Eamining and Comparing Data Mining-Based Techniques for Hepatitis Diagnosis
Increasing advances in information technology has led to significant growth in sciences. One of the fields in which significant changes has occurred is the medical field. Using data- mining techniques in this branch of science has helped physicians in all subjects, in particular diagnosis of sicknesses. Hepatitis diagnosis is highly difficult due to limited clinical diagnosis of the disease in its early stages. To this end, this paper tries to introduce and recommend the best way to diagnose hepatitis as well as to compare common clustering methods such as decision trees, neural networks, and SVM. Evaluation criteria of classification methods are the accuracy of each of methods and Clementine software along with data base in the University of California has been used to test each method. Obtained results show that neural network algorithm enjoys higher accuracy in comparison with other algorithms. Using neural network algorithm can accurately predict 89.74% hepatitis.