Study on Influencing Factors of Healing Music Records based on LightGBM

Abstract

At present, all kinds of music albums have become an indispensable product of spiritual consumption in public life. It can not only ease the pressure of life and work brought by external life, but also add some different interests to people's life. However, high-quality music is more popular with people, such as Dolby sound quality, and the demand for playing different types of high-quality music in different occasions and environments has greatly increased. This study will identify whether there is audience or artificial noise data during the recording of music albums, and explore the characteristics of various music albums, so that later music producers can better improve their music quality. To more specific, this study firstly collected the related dataset from Tianchi Platform. Then, data visualization e.g., correlation analysis is carried out to observe the data before passing it into the model. Subsequently, a typical machine learning algorithm called Light Gradient Boosting Machine is employed to train the dataset. The experimental results demonstrated the effectiveness of the model used in this study.