Chronic Kidney Disease Early Diagnosis Enhancing by Using Data Mining Classification and Features Selection

Chronic Kidney Disease (CKD) is currently a worldwide chronic disease with an increasing incidence, prevalence and high cost to health systems. A delayed recognition prevention often lead premature mortality due progressive incurable loss of kidney function. Data mining classifiers employment discover patterns in CKD indicators would contribute early diagnosis that allow patients prevent such severe damage. Adopting the cross Industry Standard Process Mining (CRISP-DM) methodology, this work develops classifier model support healthcare professionals patients. By building data pipeline manages different phases CRISP-DM, automated transformation, modelling evaluation applied dataset extracted from UCI ML repository. Moreover, along Scikit-learn package’s GridSearchCV used carry out exhaustive search best parameters preparation’s sub-stages like missing feature selection. Thus, AdaBoost selected as it outperforms 100% terms accuracy, precision, sensivity, specificity, f1-score roc auc, classification results obtained by related works reviewed. application selection reduces up 12 24 features which are employed developed.