- Music application has been reported as the longest time used in all user's entertainment application.
- A large Increase needs of recommender systems because it helps user to finding their own preferences from a huge data over the past few years.
- The current trend of users choosing digital format over physical format is likely to grow.
- Identify user's preference in representation of music features and recommend music content that may be relevant to that user.
- A music recommendation system can also help in promoting more new artists who might be unfamiliar to the users of recommendation system.
Figure 1. Process of music recommendation application
| Variable | Value | Description |
|---|---|---|
| Rating | 1-100 | In figure 2, The rating table describe the song recommendation value that given by another user who has the highest similarity and it representative the user input preference. In figure 3, The rating table describe the song recommendation value was come from average of all user ratings to that song and it is a collaborative recommendation for all users. |
| Songs | 6,035 | This dataset contains 6,035 songs that used to make this music recommendation and stored in database |
| Users | 20,000 | 20,000 users stored in database and used to this recommendation application |
Figure 2. Example of Recommendation List for the User
Figure 3. List of songs sort by the highest rating
This application is a recommendation application for music based on the principle of collaborative. So, this application provides user recommendation based on the similarity of songs played by users to all users on this system. This recommendation is also based on the number of times a song is played by a user. In this system, there are 20,000 users, 6,035 songs and there are about 1.2 million data that describe the number of times a user has played a song. This data comes from public datasets that we adjust to the needs.
This work is supported by the Directorate General of Strengthening for Research and Development, Ministry of Research, Technology, and Higher Education, Republic of Indonesia, as a part of Penelitian Terapan Unggulan Perguruan Tinggi Research Grant to Binus University titled “Pengembangan Sistem Rekomendasi Lagu Menggunakan Neural Collaborative Filtering” or “Development of Song Recommendation System Using Neural Collaborative Filtering” 2019.
By clicking the "Demo" button, you can try this system.
User ID
= user* (* = 2 until 20000, for example user11)
Password = 'P'