Diabetes mellitus is one of the most common diseases among people of all age groups, affecting children, adolescents and young adults. There is an increasing interest in using machine learning techniques to diagnose these chronic diseases. However, the poor quality of most medical data sets inhibits construction of efficient models for prediction of diabetes mellitus. Without efficient preprocessing methods, dealing with these kinds of data sets leads to unreliable results. This paper presents an efficient preprocessing technique including a combination of missing value replacement and attribute subset selection methods on a well-known diabetes mellitus data set. The results show that the proposed technique can improve the performance of applied classifier and outperforms the traditional methods in terms of accuracy and precision in diabetes mellitus prediction.
展开▼
