Web App Url: http://ec2-52-53-244-199.us-west-1.compute.amazonaws.com/web/
Abstract 3 - Music Analysis Web App - APPROVED
Traditional music players mostly focus on play music and music recommendation, but we want to give you more. Our music analysis web app combines a traditional music player with place recommendation using our unique analysis algorithm based on machine learning, more specifically, neural network.
UI: there are three parts of User Interface, one part is for music play, second part is for recommendation info bulletin and the last part is left blank for commercial ads or business fairs.
Backend: Using AWS EC2 container to hold our app online and S3 to store our data.
Algorithm: A neural network specifically trained for music analysis. Music files uploaded by users are fed to the network which output the type of the music and user's mood. The recommendation system will then suggest point of interest based on user’s mood and location
The model will be trained using pre-labelled music data of each type and mood and then stored in the cloud. When a user uploads a music file to our app, it will analyse it and determine the user’s mood. By combining user’s mood, current location, weather and time, the app will give related place suggestions based on our recommendation algorithm.
