Research on popularity of American pop singers songs based on machine learning

Abstract

The U.S. market of music is the largest market in the globe with its huge influence spreading around, which enables it to be the dominant of the world music industry. The article is produced due to the prevailing music market in U.S. that has phenomenally influence around the world. Therefore, this article takes Taylor Swift as an example thanks to the significant influence power to calculate whether some factors such as acousticness that might directly affect the popularity of singers songs for the purpose of formulating some market strategies by corresponding suggestions and advice. The research methods include 3 common mathematical model: linear regression, decision tree and random forest, which indicates that the year of release has the most contribution to the prediction with the value of importance of 0.549929. However other factors seem to have less relativity given the small values for the following 2 factors folklore and reputation with values of importance both below 0.2.