Analysis of the Neonatal Sepsis Data Set with Data Mining Methods

Abstract

With the developing technologies, many data that are not measured or measurable in the past but not recorded are now recorded and stored in the databases. Classical statistical knowledge has made it impossible to extract meaningful data from a small number of data, but data mining methods have emerged by combining classical statistical methods with modern technology. With data mining methods it is now easier to establish decision support systems with meaningful data obtained from meaningful data among many data stacks. The purpose of this study is to address the relationship between data mining methods and health services. It is not a system, method or method to decide on behalf of physicians. Ultimately, the best decision is the personal observation and experience of the physician. The intensive business will work in the sense of developing an intelligent system that will support the tempo of decision making and the complexity of data among the tempo. From this, the KNN algorithm was applied to the Sepsis data set and 121 samples were classified correctly in a total of 128 samples. The accuracy of the KNN algorithm is 94.53% for this data set. Naive Bayesian algorithm was applied to the Sepsis data set and the accuracy rate was calculated as 93.73%. © 2019 IEEE.