Stunting Factor Analysis using Machine Learning for Under Aged Children in the Developing Countries

Abstract

Malnutrition and stunting which comes from it can be put in the list of the top problems of any developing countries. A country like Bangladesh is no different. The principal contribution of this study is to predict the Stunting status of the children and also to search for the related affecting factors which affect the nutrition status. In this research, data has been collected from Bangladesh Demographic Health Survey (BDHS)’2017-18. Nutrition status correlated with the child’s age, mother’s education, father’s education, father’s employment, family wealth index, currently breastfeeding, place of residence and division. The differential impact of some sectors like demographical and socioeconomical, environment plus health-affecting determinants on the nutritional rank within the population of under-five children in Bangladesh has been taken into account. To measure the child nutritional condition of under-five children among various methods, this Z-score method is one which we have used in this paper to find the statistics of malnutrition in Bangladesh. Methods provided by WHO (World Health Organization) was followed to find out the necessary outcomes. Here, Chi-square statistics algorithm has been applied to find out the factors which are most responsible for stunting by ranking features and then applied machine learning algorithms for a prediction model. Two algorithms have been used here and based on the performance results and parameter we find Logistic Regression gave most accuracy out of them all. It’s found that based on all features the accuracy is 77.00% and based on top 5 features the accuracy is 63.00% and based on top 7 features the accuracy is 77.00% which is better than the outcomes of Random Forest Algorithm. The study is suggested to focus upon the factors which are responsible for malnutrition and it would ensure healthy nation