A Machine Learning Study to Classify the Type of Anemia in Children Under 5 Years of Age

Peña-Cáceres O.J.M.; Chunga-Zapata E.A.; Basurco-Chambilla T.R.; Cayo-Cabrera G.H.; Villegas-Paz A.D. · Smart Innovation, Systems and Technologies (2023) · DOI: 10.1007/978-3-031-31007-2_46

Resumen

In the last years, machine learning has emerged as a support in decision-making for the diagnosis of various diseases, particularly in geographical areas with a high population incidence. The objective of this study was to make use of machine learning algorithms, including Decision Tree, Random Forest, Naive Bayes, Vector Support Machine, and Logistic Regression, to classify the type of anemia in children under 5 years. The data set used corresponds to the year 2019 and represents 45,764 rates of anemia diagnoses in the Piura region, in Peru. Information was obtained through the Open Data platform of the Peruvian Government. The Python programming language and the Google Golab cloud service were used for the coding and execution of each of the models. The results obtained show us that the random forest algorithm represents an accuracy of 100% for the training data set and 93% for the test data. The model was validated by performing 400 tests, the results of which were acceptable. It is still necessary to strengthen this type of study, introducing new variables that enrich it by developing these computational techniques associated with the experience of the human being.

Instituciones: Universidad Nacional del Altiplano; Universidad César Vallejo