A Study on Cluster Variation of Twitter Hashtags during a Whole Day
Nartlada Bhakdisuparit, Iwao Fujino · Institutional Repositories DataBase (IRDB) · 2019
Our study aims to verify the cluster variation of Twitter hashtags according to their word distributions.By gathering tweets from Twitter sample data and saving tweets in a time period of each hour separately, we extract hashtags and calculate their word distributions for each time period.Furthermore, by using Ward's hierarchical method, we cluster hashtags to a structure of groups according to the Jensen-Shannon divergence between any two hashtags and show the clustering results with dendrogram, which represents the hierarchical structure of all hashtags.Finally, we verify the differences of clustering results between each time period to analyze the details of cluster variation.The achievement of this study will contribute to develop marketing campaign or advertisement on Twitter.