P. Shanthini

Performance Comparison of Support Vector Machine and Naïve Bayes Classifiers on a Synthetic Dataset: A Comprehensive Statistical Analysis

Under resource limits, NB is the preferred choice even though predictive accuracy does not differ significantly from SVM, and the practical distinction is computational: NB trained 22.5× faster, predicted 12× faster, used 2.9× less memory, and produced a 12× smaller model file.