Characterizing Depressive Related Speech with MFCC
Conference proceedings article
Authors/Editors
Strategic Research Themes
No matching items found.
Publication Details
Author list: Suwannakhun S., Yingthawornsuk T.
Publisher: Hindawi
Publication year: 2019
ISBN: 9781728156316
ISSN: 0146-9428
eISSN: 1745-4557
Languages: English-Great Britain (EN-GB)
Abstract
The experimental results from comparative study of acoustical properties in speech as emotional indicator based on spectral characteristics of speech signal have formerly been studied and reported for its quantitative information in association with the emotional states in persons suffering depression. This symptom affects speech production system of speaker, which modulates in spoken sound. MFCC has been reported for its characteristic change corresponding to severity of depression. The sixteenth MFCCs from remitted, depressed and suicidal patient groups were extracted, statistically tested and classified in pairwise fashion by using ML, LS and LMS classifiers. The best score of classification can be obtained at 0.2487 in error based on ML classifier with 80% of MFCC samples in testing phase. Results suggest the dominant property of MFCC in separation between suicidal and recovering speakers from depression. ฉ 2019 IEEE.
Keywords
maximum likelihood
Contact Information