Undesired Child Nexus Poor Health Condition: An Application of Machine Learning
Author: Urva Zanaib

ABSTRACT

One of the key indicators of human development in a child’s poor health condition practically is childhood morbidity and mortality. Although child morality is not desired, Pakistan nevertheless has a high rate of child morbidity and mortality. One of the psychological factors that may be associated with child morality is an unplanned pregnancy, whether unwanted (the parent did not want any more or more children) or mistimed (the pregnancy occurred sooner than hoped). Unplanned pregnancies and births are two psychological issues that harm children’s nutritional health (Shapiro-Mendoza et al., 2005); (Shaka et al., 2020). The goal was to evaluate the impact of mothers’ aspired status on the morbidity and mortality of children in Pakistan. We limited our analysis to children under 5 before the survey and used Pakistan demographic and health survey (2017-18), a national representative cross-sectional survey. By reducing the optimal number of children from total live births, we were able to estimate the undesired status (excess in boys, girls, both, and parity). We assessed morbidity (fever, diarrhea, cough, ARI, and SBR), nutritional status, and under 5 moralities. Finally, we perform machine learning techniques LDA, RF, SVM, and NN in the analysis of the data. The findings revealed that the overall percentage of the undesired child was 8%, 4%, 15%, and 27% for boys, girls, parity, and dual excess respectively. All the variable was associated with the undesired child. Child morbidity, fever, and cough were higher among the undesired children. We found little evidence that undesired children have acute respiration infection (ARI). We found very little evidence that an undesired child has a significant impact on childhood diseases. The ratio of child morality was less for boys but higher for girls.

Meta Data

Keywords : Linear discrimination Analysis, Machine learning, Neural Network, Random Forest, Support Vector Machine
Supervisor: Amena Urooj

Related Thesis​