Klasifikasi Risiko Stunting pada Balita Berdasarkan Data Kesehatan Menggunakan Algoritma Naïve Bayes

Abstract

Stunting remains a major public health concern in Indonesia due to its impact on children's physical growth and cognitive development. Early detection of stunting risk is essential to support targeted prevention. This study aims to implement the Naïve Bayes algorithm to classify stunting risk in children under five based on four basic anthropometric attributes: age, gender, weight, and height, using secondary data from 300 toddlers obtained from Labuhan Ruku Public Health Center, Batu Bara Regency. The stunting category label was computed using the Height-for-Age Z-score (HAZ) index based on the WHO Child Growth Standards 2006. The research stages included data collection, preprocessing (missing value and duplicate checking), discretization of numerical attributes, label encoding, an 80:20 train-test split, Naïve Bayes model development with Laplace smoothing, and evaluation using a confusion matrix and 10-fold cross validation. Testing on 60 test samples produced a confusion matrix with TP=27, TN=28, FP=5, and FN=0, yielding an accuracy of 91.67%, precision of 92.97%, recall of 91.67%, and F1-score of 91.68%. The 10-fold cross validation produced an average accuracy of 92.67% with a standard deviation of 7.42%. All misclassifications were false positives, meaning no stunted children were missed. These results indicate that the Naïve Bayes algorithm is effective and stable for classifying stunting risk based on anthropometric data and has strong potential to support early detection by healthcare workers.