Anemia Detection Model for Pregnant Women in Peru Using Machine Learning and Deep Learning

Gamarra S.; Velasquez R.; Ticona W. · Lecture Notes in Networks and Systems (2025) · DOI: 10.1007/978-3-031-96775-7_25

Resumen

Anemia is a serious health problem characterized by low blood red blood cell counts, affecting 37% of pregnant women worldwide. The objective of this study was to predict anemia in pregnant women using machine learning and deep learning techniques; the research can contribute to prenatal care and reduce anemia in pregnant women in Peru. The methodology was based on four (4) phases: Obtaining the dataset; Preprocessing; Model training: Machine Learning (SVM, Random Forest and Gradient Boosting); and Deep Learning (CNN and MLP) and Evaluation (Accuracy, Precision, Recall and F1-Score). The feature selection process using LightGBM indicated that variables such as altitude and the poverty index of the district of residence were closely related to the diagnosis of anemia, whereas variables such as the type of pregnancy were of little significance. The model trained with CNN obtained an accuracy of 82%, recall of 92%, precision of 75%, and an F1-score of 83%, suggesting the need to use datasets with variables that consider economic and geographic aspects. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

Instituciones: Universidad Tecnológica del Perú