Stunting Early Warning Application Using KNN Machine Learning Method

Abstract

Stunting in toddlers is defined as a condition of failure to thrive due to chronic malnutrition in the long term. The problem of stunting in Indonesia is an issue that is still a concern for the Indonesian government. The prevalence of stunting in Indonesia is still quite high, coupled with the COVID-19 pandemic which has had quite an impact on the economic sector. For this reason, research on stunting is still a very important topic. This study aims to classify toddler stunting using the k-Nearest Neighbor classification algorithm, as well as build a website-based early detection application for stunting toddler cases using the CodeIgniter framework with the PHP programming language. The results of the research using the k-Nearest Neighbor Algorithm trial obtained a fairly high accuracy of 92.45%. The implementation of an early detection system for stunting cases has proven to help and facilitate health workers in classifying toddlers as stunted or not. This application is also useful as an archive and facilitates data reporting. In the application there are 8 main menus, namely the Puskesmas data menu, Posyandu data, toddler data, weighing, weighing results, development menu, stunting early warning menu which contains malnourished toddlers, stunted toddlers.

Keywords

• early warning application
• k-nn method
• classification
• stunting