Implementation of Machine Learning Model for Pneumonia Classification Based on X-Ray Images

Abstract

This study applies a Machine Learning learning model for Pneumonia classification based on x-ray images. This study uses two classes, namely Pneumonia Class and Normal Class, and uses Epoch 10, Epoch 50, Epoch 250 and Epoch 700, Learning Rate 0.001, and Batch Size 16. Learning carried out using Epoch 10 to get accuracy results per class is Pneumonia Class 0.97 and Class 0.95. While learning using Epoch 50 gets accuracy results per class, namely Pneumonia Class 0.97 and Normal class 0.97, and for learning, using Epoch 250 gets accuracy results for Pneumonia Class 1.00 and Normal Class 0.97. By using Epoch 700, the accuracy results were obtained for Pneumonia Class 1.00 and Normal Class 1.00. From the results of tests carried out using Learning Rate 0.001, Batch Size 16 and Epoch 10 received an accuracy of 64%. For Learning Rate 0.001, Batch Size 16 and Epoch 50 obtained 86% accuracy, and for Learning Rate 0.001, Batch Size 16 and Epoch 250 got 87% accuracy, while for Learning Rate 0.001, Batch Size 16 and Epoch 700 get 92% accuracy. From this study, the results show the highest precision using Epoch 700.