A Study On User Portraits Based On The Top 100 UP Video Comments In The Bilibili Online Video Websit

Bilibili,as one of the most influential online video platforms in China,has attracted many young users with its unique bullet screen function and rich content.This article takes Bilibili’s top 100 UPs in2022 as the research object,uses text mining and machine learning methods to process their video bullet screens through data cleaning,word segmentation,frequency statistics,sentiment analysis,topic modeling and other methods.The results are presented and interpreted using visualization tools.By comparing and analyzing the bullet characteristics of different UPs and their affiliated sections on Bilibili platform,this article describes a portrait of Bilibili users in terms of preference,emotional tendencies,and explores factors that affect user portraits formation.This article provides a novel yet effective way to create user portraits for online video platforms while also providing researchers in related fields with a comprehensive insight into user behavior.