Title: An efficient deep neural network model for music classification
Authors: Jagendra Singh
Addresses: School of Computer Science Engineering and Technology, Bennett University, Greater Noida-203206, India
Abstract: Combining with music recommendations that may be received through mobile devices, in comparison to past periods, the usage of digital music has risen in recent years. Searching for all this digital music may be terribly long and causes data exhaustion. Thus, it is helpful to build a music recommendation model which will retrieve appropriate music to users from available library of music automatically. The music recommendations service provider will supply or forecast acceptable songs to consumers depending on their song preferences based on the specified attributes of music or songs. In this study, we develop a music recommendation algorithm that assists users in automatically recommending music based on similarity of the available music. In this paper, we first extract the properties of the music using our suggested neural network model, which is then utilised to predict the rank of music as a music recommendation. The various evaluation parameters like precision, recall and F1 measure are used to evaluate our proposed system. The experimental results showed that both proposed recommendations models performed well for all six categories of songs but our proposed approach using both the feature of frequency in spectrogram and time sequence pattern performed better. Finally, t-test performed to show the statistical significance of our models.
Keywords: recommendation system; collaboration filtering; content-based filtering; music recommendation; machine learning; classification.
International Journal of Web Science, 2022 Vol.3 No.3, pp.236 - 248
Received: 22 May 2021
Accepted: 19 Feb 2022
Published online: 19 May 2022 *