Predicting internal diseases in humans using machine learning: A systematic literature review

Human health is the main focus of clinical medicine, especially in understanding internal diseases involving the body's organs. Identifying and predicting disease at an early stage is essential to prevent the development of more severe disease. These challenges encourage using the latest technologies, especially machine learning techniques. This technology is used to ensure accurate disease predictions. The results of the research identified various types of internal diseases, including heart, kidney, lung and liver cancer, and highlighted the associated symptoms and risk factors. Several algorithms are used to classify internal diseases, including the classification of heart disease. The logistic regression algorithm is the most efficient, with accuracy results of 88.52%. Meanwhile, CHIRP kidney disease classification provides the most efficient results with an accuracy of 99.75%. MobileLungNetV2 has an accuracy of 96.97% for lung disease classification, and classification for liver disease produces the highest accuracy in logistic regression at 72.50%. Machine learning in disease prediction significantly contributes, especially in increasing accuracy and efficiency in diagnosis and risk prediction. Despite significant progress, challenges such as dataset size, data quality, and model validation need to be addressed to maximise the potential of this technology in clinical practice.