An Ensemble Classification Model for Medical Databases Using Hybrid Weights

Large databases are now frequently utilized to identify and diagnose medical disorders using extreme learning procedures. Due to its promising implementation and processing speed, the fundamental model of the study utilized for real-time applications is the ensemble classifier. Standard approaches associated with extreme learning practices project the inability for error prediction based on output layer hidden's selection under static weight selection. In this study, a unique weighted extreme learning machine (WELM) for predicting medical conditions is introduced. This research will help to solve the classification problems in the field of healthcare and medicine holds significant promise for improving the accuracy, validation, accountability, and reliability of medical data classification tasks. Moreover, the key objective for the weighted extreme learner's approach is to predict the illness by outlining the high-dimensional data applied for the case study. Typically, the practice proposed an ensemble model which functions the enhances the accurate predictions of high-dimensional cancer showing significant impact of the diagnosis and treatment planning. Furthermore, the WELM model effectiveness is validated through other ensemble learning simulations that include neural networks, random forest, PSO + NN, and ACO + NN techniques. Evaluation for the outcomes is verified through various medical datasets with attributes of the liver, ovarian, lung, diabetes, and DLBCL-Stanford. The results show that the WELM described is very computationally efficient which is related to true positive rate, accuracy, and error rate.