Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Analysis of voice recordings features for Classification of Parkinson's Disease

Beatriz Pérez-Sánchez, Noelia Sánchez-Maroño, Miguel A. Díaz-Freire

TL;DR

This paper tackles early Parkinson's disease diagnosis from voice recordings by leveraging feature selection to cope with a very high-dimensional feature set. It benchmarks SVM and artificial neural networks, integrating multiple feature-selection methods and a stable 30–50 feature subset via a two-stage union flow, with a validation scheme that preserves patient-exclusive splits. The best ANN configurations achieve MCC around 0.97 using about 44–50 features, with MFCC and TQWT features driving the performance, and SVM underperforms in comparison. The work demonstrates that strong diagnostic performance is achievable with compact, explainable feature sets, offering a non-invasive, cost-effective tool for PD screening and guiding clinical decision-making.

Abstract

Parkinson's disease (PD) is a chronic neurodegenerative disease. Early diagnosis is essential to mitigate the progressive deterioration of patients' quality of life. The most characteristic motor symptoms are very mild in the early stages, making diagnosis difficult. Recent studies have shown that the use of patient voice recordings can aid in early diagnosis. Although the analysis of such recordings is costly from a clinical point of view, advances in machine learning techniques are making the processing of this type of data increasingly accurate and efficient. Vocal recordings contain many features, but it is not known whether all of them are relevant for diagnosing the disease. This paper proposes the use of different types of machine learning models combined with feature selection methods to detect the disease. The selection techniques allow to reduce the number of features used by the classifiers by determining which ones provide the most information about the problem. The results show that machine learning methods, in particular neural networks, are suitable for PD classification and that the number of features can be significantly reduced without affecting the performance of the models.

Analysis of voice recordings features for Classification of Parkinson's Disease

TL;DR

This paper tackles early Parkinson's disease diagnosis from voice recordings by leveraging feature selection to cope with a very high-dimensional feature set. It benchmarks SVM and artificial neural networks, integrating multiple feature-selection methods and a stable 30–50 feature subset via a two-stage union flow, with a validation scheme that preserves patient-exclusive splits. The best ANN configurations achieve MCC around 0.97 using about 44–50 features, with MFCC and TQWT features driving the performance, and SVM underperforms in comparison. The work demonstrates that strong diagnostic performance is achievable with compact, explainable feature sets, offering a non-invasive, cost-effective tool for PD screening and guiding clinical decision-making.

Abstract

Parkinson's disease (PD) is a chronic neurodegenerative disease. Early diagnosis is essential to mitigate the progressive deterioration of patients' quality of life. The most characteristic motor symptoms are very mild in the early stages, making diagnosis difficult. Recent studies have shown that the use of patient voice recordings can aid in early diagnosis. Although the analysis of such recordings is costly from a clinical point of view, advances in machine learning techniques are making the processing of this type of data increasingly accurate and efficient. Vocal recordings contain many features, but it is not known whether all of them are relevant for diagnosing the disease. This paper proposes the use of different types of machine learning models combined with feature selection methods to detect the disease. The selection techniques allow to reduce the number of features used by the classifiers by determining which ones provide the most information about the problem. The results show that machine learning methods, in particular neural networks, are suitable for PD classification and that the number of features can be significantly reduced without affecting the performance of the models.
Paper Structure (14 sections, 4 equations, 3 figures, 5 tables)

This paper contains 14 sections, 4 equations, 3 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Characteristics of voice recordings
  3. State of the art
  4. Methodology
  5. Dataset
  6. Dimensionality reduction
  7. Classification models
  8. Experiment flow
  9. Validation
  10. Performance metrics
  11. Results and discussion
  12. ANN performance results
  13. Analysis of selected characteristics for ANN
  14. Conclusions

Figures (3)

  • Figure 1: Experiment flow.
  • Figure 2: Distribution of selected features in each group in the top 15 ANN models.
  • Figure 3: Distribution of selected features for the best FS algorithms.