Predicting tweet impact using a novel evidential reasoning prediction method
作者:
Highlights:
• MAKER-RIMER provides a better interpretability and transparency of the inference reasoning process and results.
• The MAKER-RIMER prediction performs better in terms of misclassification error.
• Highly shared tweets contain mentions to contender, with positive/negative emotions.
• Tweets written at the end of the campaign tend to be high impact.
• Presence of hashtags do not have influence on reaching high impact tweets.
摘要
•MAKER-RIMER provides a better interpretability and transparency of the inference reasoning process and results.•The MAKER-RIMER prediction performs better in terms of misclassification error.•Highly shared tweets contain mentions to contender, with positive/negative emotions.•Tweets written at the end of the campaign tend to be high impact.•Presence of hashtags do not have influence on reaching high impact tweets.
论文关键词:Evidential reasoning rule,Belief rule-based inference,Maximum likelihood data analysis,Twitter,Retweet,Prediction
论文评审过程:Received 29 October 2019, Revised 24 November 2020, Accepted 27 November 2020, Available online 13 December 2020, Version of Record 25 December 2020.
论文官网地址:https://doi.org/10.1016/j.eswa.2020.114400