RweetMiner: Automatic identification and categorization of help requests on twitter during disasters

作者:

Highlights:

• Redefining request under the term “rweet” in the context of social networking sties, as well as defining its primary types and subtypes.

• Proposing optimized and effective preprocessing strategy.

• Generating n-grams (bag of words) with n = 1, 2, and 3, combining them with each other and rule based features for learning subtle differences between request and non-request tweets, as well as six different types of request tweets.

• Store intermediate data to speed up the machine learning development life cycle.

• Performance improvement on the request identification and request categorization on Twitter.

摘要

•Redefining request under the term “rweet” in the context of social networking sties, as well as defining its primary types and subtypes.•Proposing optimized and effective preprocessing strategy.•Generating n-grams (bag of words) with n = 1, 2, and 3, combining them with each other and rule based features for learning subtle differences between request and non-request tweets, as well as six different types of request tweets.•Store intermediate data to speed up the machine learning development life cycle.•Performance improvement on the request identification and request categorization on Twitter.

论文关键词:Disaster response,Machine learning,Social networking sites,Intermediate data,Request tweets,Intermediate results,Relief efforts

论文评审过程:Received 6 October 2019, Revised 30 January 2021, Accepted 22 February 2021, Available online 2 March 2021, Version of Record 1 April 2021.

论文官网地址:https://doi.org/10.1016/j.eswa.2021.114787