| With the advent of the Internet era,a variety of online communities have flourished,and the Internet has gradually transformed from a static medium for information acquisition to a dynamic platform for information sharing and interaction.For example,when listening to music,users often resonate with music and other users' comments,and spontaneously publish comments to interact with other users.Thus,an online music community based on emotional communication is established.The music platform is no longer a pure music provider,but is increasingly moving towards the online community.There are online music communities such as Netease Cloud Music and Xiami Music,which are user-centric and socially friendly.So if you want to enhance the competitiveness of the online music community,attract a large number of users and increase user stickiness,we can start from enhancing their social attributes.At present,how to enhance the social attributes of the online music community faces two problems that need to be solved.One is how to analyze the differences in the user's discourse patterns,and then put forward opinions and suggestions that can meet the needs of individual users,thereby improving the inclusiveness of the online community;Another point is how to classify and segment the emotions of the online music community to find out how to grasp the most basic emotional tone,emotional distribution and how to enhance the emotional atmosphere of the online music community.Currently,in this online music community based on emotional resonance,the amount of data generated by users has exploded.The increase in user reviews has provided very favorable conditions for research.On the other hand,the comments of the online music community are mostly presented in short text form.A large part of these massive text messages is information that expresses users' opinions and emotions,such as support,opposition,joy,anger,sadness,and joy.And other text information.These emotional text messages are invaluable resources for opinions,including people's different opinions and emotional attitudes.Related methods of sentiment analysis are currently widely used in online social networking sites such as Weibo and Twitter,but there are fewer studies in the online music community.Based on this,there are three main research points in this paper.The first point is to explore the discourse patterns reflected by the user groups of different ages,genders,and regions in the online music community through the Chinese word segmentation and part-of-speech tagging operations,and the slice operations of different dimensions of the comment data.The second point is to take the Netease cloud music platform and the Xiami music platform as examples to study the most frequently used word frequency and part of speech in different music communities,to study the star-hunting atmosphere of different platforms,the enthusiasm for lyrics,the user's activity and the distribution of active time,etc.The third point is to establish a machine learning model to study the emotional classification of different online music community comment texts,and to study the emotional atmosphere of different online music communities from the emotional perspective of the text.In view of the first point of the research,this paper draws from the research results of the word frequency and part of speech of each user group that the phenomenon of discourse patterns in the real society extends to the online music community environment.Specifically,the age discourse model is mainly reflected in different words,different concerns,and different emotional expressions caused by different ages;the gender discourse model is mainly reflected in the phenomenon of gender-induced word limitation and related to gender attributes.Differences in word use caused by different genders;the regional discourse model mainly reflects the influence of regional cultural atmosphere on commentary.In addition,the nuances of the proportions of various parts of speech in each user group's comments on various types of songs.The above work extends the research on behavioral theory and discourse model theory on the online social network platform,which helps the online music platform to understand more dimensions of the user,so that the user can be targeted and recommended,and the platform can be improved.The retention and stickiness of the user on it.In view of the second point of the research,this paper explores the characteristics of the review data of the two platforms from different angles,studies and compares the characteristics of different music platform reviews and the atmosphere characteristics of the platform,and finds the similarities and differences of the review styles of the two platforms and the two platforms.The difference between the user's activity and active time.The platform can be targeted to launch platform albums and other activities in order to achieve better publicity.For the third point of the research,this paper first establishes a positive and negative sentiment classification model based on LSTM model,and completes the division of positive and negative emotion semantics with a high classification accuracy.The research results show that the emotional tone of both platforms is positive.of.Subsequently,we have established several machine learning algorithm models and compared the superiority of each model,and selected a model with good predictive effect to classify negative emotions and understand the specific composition ratio of negative emotions.We use machine learning technology to explore the emotional atmosphere characteristics of the online music community from the perspective of sentiment classification analysis,make up for the gaps in the current emotional analysis of online music community reviews,and enrich the research on sentiment classification theory in the field of online social networks.And this will also help enterprises to control the platform and emotional orientation of the platform,and contribute a certain reference value to the company to create a healthy and positive human music community. |
PDF Full Text Request