Public Health Nurse Perspectives on Predicting Preterm Labor Using Risk Factors and Simple Machine Learning Algorithms

Preterm birth (PTB) is one of the most precarious obstetrical conditions, which is one of the leading causes of infant mortality. The capacity to predict PTB during both the first and second trimesters provides great promise for improving pregnancy outcomes. The objective of this study was to predict the preterm labor using various risk factors and machine learning models. Executing a novel methodology, researchers used risk factors of preterm labour and machine learning algorithms to predict premature labour. Our data was normalized using the continuous-discrete variables technique, resulting in a single feature value. Our investigation used various prediction models, such as Naïve Bayes, Neural Network (NN), Stochastic Gradient Descent (SGD), AdaBoost, Gradient Boosting (GB), CN2 rule inducer, and k-nearest neighbors (KNN), to predict PTB. This study includes a total of 300 samples for complete analysis. The results showed that GB, AdaBoost, and the CN2 rule inducer have better accuracy ratings of 0.950, 0.947, and 0.930, respectively, along with outstanding Area Under Curve (AUC) values of 0.996, 0.996, and 0.983. Furthermore, with Precision scores of 0.950, 0.947, and 0.930, these models showed strong performance in predicting the probability of PTB. To conclude, GB, adaboost, and CN2 rule inducer models accurately predict PTB, with high AUC and precision scores, making them useful tools for predicting PTB.