Application Of The Random Forest Algorithm For Classification Of Stunting Factors In Toddlers

Abstrak

Stunting is a type of growth failurecaused by chronic malnutrition, a condition that causes children toexperience stunted growth at a certain age. The high rate of stuntinghighlights the need for intensive and innovative efforts to address it,including through data-driven and technology-based approaches. This studyaims to determine the accuracy of the Random Forest algorithm to be implemented for stunting risk classification in infants, using SMOTE for data balancing. Evaluation was conducted using accuracy, precision, recall, and F1-Score metrics. The results show that the Random Forest algorithm provides fairly high performancewith an accuracy of 79.17%, precision of 90.54%, recallof 82.43%, and an F1-Score of 86.29%. Based on these results, it can be seen that proper preprocessingplays a significant role in improving classification performance. Based on this,it can be concluded that the Random Forest model is capable of building a classification model that is quite accurate inidentifying stunting status in toddlers. With a total of 10,000 entries, this model successfully recognized patterns in the datawith fairly good performance results.