An Electronic Music Dataset and a Multilabel Neural Network for Genre Subcategory Classification Unavailable

In the context of genre classification, music subcategories are a special case. In this article, we focus on the problem of music subcategory classification in the genre of electronic music. We have assembled a publicly available electronic music dataset containing 13 subcategories and 3,047 audio tracks. For the collected dataset, we also propose a baseline classification method consisting of a multilabel neural network for feature processing and a random-forest-based classifier. The baseline method achieves an accuracy of 79%.