Optimizing Diabetes Forecasting by Comparing Random Forest and Support Vector Machine Algorithms

This research compares the Random Forest Algorithm with the SVM for predicting diabetes. Using a sample size of 133, we evaluated the accuracy of both algorithms with varying training and testing splits. The Random Forest Algorithm achieved an accuracy of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 8. 4 9 \%}$</tex>, while the SVM reached <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8 1. 2 7 \%}$</tex>. Statistical analysis showed a significant difference between the two methods <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(p&lt;0.05)$</tex>, with a power of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8 0 \%}$</tex>. The findings indicate that the Random Forest Algorithm provides superior accuracy in diabetes prediction compared to the SVM. This study is significant because accurate diabetes prediction can lead to better early diagnosis and treatment, potentially improving patient outcomes and reducing healthcare costs. By demonstrating the effectiveness of the Random Forest Algorithm, this research highlights a promising tool for enhancing diabetes prediction and management.