Prediction of Chronic Kidney Disease using Data Mining Feature Selection and Ensemble Method

Authors: ,

Abstract: The failure of the kidney is affected the whole human body and it can be a cause of the seriously ill and cause of deaths. Machine learning and data mining techniques are the most significant role in disease prediction with high-performance rate and used to help decision makers to assemble and understand information. The performance of classification techniques depends on the feature of the data set. To improve the accuracy of classification used feature selection method by reducing the dimensions of the feature and used ensemble or combine a model of the algorithm. In this research K-Nearest Neighbor, J48, Artificial Neural Network, Naïve Bayes and Support Vector Machine classification techniques were used to diagnose Chronic Kidney Disease. To predict chronic kidney disease, build two important models. Namely, feature selection method and ensemble model. To build chronic kidney disease prediction, used Info gain attributes evaluator with ranker search engine and wrapper subset evaluator with the best first engine was used. The result showed that the K-nearest neighbor classifier by using Wrapper Sub set Evaluator with Best first search engine feature selection method has 99% accuracy, J48 with Info Gain Attribute Evaluator with ranker search engine has 98.75, Artificial Neural Network with Wrapper Sub set Evaluator with Best first search engine has 99.5% accuracy, Naïve Bayes with Wrapper Sub set Evaluator with Best first search engine has 99% accuracy, Support Vector Machine with Info Gain Attribute Evaluator with ranker has 98.25% accuracy in prediction of chronic kidney disease compared to other with and without feature section method. The second model building method ensemble model by combing the five heterogeneous classifiers based on a voting algorithm. The effectiveness of the proposed ensemble model was examined by comparison of the base classifier. The experimental result showed that the proposed ensemble model achieved 99% accuracy.

WSEAS Transactions on Information Science and Applications, ISSN / E-ISSN: 1790-0832 / 2224-3402, Volume 15, 2018, Art. #19

PDF
XML
Cite
×
Citation Tools
WSEAS AMA
Sirage Zeynu, Shruti Patil, "Prediction of Chronic Kidney Disease using Data Mining Feature Selection and Ensemble Method," WSEAS Transactions on Information Science and Applications, vol. 15, pp. 155-167, 2018, DOI:
Sirage Zeynu, Shruti Patil. Prediction of Chronic Kidney Disease using Data Mining Feature Selection and Ensemble Method. WSEAS Transactions on Information Science and Applications. 2018;15:155-167.
Copy to Clipboard
Citation copied to Clipboard

Certification

---