Electromyography Signal Analysis for the Detection of TMJ Disorder Using Classification Models and Multivariate Analysis

Abstract

Temporomandibular joint (TMJ) disorder is a wide term that encompasses a variety of disorders with varying etiologies. The purpose of this study was to analyses electromyographic signals with wavelet transform for the diagnostic methods which help in TMJ issues in patients who visited the Dental Centre. Due to the increasing importance of electromyography signals in diagnosing muscular disorders, such as temporomandibular joint disorder, it has been widely used. Through various techniques, such as discrete wavelet transform and power spectral density, it is possible to identify and minimize the noise in the signals, which can be very useful in the diagnosis of the disorder. This paper presents an algorithm that combines the features of discrete wavelet transform and multivariate analysis in order to detect temporomandibular joint disorder. Support vector machine model is giving the better performance in terms of training, testing time and accuracy with 93% compared to other models. Multivariate analysis shows the significant difference in the feature variable chosen.