| With the development of information technology and the popular ity of the Internet,the music industry has transformed a lot while the way people obtain music information has also been changed.After music has undergone many changes in storage medium,transmission mode,and payment models,online music platforms have emerged and become the mainstream choice for people to enjoy copyrighted music.With the increasingly fierce competition for copyrights,Internet user dividends have diminished,therefore,the online music platforms introduced the online community based on the concept of music social.Through encouraging users to participate in platform interaction,encouraging users to generate content(UGC),and creating a community atmosphere with musical elements,the online music platforms have established a good ecosystem of music content with the help of music comment,playlist sharing,and users’ personal homepage..It has become an important way for online music platforms to conduct fine management of users by deeply cultivating current users,observing user behavior,and inventing user requirement.By studying the behavioral characteristics of user groups in the online music platform,it can promote the platforms to better understand its user requirements and adjust service strategies to continuously enhance user stickiness and increase the user’s sense of belonging.The research focus of this paper is using the public user data provided by the online music platforms,to cluster the user groups by the method of user portrait,and to explore the behavioral characteristics of users in the whole user group and different user groups by time series analysis,power law distribution theory,regression analysis and other methods.First of all,this study summarizes the literature on the relevant concepts and development stages of online music platforms.Based on the existing research results of user behavior analysis and user portraits of online music platforms,combined with the rapid propagation of IP in the field of Internet pan-entertainment,the user is regarded as an IP with personality,content,originality and traffic attributes.Based on the “super IP” theory,the user portrait system and its generated content is constructed.In this study,QQ Music platform is used as an empirical analysis sample.In the empirical pro cess,the principal component analysis is used to reduce the dimensionality of the multi-dimensional features in users’ public data,the K-means algorithm is used to cluster the users,and the radar chart generated by the original data is combined to divid e the users into five types——high-flow head users,high-quality playlist creation users,high purchase intention users,pure music users,and marginal silent users.And the characteristics of these types are analyzed.Secondly,this study analyzes the specific interaction behaviors between users and online music platforms,exploring the characteristics of user comment behaviors,user playlist behaviors,and user purchase behaviors.Through researching on behavior occurrence time and behavior distribution rules,with the time series analysis,power law distribution theory,correlation systems,etc.,this study discovers the behavior characteristics of online music platform users: they are affected by external factors that make their behaviors influence the popularity of songs,among these playlists of songs,some are periodically popular,some are accidentally popular,and some are short-lived.There are a large number of dive users in the online music platforms,and very few usergenerated content attracts the attention of most users.The natural time interval for two consecutive user comments follows the power distribution law,with slight changes in different types of songs.The behavior and frequency of users’ participation in comments also follow the power distribution law. |
PDF Full Text Request