The high prevalence of chronic kidney disease (CKD) is a significant public health concern globally. The condition has a high mortality rate, especially in developing countries. CKD often go undetected since there are no obvious early-stage symptoms. Meanwhile, early detection and on-time clinical intervention are necessary to reduce the disease progression. Machine learning (ML) models can provide an efficient and cost-effective computer-aided diagnosis to assist clinicians in achieving early CKD detection. This research proposed an approach to effectively detect CKD by combining the information-gain-based feature selection technique and a cost-sensitive adaptive boosting (AdaBoost) classifier. An approach like this could save CKD screening time and cost since only a few clinical test attributes would be needed for the diagnosis. The proposed approach was benchmarked against recently proposed CKD prediction methods and well-known classifiers. Among these classifiers, the proposed cost-sensitive AdaBoost trained with the reduced feature set achieved the best classification performance with an accuracy, sensitivity, and specificity of 99.8%, 100%, and 99.8%, respectively. Additionally, the experimental results show that the feature selection positively impacted the performance of the various classifiers. The proposed approach has produced an effective predictive model for CKD diagnosis and could be applied to more imbalanced medical datasets for effective disease detection.

In recent times, several machine learning models have been built to aid in the prediction of diverse diseases and to minimize diagnostic errors made by clinicians. However, since most medical datasets seem to be imbalanced, conventional machine learning algorithms tend to underperform when trained with such data, especially in the prediction of the minority class. To address this challenge and proffer a robust model for the prediction of diseases, this paper introduces an approach that comprises of feature learning and classification stages that integrate an enhanced sparse autoencoder (SAE) and Softmax regression, respectively. In the SAE network, sparsity is achieved by penalizing the weights of the network, unlike conventional SAEs that penalize the activations within the hidden layers. For the classification task, the Softmax classifier is further optimized to achieve excellent performance. Hence, the proposed approach has the advantage of effective feature learning and robust classification performance. When employed for the prediction of three diseases, the proposed method obtained test accuracies of 98%, 97%, and 91% for chronic kidney disease, cervical cancer, and heart disease, respectively, which shows superior performance compared to other machine learning algorithms. The proposed approach also achieves comparable performance with other methods available in the recent literature.