Stunting Detection in Toddlers Using Support Vector Machine (SVM) Classification

Project Portofolio Machine Learning Prediction Model

Stunting is the impaired growth and development that children experience from poor nutrition, repeated infection, and inadequate psychosocial stimulation. Children are defined as stunted if their height-for-age is more than two standard deviations below the WHO Child Growth Standards median. — World Health Organization (WHO)

Stunting detection in toddlers is a critical aspect of early childhood health monitoring, as it directly correlates with long-term physical and cognitive development. One of the emerging techniques for stunting detection is the application of Support Vector Machine (SVM) classification. SVM is a supervised machine learning algorithm that is widely used for classification tasks due to its effectiveness in handling high-dimensional data and providing accurate results even with small datasets. In the context of stunting detection, SVM can be used to classify toddlers into categories like severely stunted, stunted, normal, and tall based on various input features such as such as age, height, weight, and other nutritional indicators. The SVM model works by finding the optimal hyperplane that separates different classes with the maximum margin, thus ensuring the most reliable prediction.

The dataset used in this study is sourced from Kaggle.com. This dataset contains detailed information relevant to stunting detection in toddlers, including variables such as age, gender, height, and nutritional status. You can access the dataset through the following link: https://www.kaggle.com/datasets/rendiputra/stunting-balita-detection-121k-rows/data

Goals

The study aims to to develop a reliable system for detecting stunting in toddlers using Support Vector Machine (SVM) classification. By utilizing features such as age, gender, height, and nutritional status, the system aims to accurately classify toddlers into nutritional categories, enabling healthcare professionals to identify at-risk children and implement timely interventions to support their growth and development.

Result and Conclusion

1. The SVM model with optimized hyperparameters (C=100, gamma=1, kernel='rbf') delivers highly accurate results for stunting detection in toddlers.
2. The model achieves 99% accuracy, demonstrating its reliability in predicting stunting categories, including "severely stunted," "stunted," "normal," and "tall."
3. Precision, recall, and f1-scores of 0.99 for all categories highlight the model's consistency and balance in its predictions, with minimal classification errors across all labels.

References

1. https://ourworldindata.org/grapher/share-of-children-younger-than-5-who-suffer-from-stunting
2. https://www.kaggle.com/datasets/rendiputra/stunting-balita-detection-121k-rows/data
3. www.geeksforgeeks.org/support-vector-machine-algorithm
4. Banurea, M., Hutagaol, D. B., & Sihombing, O. (2023). Klasifikasi Penyakit Stunting dengan Menggunakan Algoritma Support Vector Machine dan Random Forest. Jurnal Tekinkom (Teknik Informasi dan Komputer), 6(2), 540-549.
5. Nugroho, A. F. F. (2024). Klasifikasi Stunting dan Status Gizi Balita dengan Metode SVM (Support Vector Machine). Skripsi

If you have any suggestions or feedback, please don't hesitate to contact to me in direct message on Email and LinkedIn: rahfaqurniya@gmail.com and www.linkedin.com/in/rahfaquraniyatin