The Method Of Dividing Music Community Based On Deep Learning

The network community is a platform for people to communicate.We can analyze user preferences by analyzing community information and sentiments.Sentimental analysis is a classification process of feature extraction and selection of texts containing sentimental factors.We can analyze users’ sentimental tendency towards a certain topic from the comments.The sentiments contained in the text are complex and changeable.For example,in the music community,the sentiments contained in people’s evaluation of music may contain multiple categories of sentiments at the same time.How to accurately extract the sentimental features of the text has become one of the hot issues at present.In order to accurately analyze the emotions contained in the evaluation information of the music community,this thesis proposes a convolutional neural network classification model based on multiple sentiment on the netease cloud music platform.Firstly,the method of emotional division of the music community is mostly based on the emotion of the audio or the lyrics itself,without considering the user’s intuitive feeling of the song.According to the user comments of the music community,this thesis proposes a multi-sentiment division method and a method of constructing emotion vectors.It improves the accuracy of sentiment feature extraction.Secondly,aiming at the problem that the existing methods can’t solve the feature extraction of sentiment words under similar sentences,this thesis proposes the method of sentimental splicing,which improves the accuracy of the model’s emotional feature extraction.Thirdly,aiming at the phenomenon of multiple sentiment coexistence in music commentary texts,this thesis proposes a method of emotional value based on music features,which can accurately classify and comment on emotions.Finally,In order to divide the music community,this thesis proposes a convolutional neural network classification model based on multiple sentiment combined with sentimental splicing and sentimental value.The experimental results show that the method in this thesis has a good performance in accuracy and could effectively and accurately solve the division problem of music community.