<![CDATA[Kidney has an important role in human body. Decreased kidney function can result in chronic kidney disease. Detection of chronic kidney disease using medical data that use features as factors of kidney failure, this data can be processed and build an intelligent system from it, that can help detect kidney failure. In the process of chronic kidney disease data classification method can be used. Modified K-Nearest Neighbor (MKNN) is one of classification method. But the weakness of the MKNN method is in the process using all available features, so it can cause an error detection because there are some features that are less relevant. Therefore in this research, a feature selection method is added, namely Information Gain, Information Gain method calculates the Gain value for each feature, features with large Gain values ​​will be better used for the classification process. On the results of testing variations numbers of features after selection and the effect of variations of K values, produces the highest accuracy value on 4 features with K values = 2 and 4 produces an accuracy value of 97,7% and on 6 features with K values = 2 and 4 produces an accuracy value of 97,7%. For the system testing using Information Gain produces an accuracy value of 96,8% and those not using Information Gain produce an accuracy value of 79,9%.]]>
Copyrights © 2019
