Standard audiogram classification from loudness scaling data using unsupervised, supervised, and explainable machine...

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).

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.

A large auditory reference database (n = 847 ears) containing ACALOS data was used for model development and evaluation.

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.

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.