RECORD
Reproducible Machine Learning-based Voice Pathology Detection: Introducing the Pitch Difference Feature
Sound · Submitted 14 Oct 2024
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim Conclusion: Our study demonstrates that by following the proposed methodology and feature engineering, there is a potential in detection of various voice pathologies using ML models applied to the simplest vocal task, a sustained utterance of the vowel /a:/.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
Purpose: We introduce a novel methodology for voice pathology detection using the publicly available Saarbr\"ucken Voice Database (SVD) and a robust feature set combining commonly used acoustic handcrafted features with two novel ones: pitch difference (relative variation in fundamental frequency) and NaN feature (failed fundamental frequency estimation). Methods: We evaluate six machine learning (ML) algorithms -- support vector machine, k-nearest neighbors, naive Bayes, decision tree, random forest, and AdaBoost -- using grid search for feasible hyperparameters and 20480 different feature subsets. Top 1000 classification models -- feature subset combinations for each ML algorithm are validated with repeated stratified cross-validation. To address class imbalance, we apply K-Means SMOTE to augment the training data. Results: Our approach achieves 85.61%, 84.69% and 85.22% unweighted average recall (UAR) for females, males and combined results respectively. We intentionally omit accuracy as it is a highly biased metric for imbalanced data. Conclusion: Our study demonstrates that by following the proposed methodology and feature engineering, there is a potential in detection of various voice pathologies using ML models applied to the simplest vocal task, a sustained utterance of the vowel /a:/. To enable easier use of our methodology and to support our claims, we provide a publicly available GitHub repository with DOI 10.5281/zenodo.13771573. Finally, we provide a REFORMS checklist to enhance readability, reproducibility and justification of our approach
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Purpose: We introduce a novel methodology for voice pathology detection using the publicly available Saarbr\"ucken Voice Database (SVD) and a robust feature set combining commonly used acoustic handcrafted features with two novel ones: pitch difference (relative variation in fundamental frequency) and NaN feature (failed fundamental frequency estimation). | contribution 1.00 | 0.89 | 0.03 | asserted 0.98 |
| s1 | Methods: We evaluate six machine learning (ML) algorithms -- support vector machine, k-nearest neighbors, naive Bayes, decision tree, random forest, and AdaBoost -- using grid search for feasible hyperparameters and 20480 different feature subsets. | method 1.00 | 0.96 | 0.01 | asserted 0.99 |
| s2 | Top 1000 classification models -- feature subset combinations for each ML algorithm are validated with repeated stratified cross-validation. | method 1.00 | 0.93 | 0.01 | asserted 0.95 |
| s3 | To address class imbalance, we apply K-Means SMOTE to augment the training data. | method 1.00 | 0.95 | 0.01 | asserted 0.98 |
| s4 | Results: Our approach achieves 85.61%, 84.69% and 85.22% unweighted average recall (UAR) for females, males and combined results respectively. | result 1.00 | 0.97 | 0.05 | asserted 1.00 |
| s5 | We intentionally omit accuracy as it is a highly biased metric for imbalanced data. | method 0.48 | 0.87 | 0.02 | asserted 1.00 |
| s6 | Conclusion: Our study demonstrates that by following the proposed methodology and feature engineering, there is a potential in detection of various voice pathologies using ML models applied to the simplest vocal task, a sustained utterance of the vowel /a:/. | interpretation 0.99 | 0.81 | 0.97 | asserted 0.62 |
| s7 | To enable easier use of our methodology and to support our claims, we provide a publicly available GitHub repository with DOI 10.5281/zenodo.13771573. | contribution 0.96 | 0.89 | 0.02 | asserted 0.97 |
| s8 | Finally, we provide a REFORMS checklist to enhance readability, reproducibility and justification of our approach | contribution 0.97 | 0.86 | 0.02 | asserted 0.90 |
What was examined, and what was not
- arXiv listing abstract v3
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
Purpose: We introduce a novel methodology for voice pathology detection using the publicly available Saarbr\"ucken Voice Database (SVD) and a robust feature set combining commonly used acoustic handcrafted features with two novel ones: pitch difference (relative variation in fundamental frequency) and NaN feature (failed fundamental frequency estimation). Methods: We evaluate six machine learning (ML) algorithms -- support vector machine, k-nearest neighbors, naive Bayes, decision tree, random forest, and AdaBoost -- using grid search for feasible hyperparameters and 20480 different feature subsets. Top 1000 classification models -- feature subset combinations for each ML algorithm are validated with repeated stratified cross-validation. To address class imbalance, we apply K-Means SMOTE to augment the training data. Results: Our approach achieves 85.61%, 84.69% and 85.22% unweighted average recall (UAR) for females, males and combined results respectively. We intentionally omit accuracy as it is a highly biased metric for imbalanced data. Conclusion: Our study demonstrates that by following the proposed methodology and feature engineering, there is a potential in detection of various voice pathologies using ML models applied to the simplest vocal task, a sustained utterance of the vowel /a:/. To enable easier use of our methodology and to support our claims, we provide a publicly available GitHub repository with DOI 10.5281/zenodo.13771573. Finally, we provide a REFORMS checklist to enhance readability, reproducibility and justification of our approach
Human review
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.