The Use of Artificial Intelligence in Disease Diagnosis: A Systematic Literature Review

Abstract

This paper discusses the crucial role of Artificial Intelligence in improving disease diagnosis and the use of medical data in the era of big data. Through the Systematic Literature Review (SLR) approach, we review recent developments in the application of Machine Learning (ML) to disease diagnosis, evaluate the ML techniques applied, and highlight their impact on healthcare. BP-CapsNet uses the Convolutional Capsule Network (CapsNet) to diagnose cancer with advantages in overcoming invariance to image transformation. The Stacking Classifier achieves 92% accuracy in detecting heart defects with the advantage of combining weak learners. CraftNet, a combination of deep learning and handmade features, demonstrates strong recognition capabilities in cardiovascular disease. ML in infectious diseases demonstrates the ability to process big data, focusing on bacterial, viral, and tuberculosis infections. In heart disease diagnosis, ML, especially with CNN and DNN, detects disease at an early stage, despite the challenges of data imbalance. The ensemble algorithm for heart disease prediction demonstrates the superiority of categorical medical features, with SVM and AdaBoost as suitable methods. A new CNN for wide QRS complex tachycardia provides accurate results. In vestibular disease, five ML algorithms provide satisfactory results, with SVM as the best. These findings detail the development of ML in disease diagnosis, highlighting future challenges and opportunities in its use

Ähnliche Arbeiten