Standard audiogram classification from loudness scaling data using unsupervised, supervised, and explainable machine learning techniques
Standard audiogram classification from loudness scaling data using unsupervised, supervised, and explainable machine learning techniques
Objective: to address the calibration and procedural challenges inherent in remote audiogram assessment for rehabilitative audiology, this study investigated whether calibration-independent adaptive categorical loudness scaling (ACALOS) data can be used to approximate individual audiograms by classifying listeners into standard Bisgaard audiogram types using machine learning (ML).
Design: three classes of ML approaches—unsupervised, supervised, and explainable—were evaluated. Principal component analysis (PCA) was performed to extract the first two principal components, which jointly explained more than 50% of the variance. Seven supervised multi-class ML classifiers were trained and compared, alongside unsupervised and explainable methods.
Study sample: a large auditory reference database (n = 847 ears) containing ACALOS data was used for model development and evaluation.
Results: the factor map showed substantial overlap between listeners, indicating that cleanly separating participants into six Bisgaard classes based solely on their loudness patterns is challenging. Nevertheless, the ML models demonstrated reasonable classification performance. Among the supervised classifiers, logistic regression achieved the highest accuracy. In addition, the SHAP and feature permutation analyses showed that the highest predictive power of the ML models was attributable to the minimum loudness levels at 1.5 and 4 kHz.
Conclusions: the findings demonstrate that ML models can predict standard Bisgaard audiogram types—within certain limits—from calibration-independent loudness perception data. This approach may support future hearing aid fitting in remote or resource-limited settings without requiring a traditional audiogram.
Bisgaard profiles prediction, Explainable machine learning, big data, loudness scaling test
Xu, Chen
73268368-81b7-46b9-b752-5d0392977212
Schell-Majoor, Lena
78187eff-316c-4cfc-ba26-2a4f2a0cf893
Kollmeier, Birger
6de18374-5a52-4ca2-8d1f-08b32ca43942
Xu, Chen
73268368-81b7-46b9-b752-5d0392977212
Schell-Majoor, Lena
78187eff-316c-4cfc-ba26-2a4f2a0cf893
Kollmeier, Birger
6de18374-5a52-4ca2-8d1f-08b32ca43942
Xu, Chen, Schell-Majoor, Lena and Kollmeier, Birger
(2026)
Standard audiogram classification from loudness scaling data using unsupervised, supervised, and explainable machine learning techniques.
International Journal of Audiology, 0 (0).
(doi:10.1080/14992027.2026.2642765).
Abstract
Objective: to address the calibration and procedural challenges inherent in remote audiogram assessment for rehabilitative audiology, this study investigated whether calibration-independent adaptive categorical loudness scaling (ACALOS) data can be used to approximate individual audiograms by classifying listeners into standard Bisgaard audiogram types using machine learning (ML).
Design: three classes of ML approaches—unsupervised, supervised, and explainable—were evaluated. Principal component analysis (PCA) was performed to extract the first two principal components, which jointly explained more than 50% of the variance. Seven supervised multi-class ML classifiers were trained and compared, alongside unsupervised and explainable methods.
Study sample: a large auditory reference database (n = 847 ears) containing ACALOS data was used for model development and evaluation.
Results: the factor map showed substantial overlap between listeners, indicating that cleanly separating participants into six Bisgaard classes based solely on their loudness patterns is challenging. Nevertheless, the ML models demonstrated reasonable classification performance. Among the supervised classifiers, logistic regression achieved the highest accuracy. In addition, the SHAP and feature permutation analyses showed that the highest predictive power of the ML models was attributable to the minimum loudness levels at 1.5 and 4 kHz.
Conclusions: the findings demonstrate that ML models can predict standard Bisgaard audiogram types—within certain limits—from calibration-independent loudness perception data. This approach may support future hearing aid fitting in remote or resource-limited settings without requiring a traditional audiogram.
Text
Standard audiogram classification from loudness scaling data using unsupervised supervised and explainable machine learning techniques
- Version of Record
More information
Accepted/In Press date: 27 February 2026
e-pub ahead of print date: 1 April 2026
Additional Information:
PMID: 41921539
Keywords:
Bisgaard profiles prediction, Explainable machine learning, big data, loudness scaling test
Identifiers
Local EPrints ID: 511515
URI: http://eprints.soton.ac.uk/id/eprint/511515
ISSN: 1499-2027
PURE UUID: 426c432b-911f-4713-9330-e24f46bf1268
Catalogue record
Date deposited: 18 May 2026 16:51
Last modified: 19 May 2026 02:13
Export record
Altmetrics
Contributors
Author:
Chen Xu
Author:
Lena Schell-Majoor
Author:
Birger Kollmeier
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics