Regression Analysis of Song Popularity based on Ridge, K-Nearest Neighbors and Multiple-Layers Neural Networks

Abstract

Contemporarily, human beings are working on the implementation of artificial intelligence technology in the arts fields, where the music is one of the directions. Before humans can create a song with artificial intelligence, it is necessary to understand the song first. This research tries to find out the relationship between the song's popularity and several selected songs' physical parameters based on statistics and machine learning. According to the analysis, this research proves that there is no significant relationship between selected physical parameters and the song's popularity. In addition, machine learning algorithms also do not find the potential relationships between them. In this case, it is safe to conclude that creating the song by considering these selected physical parameters is meaningless. On this basis, scholars should try to find out what factors make the song popular in terms of analyzing songs differently. These results shed light on guiding further exploration of future music analysis and artificial intelligence in music fields.