Font Size: a A A

Analysis Of Music Reviews Based On Topic Mining

Posted on:2021-05-24Degree:MasterType:Thesis
Country:ChinaCandidate:S Y GuoFull Text:PDF
GTID:2415330605957334Subject:Applied Statistics
Abstract/Summary:PDF Full Text Request
As the popularity and development of the Internet,more and more network applications are developing rapidly.At the same time,people's living condi-tion continues to improve.Pure material products have been difficult to sat-isfy consumers,so emotional marketing came into being.Music,one of the most expressive means to express people's feeling,has develope drapidly on the Internet which named online music.Simultaneously,various new songs have emerged endlessly,and the public's evaluation of songs directly affects our choice of songs.In order to give a reasonable social evaluation of a song,this article conducts a series of studies on a series of text data of music reviews.Topic mining is the process of discovering topics by using associations between text features in the corpus.It can discover the underlying semantic information hidden in the vocabulary of the text.Therefore,this paper retrieves and preprocesses the review data from the music application,then does the topic mining based on the LDA model.Different from the traditional theme analysis,this article builds a theme social network which based on the theme mining to analyze the derived re-lationships between themes.On the other hand,it uses topic text clustering to make various topic word cloud diagrams to optimize the results of theme model.Through the analysis of the theme mining,the public's evaluation of music is obtained,which closer to the overall social trend,and avoids the impact of per-sonal evaluation.Furthermore,the results of text topic mining can also be used for music recommendations or recommended tags,which make the promotion more efficient.Applying this research process to two types of music,which called pure music and song,such as "Town of Windmill" and "I Want to Love This World",we can get some conclusions:First,the pure music "Town of Windmill" should be widely used on various media platforms,such as games,animations,and even the soundtrack of some videos or movies.The style should be purer.Based on the different emotions of the audience's evaluation,it can be inferred that its arrangement should be very layered.Second,the song "I really want to love this world" may be created for an event.In fact,it was written for the depression group.According to the different themes of the review text,we can know that most of the listeners got hope and happiness from this song.There is not difficult to find that,It's a timely and powerful song to depression patients.
Keywords/Search Tags:Text mining, Topic mining, LDA model, Topic clustering, Music evaluation
PDF Full Text Request
Related items