| dc.description.abstract | Music is everywhere in our life. No matter what you do, you might pick up your mobile device or computer and turn on the music to enjoy in it. Because of the demand for music, there are many kinds of music streaming service appeared, like Spotify, Apple Music, Kkbox and YouTube. According to the 2018 Music Consumer Insight Report published by IFPI, on average consumers spend 17.8hrs listening to music each week globally, 52% of music streaming is on video streaming, and 47% of time spend on listening to on-demand music is on YouTube. Thus, we can conclude that YouTube is a popular and important music streaming platform.
Many researches about YouTube is to realize the user behavior on YouTube and why people participate on the platform. Some researches about the popularity of the videos on YouTube. Some researches are about the recommendation system. As far as I know within my ability range, there is no research about the relationship between the sentiment on comments expressed by listeners and the listener behavior on YouTube. Thus, the research we do is to know that whether the amounts of like will be more when the sentiment of the music and the sentiment of comments are all more positive or are all more negative based on the following three demonstrations : Comments can be a way to express our sentiments or opinions, Like is a kind of way to express our sentiments and is also call as virtual empathy that we have same feeling to others.
We use sentiment analysis to get the sentiments of lyrics and comments, and according to the sentiments of them to divide all data in four kinds of categories. Then, we respectively use regression analysis on four kinds of categories. In the end, the result of this research is that if the sentiments of lyrics and comments are all more positive or they are all more negative, the amounts of like will be more.
| en_US |