Application of rough set and support vector machines in hepatitis diagnosis
Xiaodong Zhu · Journal of Jilin University · 2011
A method applied in hepatitis diagnosis was proposed, which is based on rough set and Support Vector Machines (SVM). The method uses rough set to reduce the original features and to obtain a number of feature subsets. Then a set selection algorithm was employed to execute the feature reduction again, and a new data set was acquired in the light of the reduced subset. SVM was used for the new data set training and prediction. UCI machine learning public data set was adopted. Experiment results and data analysis show that, comparing with SVM, Neural Network (NN) and decision tree, the proposed method has higher diagnostic accuracy and can diagnose whether the data being negative or positive.