Using ANN and association rules to analyze maternal health-related variables and child growth in India: a cross-sectional study

The mother is often the main caregiver of the child and the growth and well-being of the child is often dependent on the health of the mother. The study was carried out to predict underweight among children aged 2 to 4 years in India given maternal health-related variables. An objective was also to identify maternal health-related factors that were associated with underweight. The National Family Health Survey (NFHS-5), 2019–2021 Indian dataset was used and an artificial neural network (ANN) model predicted underweight using maternal health-related variables. Association rule analysis was also used to identify maternal health-related factors associated with underweight and healthy weight in children. Accuracy and precision of over 70% were obtained from the prediction using the ANN model. Children of mothers who had prenatal doctor visits were more likely to be of healthy weight. Mothers who received supplementary nutrition during pregnancy and those who did not have prenatal doctor visits were more likely to have children who were underweight. There is a possible socioeconomic link here, where mothers who received supplementary nutrition are more likely to be in the low economic bracket. This may be driving the association with underweight among their children. Financial support beyond the gestation period should be provided to mothers who received supplementary nutrition and expectant mothers should be encouraged to have prenatal doctor visits.