Optimization of Random Forest Model with SMOTE for Fetal Health Classification Based on Cardiotocography

Fetal health during pregnancy is a crucial aspect in ensuring the optimal growth and development of a child, particularly during the golden period of life from the womb to the age of two. In the medical field, monitoring fetal conditions is vital to detect potential risks as early as possible. One of the tools commonly used in this process is cardiotocography (CTG), which provides essential data on fetal heart activity and movement. With technological advancements, machine learning-based approaches are increasingly being utilized to process CTG data more effectively. However, a major challenge in classifying medical data such as CTG lies in class imbalance, where the distribution between majority and minority classes is uneven. This study evaluates the effectiveness of the Synthetic Minority Over-sampling Technique (SMOTE) in addressing this imbalance and assesses the performance of the Random Forest algorithm in classifying fetal health conditions. The results show that the combination of SMOTE and Random Forest achieves the best performance compared to other methods, with an accuracy of 94.40%, precision of 94.45%, recall of 94.40%, and an F1-score of 94.38%. These findings indicate that SMOTE is effective in improving the representation of minority classes, while Random Forest demonstrates superior and consistent classification performance on CTG data