A comparison of the performance of data mining classification algorithms on medical datasets with the application of data normalization

Medical research has evolved considerably at this time. In the past decade, a growing number of researchers have been conducting numerous medical research works. Existing medical research generally uses medical datasets. The medical research that is currently being carried out is related to the use of computers and datasets. Data processing on datasets usually uses data mining techniques, one of which is classification. The results from a classification process are generally dependent on the model formed. Based on this, it is necessary to perform a comparison of classification algorithms in data mining. The purpose of this comparison is to find out which algorithm has the best performance. The data processing carried out on medical datasets faces various obstacles, including the problem of the distance between scattered values. Normalization is a process of simplifying the data contained in a dataset. Based on the test results from the comparison performed, it was found that the K-NN algorithm has better performance than the Naïve Bayes algorithm. The K-NN algorithm obtained a 95.30% accuracy before normalization and 95.44% after normalization. Meanwhile, the Naïve Bayes algorithm obtained an accuracy of 86.28% before normalization and 95.44% after normalization. It can be summarized that normalization was better with the use of the Naïve Bayes algorithm, with an increase of accuracy level of 6.21%. On the other hand, the K-NN algorithm increased only by 0.14%.