An Algorithm for Clustering and Classification of Medical Datasets Using k-Means and Radial Basis Function Neural Networks

Classification and clustering is an integral part of any pattern recognition and data mining system. These methods allow the using huge size of databases is a challenging issue for machine learning and data sciences. This paper presents a hybrid approach by adopting classification via clustering with simple k-means algorithm with Euclidean distance function and radial basis function network for classification. The proposed approach conventionally used in medical datasets such as diabetes diagnostics, hepatitis and liver diseases these are commonly detected and reduce the quality of human life. These medical datasets such as liver diseases, hepatitis and diabetes datasets were obtained from University of California, Irvine (UCI)data repositories and to test the classifier accuracies with other popular methods on these datasets. Experiment results show that the proposed method performance is improved classifier accuracy and robustness for data mining and machine learning applications.