SVM-based spectral analysis and classification of auditory brainstem response signals using fast Continuous Wavelet Transform

Authors

Nicolo John B. David Department of Physical Sciences and Mathematics, University of the Philippines Manila
Herbert B. Domingo Department of Physical Sciences and Mathematics, University of the Philippines Manila
Catherine Manuela Lee Ramos Manufacturing Engineering and Management, De La Salle University, Manila

Abstract

This study classifies hearing-impaired patients from normal patients using fast continuous wavelet transform (fCWT) to extract auditory brainstem response (ABR) signal spectra. The fCWT-based spectral analysis was compared to the traditional fast Fourier transform (FFT) in a support vector machine (SVM) classifier. ABR signals from both hearing-impaired and normal listeners were pre-processed and analyzed to extract the top 20% dominant frequencies using fCWT and FFT. These frequencies were then used as features for the SVM classifier. Performance metrics, including accuracy, sensitivity, and specificity, were evaluated. Results indicated that the fCWT-based SVM model achieved higher accuracy (96.76%), sensitivity (100%), and specificity (95.96%) compared to the FFT-based model. This demonstrates fCWT’s superior capability in capturing relevant frequencies for classification, making it a more effective tool for non-stationary signal analysis and rapid hearing screening.

Issue

2024: Proceedings of the 42nd Samahang Pisika ng Pilipinas Physics Conference

Article ID

SPP-2024-PB-19

Section

Poster Session B (Complex Systems, Computational Physics, and Astrophysics)

Published

2024-06-28

How to Cite

[1]

NJB David, HB Domingo, and CM Lee Ramos, SVM-based spectral analysis and classification of auditory brainstem response signals using fast Continuous Wavelet Transform, Proceedings of the Samahang Pisika ng Pilipinas 42, SPP-2024-PB-19 (2024). URL: https://proceedings.spp-online.org/article/view/SPP-2024-PB-19.