Social media analysis by innovative hybrid algorithms with label propagation

作者：

Highlights：

• Semantic values of terms improve topic based Tweet classification performance.

• Label Propagation Algorithm can be enriched by semantic weights on edges.

• Combination of semantic classifiers superior to either type.

• Preprocessing is important for short text classification.

• Label Propagation Algorithm with semantics is efficient in sentiment polarity.

摘要

•Semantic values of terms improve topic based Tweet classification performance.•Label Propagation Algorithm can be enriched by semantic weights on edges.•Combination of semantic classifiers superior to either type.•Preprocessing is important for short text classification.•Label Propagation Algorithm with semantics is efficient in sentiment polarity.

论文关键词：Label propagation algorithm,Social media analysis,Topic-based tweet classification,Sentiment polarity detection

论文评审过程：Received 3 February 2021, Revised 5 August 2022, Accepted 15 August 2022, Available online 18 August 2022, Version of Record 21 August 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.118606