M Zic
Listen what you love
With AI
01
Song recommendation
Moozic is a music recommendation app that uses AI to chat with users about their feelings and recommend or generate corresponding music that fits users' moods based on their responses.
- Tell us how you feel
- Tell us how you want to feel
- We recommend songs for you
02
Emotional Learning
Moozic train the recommendation algorithm by learning how you respond to different music genre.
As you input more ratings, we'll get better at showing the music that you'll love.
- AI play/you search a song
- AI tag the songs with emotions
- You rate/change the tag
Playlist
Moozic organizes all the music data trained into different playlists that tag with different emotions.
You can also remix new sog based on the emotional tag.
03
- Organize your playlist
- Pick the song you want
- Remix the song you love
Our Research
The effect of music on people
Music-based interventions may be beneficial for managing anxiety, depressive symptoms, and pain associated with various health conditions. [1]
​
Sad music allows listeners to disengage from distressing situations and instead focus on the beauty of the music. [2]
​
Music is also a potent trigger for aesthetic emotions like awe, wonder, transcendence, nostalgia, and tenderness. [3]
The individuality of people through their opinions on music
A global study indicates that regardless of cultural or geographical differences, certain types of music tend to appeal to specific personality types, suggesting a deep-rooted connection between music preferences and individual personality traits. [4]
​
The prevalence of specific emotions in music listening is influenced by factors such as individual characteristics, social context, and musical style.[5]
How AI could detect songs
​
AI can learn individual users' music preferences through their listening history, ratings, and behavior patterns. [6]
​
Once the AI understands individual preferences, it can group users based on similar tastes. This is typically done using clustering algorithms that find patterns and similarities among large sets of data. [7]
​
AI can use data from many users to make recommendations based on collaborative filtering. [8]
AI algorithm
We use machine learning to accurately learn your mood
Mood Coordinate System
Pick the coordinate that most accurately describes your feelings.
The bigger the coordinate number, the stronger you feel about the emotions
Replay the to make sure your feelings are right.
Machine Training Process
We trained the data from a wide range of users, categorizing your feelings to the group that has similar feelings as yours.
By repeating the process, we are making more and more subgroups and finding help you find the best subgroup you belong to, and then make a more accurate prediction of what you like based on your subgroup belongings.
Data Collection
Trained 200 songs, and each was evaluated by at least 30 people.
We scored on a valence arousal rating: Arousal scores and plotted them onto a graph by the median.
Then we utilized KNN models to determine target groups for the recommendation system and simply adjusted the sizing and distribution depending on user input.
See how it works
Watch the video demo
NOTICE: This prototype is intended only for testing interaction and visual elements. It does not feature any sound.
