An overview of microblog user geolocation methods

作者:

Highlights:

• This paper reviews the existing microblog user geolocalisation methods, and summarizes a general framework for microblog user geolocalisation. Along with the framework, we review the studies on user geolocalisation from the aspects of data acquisition, data preprocessing, location representation, user geolocalisation methods categorization, and evaluation metrics.

• Based on the input of the geolocalisation algorithm, we categorize microblog user geolocalisation methods into three categories: text-based methods, network-based methods, and multi-view based methods. We summarize the advantages and limitations of each type of method theoretically.

• We conduct a performance comparison of existing methods based on the results reported in existing literature with the most widely used real-world datasets and evaluation metrics. The advantages and disadvantages of existing methods are further uncovered by comparing them experimentally. Important research challenges that may need further attention are discussed according to our analysis.

• Survey findings conclude that multiview-based methods are superior to the text-based methods as well as the network-based methods. Besides that, existing user geolocalisation methods cannot capture the user's home location change, resulting in the misjudged results. Also, how to locate users from multiple social platforms is relatively unaddressed. Hence, the geolocalization of users across multiple social platforms is one of the problems that deserve further research.

摘要

•This paper reviews the existing microblog user geolocalisation methods, and summarizes a general framework for microblog user geolocalisation. Along with the framework, we review the studies on user geolocalisation from the aspects of data acquisition, data preprocessing, location representation, user geolocalisation methods categorization, and evaluation metrics.•Based on the input of the geolocalisation algorithm, we categorize microblog user geolocalisation methods into three categories: text-based methods, network-based methods, and multi-view based methods. We summarize the advantages and limitations of each type of method theoretically.•We conduct a performance comparison of existing methods based on the results reported in existing literature with the most widely used real-world datasets and evaluation metrics. The advantages and disadvantages of existing methods are further uncovered by comparing them experimentally. Important research challenges that may need further attention are discussed according to our analysis.•Survey findings conclude that multiview-based methods are superior to the text-based methods as well as the network-based methods. Besides that, existing user geolocalisation methods cannot capture the user's home location change, resulting in the misjudged results. Also, how to locate users from multiple social platforms is relatively unaddressed. Hence, the geolocalization of users across multiple social platforms is one of the problems that deserve further research.

论文关键词:Social media,Microblog user,Twitter,Sina weibo,Geolocalisation

论文评审过程:Received 27 March 2020, Revised 11 August 2020, Accepted 16 August 2020, Available online 25 August 2020, Version of Record 20 October 2020.

论文官网地址:https://doi.org/10.1016/j.ipm.2020.102375