EUVIMEDEUVIMED entdecken
Crossref

Prediction of Childhood Anemia using Supervised Machine Learning Models: LASSO Feature Selection Approach

Ravi Kumar, Shubhra Katara, Sabir Ali Siddiqui

INTERNATIONAL JOURNAL OF AGRICULTURAL AND STATISTICAL SCIENCES · 2026-06 · 22 (01) · 93

DOI: 10.59467/ijass.2026.22.93 ↗

Kurz gefasst

Fachartikel in der Zeitschrift INTERNATIONAL JOURNAL OF AGRICULTURAL AND STATISTICAL SCIENCES, 2026-06. Verfasst von Ravi Kumar und 2 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Childhood anemia remains a persistent public health concern in India. Accurate and interpretable predictive models are essential for early identification and targeted intervention. Since this study aims to evaluate and compare the performance of multiple machine learning models in predicting childhood anemia using NFHS-5 data for Uttar Pradesh, with a focus on feature selection through LASSO regression. Study utilized data from 26,283 children aged 6-59 months extracted from the NFHS-5 Kids Recode file for Uttar Pradesh. After applying Least Absolute Shrinkage and Selection Operator (LASSO) for feature selection, the dataset was analyzed using various machine learning models, including Naive Bayes (NB), Random Forest (RF), Support Vector Machine (SVM), Logistic Regression (LR), Decision Trees (DT), and Neural Networks (NN). The models were trained on 80% of the data and tested on the remaining 20%. Comparison of models was done through performance metrics included Accuracy, AUC, Precision, Recall, F1 Score, and Sensitivity. LASSO identified maternal anemia, stunting, child age (12-23 months) and regional factors as key predictors. Logistic Regression, SVM and Decision Tree achieved the highest accuracy (0.757) and perfect sensitivity. Random Forest and Neural Networks showed comparable recall but added complexity with limited gain. Naive Bayes had the lowest accuracy (0.67), but achieved a strong balance with precision (0.786), recall (0.776). LR, SVM and DT showed high sensitivity, while NB offered better precision, useful for minimizing over treatment. LASSO-based feature selection identified maternal anemia, stunting and child age as key predictors, with regional and age-based disparities also observed. These findings highlight the value of interpretable models and feature selection in early anemia detection, with future work recommended on deep learning, advanced feature selection, and integration with clinical systems.. KEYWORDS :Childhood anemia, Machine learning, NFHS-5, Uttar Pradesh, LASSO regression.

Zitieren

APA

Kumar, R., Katara, S., & Siddiqui, S. A. (2026). Prediction of Childhood Anemia using Supervised Machine Learning Models: LASSO Feature Selection Approach. INTERNATIONAL JOURNAL OF AGRICULTURAL AND STATISTICAL SCIENCES, 22(01), 93. https://doi.org/10.59467/ijass.2026.22.93

Vancouver

Kumar R, Katara S, Siddiqui SA. Prediction of Childhood Anemia using Supervised Machine Learning Models: LASSO Feature Selection Approach. INTERNATIONAL JOURNAL OF AGRICULTURAL AND STATISTICAL SCIENCES. 2026;22(01):93. doi:10.59467/ijass.2026.22.93

BibTeX
@article{Kumar2026,
  author = {Kumar, Ravi and Katara, Shubhra and Siddiqui, Sabir Ali},
  title = {Prediction of Childhood Anemia using Supervised Machine Learning Models: LASSO Feature Selection Approach},
  journal = {INTERNATIONAL JOURNAL OF AGRICULTURAL AND STATISTICAL SCIENCES},
  year = {2026},
  volume = {22},
  number = {01},
  pages = {93},
  doi = {10.59467/ijass.2026.22.93}
}
Kostenlos starten →