Multi-label Arabic text classification in Online Social Networks

作者：

Highlights：

• We construct a standard multi-label Arabic dataset using manual annotation and semi-supervised annotation techniques.

• We train machine-learning models for topic classification, sentiment analysis, and multilabel classification in OSNs.

• We examine the relationship between topics published in OSNs and hate speech.

• We propose a technique to filter social networks contents.

摘要

•We construct a standard multi-label Arabic dataset using manual annotation and semi-supervised annotation techniques.•We train machine-learning models for topic classification, sentiment analysis, and multilabel classification in OSNs.•We examine the relationship between topics published in OSNs and hate speech.•We propose a technique to filter social networks contents.

论文关键词：Arabic natural language processing,Arabic text classification,Arabic sentiment analysis

论文评审过程：Received 16 May 2020, Revised 4 March 2021, Accepted 30 March 2021, Available online 10 April 2021, Version of Record 16 April 2021.

论文官网地址：https://doi.org/10.1016/j.is.2021.101785