Predicting Pregnancy Delivery Outcomes using Machine Learning Algorithms
DOI:
https://doi.org/10.5281/zenodo.13272742Keywords:
WHO, KDHS, ANC, Machine Learning, ANNs, SVMs, GLMs, Random Forest Gini ScoreAbstract
The World Health Organization (WHO) has recognized the unpredictability of pregnancy delivery outcomes, including stillbirth, miscarriage, induced abortion, and live births, as a significant issue. Accurate prediction of these outcomes is crucial for improving healthcare treatments for mothers and children. This research uses a dataset from the Kenya Demographic and Health Survey (KDHS) to explore machine learning algorithms for predicting these outcomes. Pregnancy delivery outcomes are complex and rely on various demographic factors, including the number of antenatal care visits, the respondent's age, wealth index, marital status, place of residence, geographic location, religious affiliation, and the overall count of births. The research aims to develop reliable machine learning models that accurately predict these outcomes by examining the complex interactions between demographic factors. Various approaches, including neural network algorithms (ANNs), support vector machines (SVMs), and generalized linear models (GLMs), will be employed to find the model that best fits the needs of understanding and predicting the range of pregnancy delivery outcomes while accounting for these factors. The study also aims to reveal important correlations between Kenyan population demographic characteristics and delivery outcomes, with preliminary studies suggesting that the Random Forest Gini score may be a key factor in determining delivery results.
Downloads
Downloads
Published
Issue
Section
License
Copyright (c) 2024 Gabriel Mbatha Ngao, Abraham Mutua Matheka
This work is licensed under a Creative Commons Attribution 4.0 International License.