Randomized Vector Network Model for Thyroid Prediction Using Relief And Lasso Feature Selection Approaches

Abstract

The studies’ primary aim is to help the research scholars as a source who would like to research in the thyroid disease detection region. UC Irvin knowledge discovery provides databases files for the machine learning archives' thyroid dataset. Here, a random vector network model (RVNM) is proposed to perform classification tasks. The proposed model integrates the prior dataset information regarding the samples to train the more effective classifier. This cascaded random vector network model helps in thyroid disease prediction. The evaluation process is performed to predict and determine the respective performance concerning accuracy. The intuition is provided in this research, like forecasting the thyroid disease; it also calls attention to the process of using a Randomized Vector Network Model (RVNM) as a medium for classification. The simulation is done in the MATLAB 2020a environment and establishes a better trade-off than various existing approaches. The model gives a prediction accuracy of 96.1% accuracy compared to other models and shows a better trade than others.