Institutional Repository of Vilnius University
Classification of speech signal using functional data /
Abstract
dc:descriptionThe objective of this study is to classify Lithuanian words recorded in audio files by predicting the speaker's gender. Initially, the Hilbert transform was applied to the speech signals. Subsequently, after finding the optimal parameters, the smoothing of the speech signals was performed. Finally, the classification was done by using three classifiers: K-Nearest Neighbor, Support Vector Machine and Random Forest. All classifiers were applied to both functional and multivariate data after utilizing Functional Data Analysis. Evaluation of the results revealed that the Random Forest classifier for multivariate data was the most effective, achieving an accuracy of 82.60 % in predicting the speaker's gender.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Vengalienė, Judita,
Subjects
dc:subject × 1Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- eng
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD192057680&prefLang=en_US