Because electronic music is so varied and individualized, categorizing it presents special difficulties. Traditional approaches to classifying electronic music are simple, based on labeling, and not keeping up with modern use and tastes.
Chinese academics at Chongqing Normal University have created a novel system for categorizing electronic music. The system’s goal is to assist streaming services and music libraries in handling the massive volume of digital information that is now accessible.
The researchers suggest the approach is based on a complex decision tree framework to achieve high accuracy and optimized processing times. Principal component analysis removes noise from the recording before it is divided into manageable pieces. The short-time Fourier transform is then used to extract the features from each piece. The team then refines their decision tree model to obtain the most accurate classification possible.
As demonstrated by their tests, which have a 98.6% classification accuracy, their method can be quite effective. The implications have potential applications in the music industry and other fields far beyond academics. Accurate genre classification is crucial for online music libraries and streaming services, which may benefit from this new strategy to arrange their collections and more discreetly promote songs to their listeners. Users may include regular music listeners or people working in the media or other creative fields who require a certain genre of music to go along with their work.
For example, the classification approach should facilitate music exploration, sound discovery, and retrieval for casual users. Understanding music preferences based on genre classification is essential for targeted efforts based on musical taste in marketing, advertising, and other fields.
Journal Reference:
- Hongyuan Wu; Lin Zhu. Adaptive classification method of electronic music based on improved decision tree. International Journal of Arts and Technology (IJART). DOI: 10.1504/IJART.2024.137296