Hybrid CNN-SVM model for enhanced early detection of Chronic kidney disease

Dimensions

Ramu, K. and Sridhar, P. and Prajapati, Yogendra Narayan and Ramesh, Janjhyam Venkata Naga and Banerjee, Sudipta and Brahma Rao, K. B.V. and Alzahrani, Saleh I. and Rajaram, Ayyasamy (2025) Hybrid CNN-SVM model for enhanced early detection of Chronic kidney disease. Biomedical Signal Processing and Control, 100. ISSN 17468108; 17468094

Full text not available from this repository.

Abstract

Chronic kidney disease (CKD) progresses silently to extensive organ damage and end-stage kidney disease. Existing ML methods face issues like overfitting, slow computation, and class imbalance. This study proposes a hybrid CNN-SVM model to improve CKD prediction. CNN extracts features and SVM classifies them. Using a dataset with 10 medical indicators and SMOTE for balancing, the model achieved 96.8% accuracy, outperforming standalone SVM (94.8%) and Random Forest (94.6%). CKD recall was 1.00. The model overcomes overfitting and mitigates class imbalance, though computational load remains significant. The hybrid approach is a robust candidate for clinical application and future research.

Item Type: Article
Additional Information: Cited by: 11
Uncontrolled Keywords: Clinical research; Contrastive Learning; Deep neural networks; Diseases; Support vector machines; Chronic kidney disease; Chronic kidney disease classification; Convolutional neural network; Convolutional neural network-support vector machine; Disease classification; Early detection; Features extraction; Machine-learning; Network support; SMOTE; Support vectors machine; �chronic kidney disease; Convolutional neural networks; albumin; creatinine; glucose; hemoglobin; potassium; sodium; urea; age; albumin blood level; anemia; appetite; Article; artificial neural network; back propagation neural network; bacterium; Bayesian learning; Bayesian network; binary particle swarm optimization; blood pressure; chronic kidney failure; controlled study; convolutional neural network; coronary artery disease; creatinine blood level; cross validation; decision tree; deep belief network; diabetes mellitus; diagnostic accuracy; diagnostic test accuracy study; discriminant analysis; disease classification; erythrocyte; feature extraction; feature selection; foot edema; glucose blood level; hematocrit; hemoglobin blood level; human; hypertension; k nearest neighbor; kernel method; leukocyte count; logistic regression analysis; long short term memory network; multilayer perceptron; nonhuman; potassium blood level; prediction; principal component analysis; radial basis function neural network; random forest; recurrent neural network; relative density; sensitivity and specificity; sodium blood level; support vector machine; urea blood level
Subjects: Computer Science > Artificial Intelligence
Divisions: Medicine > Aarupadai Veedu Medical College and Hospital, Puducherry > Microbiology
Depositing User: Unnamed user with email techsupport@mosys.org
Date Deposited: 26 Nov 2025 09:22
Last Modified: 26 Nov 2025 09:22
URI: https://vmuir.mosys.org/id/eprint/296

Actions (login required)

View Item
View Item