A Novel Biased Probability Neural Network (BPNN) and Regularized Extreme Learning Machine (RELM) based Hearing Loss Prediction System

• Published in: Iraqi Journal for Computer Science and Mathematics 2023-02, Vol.4 (2)
• Main Authors: Selvaganesh N, Shanthi D, Rajesh Pandian N
• Format: Article
• Language: English
• Subjects: Hearing Impairment, Acoustic Stimulated Signals, Age Prediction, Biased Probability Neural Network (BPNN), Regularized Extreme Learning Machine (RELM), and Cross-fold validation
• ISSN: 2958-0544; 2788-7421
• DOI: 10.52866/ijcsm.2023.02.02.005

: Hearing loss or hearing impairment is one of the most leading cause highly affecting the people around the world in present days. Also, the youngsters and adults are mostly affecting by this disease, which affects their normal/social life, carrier, education, and etc. Hence, it should be properly identified and diagnosed for providing an earlier treatments to save the people live. For this purpose, the different types of automated hearing loss prediction/detection systems are developed in the conventional works. The existing works are mainly focused on deploying the machine learning/deep learning based prediction approaches for disease identification. However, it lacks with the flaws of difficult computational steps, more training & testing time, increased mis-prediction results, and error outputs. Therefore, the proposed work objects to develop a Human Age - Hearing Impairment & Level (HAHIL) prediction system by using the machine learning methodologies. Here, three distinct prediction models are deployed for age prediction, hearing loss detection, and its severity level prediction. The Biased Probability Neural Network (BPNN) technique is utilized to predict the age based on simulated human acoustical signals. Then, the Regularized Extreme Learning Machine (RELM) mechanism is deployed for predicting the hearing impairment by constructing the weight and target matrices. During evaluation, the performance of the proposed HAHIL prediction system is validated and tested by using various evaluation indicators.