|
|
Music recommendation based on music information retrieval
Semela, René ; Schimmel, Jiří (referee) ; Kiska, Tomáš (advisor)
This thesis deals with the design, implementation and testing of the content-based music recommender system based on music information retrieval. In the introduction the attention is paid to issues of music information retrieval and to areas of their utilization, it also focuses on tools of their retrieving. Aferwards the most used types of recommender systems are described, including their typical problems. Options of the hybridization of these systems as well as examples of the popular music recommender systems are mentioned. There also is an outline of their functioning. The following section is focused on the parameterization of musical pieces and is devoted to the description of particular most used parameters. The next section is devoted to the content-based music recommender system design itself, including the defining of particular parameters that are used to differentiate musical recordings using the algorithm mRMR and other procedures. The recommender system design as such is oriented to the classification method k-nearest neighbors. The attention is also paid to the model of user taste recorded by Rocchio algorithm. In the next section the system is implemented according to the design. There is also described its functionality including the background processes. The final part of this work is focused on system testing and evaluation.
|
|
|
Automatic tagging of musical compositions using machine learning methods
Semela, René ; Galáž, Zoltán (referee) ; Kiska, Tomáš (advisor)
One of the many challenges of machine learning are systems for automatic tagging of music, the complexity of this issue in particular. These systems can be practically used in the content analysis of music or the sorting of music libraries. This thesis deals with the design, training, testing, and evaluation of artificial neural network architectures for automatic tagging of music. In the beginning, attention is paid to the setting of the theoretical foundation of this field. In the practical part of this thesis, 8 architectures of neural networks are designed (4 fully convolutional and 4 convolutional recurrent). These architectures are then trained using the MagnaTagATune Dataset and mel spectrogram. After training, these architectures are tested and evaluated. The best results are achieved by the four-layer convolutional recurrent neural network (CRNN4) with the ROC-AUC = 0.9046 ± 0.0016. As the next step of the practical part of this thesis, a completely new Last.fm Dataset 2020 is created. This dataset uses Last.fm and Spotify API for data acquisition and contains 100 tags and 122877 tracks. The most successful architectures are then trained, tested, and evaluated on this new dataset. The best results on this dataset are achieved by the six-layer fully convolutional neural network (FCNN6) with the ROC-AUC = 0.8590 ± 0.0011. Finally, a simple application is introduced as a concluding point of this thesis. This application is designed for testing individual neural network architectures on a user-inserted audio file. Overall results of this thesis are similar to other papers on the same topic, but this thesis brings several new findings and innovations. In terms of innovations, a significant reduction in the complexity of individual neural network architectures is achieved while maintaining similar results.
|
|
|
Automatic tagging of musical compositions using machine learning methods
Semela, René ; Galáž, Zoltán (referee) ; Kiska, Tomáš (advisor)
One of the many challenges of machine learning are systems for automatic tagging of music, the complexity of this issue in particular. These systems can be practically used in the content analysis of music or the sorting of music libraries. This thesis deals with the design, training, testing, and evaluation of artificial neural network architectures for automatic tagging of music. In the beginning, attention is paid to the setting of the theoretical foundation of this field. In the practical part of this thesis, 8 architectures of neural networks are designed (4 fully convolutional and 4 convolutional recurrent). These architectures are then trained using the MagnaTagATune Dataset and mel spectrogram. After training, these architectures are tested and evaluated. The best results are achieved by the four-layer convolutional recurrent neural network (CRNN4) with the ROC-AUC = 0.9046 ± 0.0016. As the next step of the practical part of this thesis, a completely new Last.fm Dataset 2020 is created. This dataset uses Last.fm and Spotify API for data acquisition and contains 100 tags and 122877 tracks. The most successful architectures are then trained, tested, and evaluated on this new dataset. The best results on this dataset are achieved by the six-layer fully convolutional neural network (FCNN6) with the ROC-AUC = 0.8590 ± 0.0011. Finally, a simple application is introduced as a concluding point of this thesis. This application is designed for testing individual neural network architectures on a user-inserted audio file. Overall results of this thesis are similar to other papers on the same topic, but this thesis brings several new findings and innovations. In terms of innovations, a significant reduction in the complexity of individual neural network architectures is achieved while maintaining similar results.
|
|
|
Music recommendation based on music information retrieval
Semela, René ; Schimmel, Jiří (referee) ; Kiska, Tomáš (advisor)
This thesis deals with the design, implementation and testing of the content-based music recommender system based on music information retrieval. In the introduction the attention is paid to issues of music information retrieval and to areas of their utilization, it also focuses on tools of their retrieving. Aferwards the most used types of recommender systems are described, including their typical problems. Options of the hybridization of these systems as well as examples of the popular music recommender systems are mentioned. There also is an outline of their functioning. The following section is focused on the parameterization of musical pieces and is devoted to the description of particular most used parameters. The next section is devoted to the content-based music recommender system design itself, including the defining of particular parameters that are used to differentiate musical recordings using the algorithm mRMR and other procedures. The recommender system design as such is oriented to the classification method k-nearest neighbors. The attention is also paid to the model of user taste recorded by Rocchio algorithm. In the next section the system is implemented according to the design. There is also described its functionality including the background processes. The final part of this work is focused on system testing and evaluation.
|
guest ::
