Volume : VIII, Issue : VI, June - 2019
DIABETES PREDICTION USING STACKING AND COST-SENSITIVE LEARNING
Ferhat Avdic, Nejdet Dogru
Abstract :
According to the World Health Organization diabetes mellitus is the cause of millions of deaths around the world. Diabetes-related issues may be avoided when the patient is diagnosed and treated on time. Machine learning is used to create decision support systems in healthcare which aid in diagnosis. The Pima Indian dataset was used for patient classification. Algorithms such as Logistic Regression, Support Vector Machine, k-Nearest Neighbor, Naive Bayes, C4.5 Rules were used to create the prediction models. This study aimed to achieve a higher prediction accuracy for the onset of diabetes using a cost-sensitive stacking model. It performed better than the models in the literature, reaching approximately 85% accuracy, 86% specificity, and 85% sensitivity.
Keywords :
Article:
Download PDF
DOI : https://www.doi.org/10.36106/gjra
Cite This Article:
DIABETES PREDICTION USING STACKING AND COST-SENSITIVE LEARNING, Ferhat Avdic, Nejdet Dogru GLOBAL JOURNAL FOR RESEARCH ANALYSIS : Volume-8 | Issue-6 | June-2019
Number of Downloads : 298
References :
DIABETES PREDICTION USING STACKING AND COST-SENSITIVE LEARNING, Ferhat Avdic, Nejdet Dogru GLOBAL JOURNAL FOR RESEARCH ANALYSIS : Volume-8 | Issue-6 | June-2019