Spectral Sub-Band Filter Dependent Windowing Music Mood Classification Machine Learning Dataset

Published: 28 July 2022| Version 2 | DOI: 10.17632/n6fnpc8jm6.2
Fabi Prezja


This dataset contains the generated features and targets used for training music mood classification systems in [1]. The data contains mean and standard deviation summaries of the sub-band filter-dependent windowing features extracted from the ‘Mirex-Like Mood' - 'PandaMood’ Dataset [2]. Data is in CSV format; feature names contain the following suffixes: f (1-10), mean or std. The 'f' suffix specifies the sub-band filter index while 'mean' or 'std' is the statistical summary. Please refer to the source [1] for the extraction specification details. References: [1] F. Prezja, “Developing and testing sub-band spectral features in music genre and music mood machine learning,” Master Thesis, University of Jyväskylä, Jyväskylä, November. 2018. [Online]. Available: https://jyx.jyu.fi/bitstream/handle/123456789/60963/1/URN%3ANBN%3Afi%3Ajyu-201901081104.pdf [2] R. Panda, R. Malheiro, B. Rocha, A. P. Oliveira, and R. P. Paiva, “Multi-Modal Music Emotion Recognition: A New Dataset, Methodology and Comparative Analysis,” 10th Int. Symp. Comput. Music Multidiscip. Res. – C. 2013, pp. 570–582, 2013.


Steps to reproduce

All feature extraction specification details can be found in [1].


Jyvaskylan Yliopisto


Music Computing, Artificial Intelligence in Music, Machine Learning, Supervised Learning, Spectral Analysis of Signal