Abstract:

This research endeavors to leverage advanced AI algorithms for the differentiation between typically developing children and those afflicted with Intrauterine Growth Restriction (IUGR) disease. We deployed three distinct AI algorithms: Quadratic Discriminant Analysis (QDA), Linear Discriminant Analysis (LDA), and Support Vector Machine (SVM). These algorithms were applied to the task of classifying the two target groups while discerning the pivotal parameters contributing to their classification. The obtained results yielded classification accuracy scores of 92.89%, 89.89%, and 87.67% for QDA, LDA, and SVM, respectively. Notably, the analysis revealed that parameters related to birth weight held the most substantial influence in distinguishing between the two cohorts. In light of our conclusive findings, we recommend the utilization of Quadratic Discriminant Analysis (QDA) as a valuable tool for clinicians seeking to identify children at risk of IUGR disease. This research contributes to the enhancement of diagnostic methodologies in pediatric medicine, fostering more accurate and timely interventions for affected individuals.

This was presented as a poster in CinC 2023, Atlanta, Georgia. The link to the full paper is: