摘要：In recent years, social media is growing at anunprecedented rate, and more people have become influencers.Understanding popularity helps ordinary users to boostpopularity, and business users to choose better influencers.There were studies to predict the popularity of posted imageson social media, but there was none on the user's popularity asa whole. Furthermore, existing studies have not taken hashtaganalysis into consideration, one of the most useful social mediafeature. This research aims to create a model to predict a user'spopularity, which is defined by a combination of engagementrate and followers growth. There were six machine learningregression models tested. The proposed model successfullypredicted the users’ popularity, with R 2 up to 0.852, usingRandom Forest with 10-fold cross-validation. The additionalstatistical analysis and features analysis results revealed factorsthat can boost popularity, such as actively posting andfollowing users, completing user's metadata, and using 11hashtags. In contrast, it was also found that having a largenumber of posts and following in the past will not help ingrowing popularity, as well as the use of popular hashtags.
关键词：Machine Learning; Regression Analysis;
Social Media; Predictive Model
