Abstract:

Hepatitis is one of the most common diseases worldwide. Hepatic diseases can pose great dangers in the body. For these reasons, it is vital to diagnose and treat in a very early stage possible liver diseases such as alcohol-independent liver fertilization, alcoholic liver fertilization, hepatitis A, hepatitis B, hepatitis C, hepatitis D, liver growth, liver cancer, liver and intrahepatic gall canal cancer, liver insufficiency and sirosis. Traditional methods of diagnosis are still used in medicine. But today, thanks to the advanced artificial intelligence technologies, we can provide powerful tools to support doctors in the diagnosis, diagnosis and treatment processes of diseases. In this study, using the WEKA data mining tool, the J48, Logistics Model Tree (LMT), Decision Stump, Hoeffding Tree, REP Tree, Random Forest, Random Tree and IBk machine learning algorithms have been developed on the Liver Patient Data Set (ILPD). These algorithms have been made to get the best diagnosis results. For the performance and success of the machine learning algorithms used in the study, first mixture matrix was created and then calculations were made according to the standards that were accepted in general validity in medicine. The performance of these algorithms on the ILDP data set is calculated according to accuracy, sensitivity, specificity, ROC, MCC, memory, accuracy, and F-measurement criteria. The best models obtained from machine learning algorithms can be used to create decision-making mechanisms for intelligent systems that will be developed for diagnosis and treatment. The models we have achieved in this study have the potential to create a sub-structure of a smart system that will enable the early stages of the world’s widespread hepatic diseases to be captured in all hospitals.

Keywords: