A comparative study of different classifiers for detecting depression from spontaneous speech

Sharifa Alghowinem, Roland Goecke, Michael Wagner, Julien Epps, Tom Gedeon, Michael Breakspear, Gordon Parker

Research output: A Conference proceeding or a Chapter in BookConference contribution

27 Citations (Scopus)

Abstract

Accurate detection of depression from spontaneous speech could lead to an objective diagnostic aid to assist clinicians to better diagnose depression. Little thought has been given so far to which classifier performs best for this task. In this study, using a 60-subject real-world clinically validated dataset, we compare three popular classifiers from the affective computing literature – Gaussian Mixture Models (GMM), Support Vector Machines (SVM) and Multilayer Perceptron neural networks (MLP) – as well as the recently proposed Hierarchical Fuzzy Signature (HFS) classifier. Among these, a hybrid classifier using GMM models and SVM gave the best overall classification results. Comparing feature, score, and decision fusion, score fusion performed better for GMM, HFS and MLP, while decision fusion worked best for SVM (both for raw data and GMM models). Feature fusion performed worse than other fusion methods in this study. We found that loudness, root mean square, and intensity were the voice features that performed best to detect depression in this dataset.
Original languageEnglish
Title of host publicationICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
EditorsRabab Ward, Li Deng
Place of PublicationVancouver, British Columbia, Canada
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages8022-8026
Number of pages5
ISBN (Electronic)9781479903566
DOIs
Publication statusPublished - 2013
Event38th International Conference on Acoustics, Speech and Signal Processing ICASSP2013 - Vancouver, Vancouver, Canada
Duration: 26 May 201331 May 2013

Conference

Conference38th International Conference on Acoustics, Speech and Signal Processing ICASSP2013
Abbreviated titleICASSP2013
CountryCanada
CityVancouver
Period26/05/1331/05/13
OtherThe ICASSP meeting is the world's largest and most comprehensive technical conference focused on signal processing and its applications. The conference will feature world-class speakers, tutorials, exhibits, and over 50 lecture and poster sessions

Fingerprint

Classifiers
Fusion reactions
Support vector machines
Multilayer neural networks
Neural networks

Cite this

Alghowinem, S., Goecke, R., Wagner, M., Epps, J., Gedeon, T., Breakspear, M., & Parker, G. (2013). A comparative study of different classifiers for detecting depression from spontaneous speech. In R. Ward, & L. Deng (Eds.), ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings (pp. 8022-8026). Vancouver, British Columbia, Canada: IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ICASSP.2013.6639227
Alghowinem, Sharifa ; Goecke, Roland ; Wagner, Michael ; Epps, Julien ; Gedeon, Tom ; Breakspear, Michael ; Parker, Gordon. / A comparative study of different classifiers for detecting depression from spontaneous speech. ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. editor / Rabab Ward ; Li Deng. Vancouver, British Columbia, Canada : IEEE, Institute of Electrical and Electronics Engineers, 2013. pp. 8022-8026
@inproceedings{1eb61a2b384d4f71ad67bea9db00d7ab,
title = "A comparative study of different classifiers for detecting depression from spontaneous speech",
abstract = "Accurate detection of depression from spontaneous speech could lead to an objective diagnostic aid to assist clinicians to better diagnose depression. Little thought has been given so far to which classifier performs best for this task. In this study, using a 60-subject real-world clinically validated dataset, we compare three popular classifiers from the affective computing literature – Gaussian Mixture Models (GMM), Support Vector Machines (SVM) and Multilayer Perceptron neural networks (MLP) – as well as the recently proposed Hierarchical Fuzzy Signature (HFS) classifier. Among these, a hybrid classifier using GMM models and SVM gave the best overall classification results. Comparing feature, score, and decision fusion, score fusion performed better for GMM, HFS and MLP, while decision fusion worked best for SVM (both for raw data and GMM models). Feature fusion performed worse than other fusion methods in this study. We found that loudness, root mean square, and intensity were the voice features that performed best to detect depression in this dataset.",
keywords = "Mood detection, affective sensing, classifier comparison, Depression",
author = "Sharifa Alghowinem and Roland Goecke and Michael Wagner and Julien Epps and Tom Gedeon and Michael Breakspear and Gordon Parker",
year = "2013",
doi = "10.1109/ICASSP.2013.6639227",
language = "English",
pages = "8022--8026",
editor = "Rabab Ward and Li Deng",
booktitle = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
address = "United States",

}

Alghowinem, S, Goecke, R, Wagner, M, Epps, J, Gedeon, T, Breakspear, M & Parker, G 2013, A comparative study of different classifiers for detecting depression from spontaneous speech. in R Ward & L Deng (eds), ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. IEEE, Institute of Electrical and Electronics Engineers, Vancouver, British Columbia, Canada, pp. 8022-8026, 38th International Conference on Acoustics, Speech and Signal Processing ICASSP2013, Vancouver, Canada, 26/05/13. https://doi.org/10.1109/ICASSP.2013.6639227

A comparative study of different classifiers for detecting depression from spontaneous speech. / Alghowinem, Sharifa; Goecke, Roland; Wagner, Michael; Epps, Julien; Gedeon, Tom; Breakspear, Michael; Parker, Gordon.

ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. ed. / Rabab Ward; Li Deng. Vancouver, British Columbia, Canada : IEEE, Institute of Electrical and Electronics Engineers, 2013. p. 8022-8026.

Research output: A Conference proceeding or a Chapter in BookConference contribution

TY - GEN

T1 - A comparative study of different classifiers for detecting depression from spontaneous speech

AU - Alghowinem, Sharifa

AU - Goecke, Roland

AU - Wagner, Michael

AU - Epps, Julien

AU - Gedeon, Tom

AU - Breakspear, Michael

AU - Parker, Gordon

PY - 2013

Y1 - 2013

N2 - Accurate detection of depression from spontaneous speech could lead to an objective diagnostic aid to assist clinicians to better diagnose depression. Little thought has been given so far to which classifier performs best for this task. In this study, using a 60-subject real-world clinically validated dataset, we compare three popular classifiers from the affective computing literature – Gaussian Mixture Models (GMM), Support Vector Machines (SVM) and Multilayer Perceptron neural networks (MLP) – as well as the recently proposed Hierarchical Fuzzy Signature (HFS) classifier. Among these, a hybrid classifier using GMM models and SVM gave the best overall classification results. Comparing feature, score, and decision fusion, score fusion performed better for GMM, HFS and MLP, while decision fusion worked best for SVM (both for raw data and GMM models). Feature fusion performed worse than other fusion methods in this study. We found that loudness, root mean square, and intensity were the voice features that performed best to detect depression in this dataset.

AB - Accurate detection of depression from spontaneous speech could lead to an objective diagnostic aid to assist clinicians to better diagnose depression. Little thought has been given so far to which classifier performs best for this task. In this study, using a 60-subject real-world clinically validated dataset, we compare three popular classifiers from the affective computing literature – Gaussian Mixture Models (GMM), Support Vector Machines (SVM) and Multilayer Perceptron neural networks (MLP) – as well as the recently proposed Hierarchical Fuzzy Signature (HFS) classifier. Among these, a hybrid classifier using GMM models and SVM gave the best overall classification results. Comparing feature, score, and decision fusion, score fusion performed better for GMM, HFS and MLP, while decision fusion worked best for SVM (both for raw data and GMM models). Feature fusion performed worse than other fusion methods in this study. We found that loudness, root mean square, and intensity were the voice features that performed best to detect depression in this dataset.

KW - Mood detection

KW - affective sensing

KW - classifier comparison

KW - Depression

U2 - 10.1109/ICASSP.2013.6639227

DO - 10.1109/ICASSP.2013.6639227

M3 - Conference contribution

SP - 8022

EP - 8026

BT - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

A2 - Ward, Rabab

A2 - Deng, Li

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - Vancouver, British Columbia, Canada

ER -

Alghowinem S, Goecke R, Wagner M, Epps J, Gedeon T, Breakspear M et al. A comparative study of different classifiers for detecting depression from spontaneous speech. In Ward R, Deng L, editors, ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. Vancouver, British Columbia, Canada: IEEE, Institute of Electrical and Electronics Engineers. 2013. p. 8022-8026 https://doi.org/10.1109/ICASSP.2013.6639227