Enhancing Accuracy in Early Pneumonia Diagnosis with a Novel Support Vector Machine and Convolutional Neural Network

This study aimed to enhance the early detection of pneumonia by using a new CNN instead of the traditional SVM technique. For this purpose, we used a dataset from Kaggle as our primary data source. We divided the study into two groups of 20 samples each: Group 1 used the CNN, while Group 2 used the SVM. Overall, the study involved 40 samples. To assess the performance of these two methods, we conducted statistical analysis using Python and clincalc.com, setting the statistical power (G-power) at 85%, the alpha level at 0.05, and the beta level at 0.2. Our analysis focused on comparing the accuracy of the CNN and the SVM. The results showed that the CNN achieved an accuracy of 96.1%, significantly higher than the 75% accuracy of the SVM. The statistical tests revealed a significant difference between the two methods, with a p-value less than 0.001.