Kodi Owen, Obadiah Musau, Mvurya Mgala

ABSTRACT

Preterm births affect millions of newborns worldwide each year and are a major public health concern. Because preterm births are complex, it is still difficult to anticipate them accurately, despite advances in medical research. In order to overcome the shortcomings of current diagnostic procedures and pursue proactive intervention measures, this study intends to improve the prediction of preterm deliveries by utilizing machine learning techniques.

Various datasets covering a range of factors, including physiological signals, ultrasound data, and maternal health records, will be utilized in this study. Decision trees, logistic regression, and support vector machines are used to study and find predictive patterns linked to premature births. To increase the precision and dependability of the prediction models, entropy-based feature selection strategies and thorough model evaluation procedures are applied.

The creation of reliable predictive models that can precisely identify pregnancies at high risk of premature births is one of the study’s anticipated results. In order to facilitate early identification and focused intervention efforts to lower the frequency of preterm births, machine learning algorithms are used to a variety of datasets in an effort to reveal hidden patterns and risk factors linked with premature deliveries.

The aim of this research is to make a substantial contribution to the field of obstetrics by showcasing the potential of machine learning to enhance the accuracy of preterm birth prediction. If these predictive models are successfully put into practice, early treatments may result, which could lessen the negative consequences linked to preterm deliveries. Furthermore, the knowledge gathered from this research may facilitate the creation of individualized treatment plans designed to reduce the risk factors associated with premature births.