Ranking influencers of social networks by semantic kernels and sentiment information
作者：
Highlights：
• LDA including proper pooling strategy can be applied to decide suitable categories.
• Semantic kernel classifiers improve topic based tweet classification performance.
• Retweet count and spread score can be used to evaluate influencers.
• Preprocessing is important for short text classification.
• The performance of semantic classifiers depends on the size of the training set.
摘要
•LDA including proper pooling strategy can be applied to decide suitable categories.•Semantic kernel classifiers improve topic based tweet classification performance.•Retweet count and spread score can be used to evaluate influencers.•Preprocessing is important for short text classification.•The performance of semantic classifiers depends on the size of the training set.
论文关键词：Social network analysis,Opinion leader detection,Flow of influence,Sentiment polarity score,PageRank algorithm,Semantic kernels
论文评审过程：Received 30 September 2020, Revised 22 December 2020, Accepted 9 January 2021, Available online 14 January 2021, Version of Record 29 January 2021.
论文官网地址：https://doi.org/10.1016/j.eswa.2021.114599
