CLASSIFICATION OF NOISE INDUCED HEARING LOSS USING BAYESIAN OPTIMIZATION SUPPORT VECTOR MACHINE

Abstract

ABSTRACT - Noise Induced Hearing Loss (NIHL) is an occupational disease is becoming leading
cause of Occupational Noise Related Hearing Disorder (ONRHD) among Malaysian workers. To lower
the number of reported NIHL cases, it is important to detect NIHL at an early stage. The primary goal of
this study is to identify the key factors contributing to diagnosis of NIHL while developing an enhanced
Support Vector Machine (SVM) prediction model using Bayesian Optimization to produce high precision
classification results. The study examined 355 secondary datasets consisting of 24 variables divided
into four segments which are 100, 200, 300 and 355.This data were analyzed to determine the best
percentage of the sample size to be classified. Using the MATLAB application, the Support Vector
Machine was used to classify the data based on the five prediction outputs based on the severity level
of the NIHL. Using the Bayesian Optimization SVM using linear kernel the training accuracy exceeded
85.00%. This improved prediction model using kernel gaussian was recorded outperformed and
achieved 100% prediction accuracy during testing. Because this improved prediction model produced
high accuracy prediction results, it has the potential to reduce misinterpretations in ingratiating cases,
thereby lowering the number of unconfirmed NIHL cases reported.