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

This dataset contains the generated features and targets used for training music genre classification systems in [1]. The data contains mean and standard deviation summaries of the sub-band filter-dependent windowing features extracted from the fault-filtered version of the GTZAN audio genre dataset [2]. Fault filtering specifications are described in [1]. 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] G. Tzanetakis and P. Cook, “Musical genre classification of audio signals,” IEEE Trans. Speech Audio Process., vol. 10, no. 5, pp. 293–302, Jul. 2002.