A Survey and Compare the Performance of IBM SPSS Modeler and Rapid Miner Software for Predicting Liver Disease by Using Various Data Mining Algorithms

Authors : Moloud ABDAR

Pages : 3230-3241

View : 52 | Download : 14

Publication Date : 2015-05-13

Article Type : Review Paper

Abstract :Abstract. Today, with the development of industry and mechanized life style, prevalence of diseases is rising steadily, as well. In the meantime, the number of patients with liver diseases (such as fatty liver, cirrhosis and liver cancer, etc.) is rising. Since prevention is better than treatment, early diagnosis can be helpful for the treatment process so it is essential to develop some methods for detecting high-risk individuals who have the chance of getting liver diseases and also to adopt appropriate solutions for early diagnosis and initiation of treatment in early stages of the disease. In this study, we tried to use common data mining techniques that are used nowadays for diagnosis and treatment of different diseases, for the diagnosis and treatment of liver disease. For this purpose, we used Rapid Miner and IBM SPSS Modeler data mining tools together. Accuracy of different data mining algorithms such as C5.0 and C4.5, Decision tree and Neural Network were examined by the two above tools for predicting the prevalence of these diseases or early diagnosis of them using these algorithms. According to the results, the C4.5  and C5.0  algorithms by using IBM SPSS Modeler and Rapid Miner tools had 72.37% and 87.91% of accuracy respectively. Further, Neural Network algorithm by using Rapid Miner had the ability of showing more details.
Keywords : Data mining techniques, Liver diseases, Rapid Miner, IBM SPSS Modeler

ORIGINAL ARTICLE URL