Enhanced fuzzy joint mutual information with Cuckoo Search Algorithm and genetic algorithm for diagnostics of chronic kidney disease prediction

Advanced methods for analyzing data are in high demand as the volume of high-dimensional data used in healthcare services increases, particularly for the early detection and management of Chronic Kidney Disease (CKD). False diagnosis and treatment of CKD have been rendered more challenging by the complicated nature and the several features detected in these collections of data. Optimizing the diagnostic result depends significantly on the efficient analysis and processing of such complex data. Recent healthcare data analysis tools, including the Feature Selection Model (FSM), face challenges when processing large volumes of historical data with numerous features. Overfitting, a result of the Curse of Dimensionality (CoD), diminishes the accuracy of algorithms developed using conventional methods. Less-than-optimal feature selection (FS) and reduced Classification Accuracy (CA) have also resulted from the failure of multiple modern methods to address the fuzziness and uncertainty inherent in healthcare data features. The current study introduces the Enhanced Fuzzy Joint Mutual Information-Cuckoo Search Algorithm-Genetic Algorithm (EFJMI + CSA-GA) integrated FSM to address the earlier faults. The method defined here integrates EFJMI's fuzzy data processing methods with the Cuckoo Search Algorithm-Genetic Algorithm (CSA-GA) optimization. Applying the University of California Repository's CKD data collection, the EFJMI + CSA-GA was developed. This model has been able to minimize the complete feature set by 62%, which is related to the size of the features. The experimental results from clinical tests using many classification algorithms validate that the FSM-recommended Recurrent Neural Network (RNN) classifier performs effectively. It performs similarly to other FSMs, achieving 99.71% CA, 99.81% sensitivity, and 99.58% specificity.