact, namely it can disrupt the physical and intellectual development of toddlers in the future. In this case, the examination of stunting status by medical personnel is still carried out manually which takes a long time and is prone to inaccuracies. This study aims to classify stunting status in toddlers by applying the K-Nearest Neighbor method using the Backward Elimination feature selection to get fast and accurate results. Based on the results of this study, the average accuracy produced by the K-Nearest Neighbor algorithm at k=5 is 91.90% with 9 attributes and the average accuracy produced by the K-Nearest Neighbor algorithm with the addition of Backward Elimination is 92.20%. with 8 attributes. These results indicate that the application of Backward Elimination can increase the accuracy value of the K-Nearest Neighbor algorithm and also perform attribute selection.