Performance Analysis of the Decision Tree Classification Algorithm on the Pneumonia Dataset

Abstract

The rapid advancements in machine learning have paved the way for innovative approaches in medical imaging diagnostics. In this context, this study explored the efficacy of the Decision Tree Classification Algorithm for distinguishing between normal and pneumonia-diagnosed X-ray images. We sourced our dataset from pediatric X-rays obtained from the Guangzhou Women and Children’s Medical Center. To enhance the classifier's performance, a methodical pre-processing strategy was adopted. This encompassed the application of the Canny segmentation technique, followed by feature extraction using humoments. The evaluation phase involved a 5-fold cross-validation, revealing a commendable average accuracy of 82.72%. These findings highlight not only the utility of Decision Trees in such specialized diagnostic tasks but also accentuate the pivotal role of systematic pre-processing in achieving optimal results. As medical diagnostics steadily move towards automation, this research provides valuable insights and benchmarks for future endeavors aiming to harness the power of machine learning in healthcare.