Deep learning based thyroid prediction with opposition learning based red panda optimization feature selection

This research introduces a novel approach for thyroid prediction by considering three various publicly available datasets. The input data from the dataset is preprocessed to ensure standardization and balance for mitigating the biased outcomes. Then, the proposed cascaded autoencoder-based simple recurrent model is employed for extracting significant spatio-temporal features. From the extracted features, the optimal feature is selected using the proposed Opposition Learning-based Red Panda Optimization (OL_RPO) algorithm, which enhances the efficiency of the predictive model. Finally, the thyroid prediction is performed using the Enhanced Transformer Model, which uses the selected features to achieve robust and accurate predictions. The analysis of the proposed model based on Accuracy, Specificity, Sensitivity, F-Score, positive predictive value (PPV), negative predictive value (NPV), and Error acquired the value of 99, 99.2, 99.01, 98.501, 98.1, 1.9, and 0.07689 respectively.