Institutional Repository of Vilnius University
Akustinių garso signalo požymių parinkimas muzikos žanro klasifikavimo uždaviniui spręsti /
Abstract
dc:descriptionMusic is an art form that has its own meaning, use, and is understood uniquely in different cultures. Since music reflects so many different emotions, ideas and ideals, it is quite varied. Because of this, music is separated into different genres. The main goal of this paper is to evaluate the impact of different acoustic features to the classification of music. To achieve this goal, the following tasks were being solved during the process of work: analyze which methods and acoustic features are most often used in music genre classification; complete the analysis of different datasets, which are used in music classification; complete the task of music genre classification task by using different acoustic features and classification algorithms; compare the differences and accuracies between the tested classification models and evaluate the obtained results. To solve the music genre classification task, k nearest member, random forest and convolutional neural network methods have been used. These methods were tested on a 58 feature vector, a 17 feature vector, spectrograms and mel spectrograms. The best results were obtained by using the random forest algorithm with the vector containing 58 features. The results of this algorithm were equal to 0,76 for precision score, 0,77 for recall value and 0,76 for the F1 score.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Tumalavičiūtė, Gabija,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD210647363&prefLang=en_US