期刊名称：Lecture Notes in Engineering and Computer Science
印刷版ISSN：2078-0958
电子版ISSN：2078-0966
出版年度：2019
卷号：2241
页码：298-303
出版社：Newswood and International Association of Engineers
摘要：Diabetes has emerged as the most chronic disease
that may cause mortality to the diabetic patients. However,
early prediction of diabetes can be helpful in reducing the
severe effects of diabetes. Several approaches have been
developed for diabetes prediction, recently, the data mining
based machine learning approaches have gained huge
attraction from research community. However, conventional
approaches suffer from the performance related issues. Hence,
in this work we introduce a novel data mining approach for
diabetes prediction and classification. The proposed approach
includes missing value imputation, data clustering, dimension
reduction and Bayesian regularized neural network
classification. The proposed approach is carried out using open
source available Pima Indian Diabetes dataset and
implemented using MATLAB simulation tool. The obtained
performance is compared with the existing techniques; this
comparative study shows a significant improvement in the
prediction performance using proposed approach.
关键词：Bayesian regularized neural network;
classification; Diabetes; Data clustering; Dimension reduction;
machine learning data clustering; dimension reduction
