SCIENCE@home

Predicting user gender on social media sites using geographical information

Rio Miura,Masaharu Hirota,Daiju Kato,Tetsu Araki,Masaki Endo,Hiroshi Ishikawa

Published 2018 in International ACM Conference on Management of Emergent Digital EcoSystems

ABSTRACT

Recently, the tourism industry has developed remarkably. Marketing for revitalizing the tourism market has attracted intense attention. To perform effective marketing, analyzing attributes such as gender, age, and residential areas of visitors is a fundamentally important approach because it is possible to present an appropriate advertisement to each user considering user attributes. As described in this paper, we propose a method to estimate user attributes based on geographical information annotated to contents posted by users in social media posts. Attributes of people visiting a specific area might be biased, such as "men visit Shimbashi frequently" and "women often visit Harajuku." Our approach assumes that "people with a certain attribute often visit a certain area" and that "such areas differ depending on attributes." Based on those assumptions, we create feature vectors based on geographical information related to social media sites. Furthermore, we propose a method to estimate user attributes with feature vectors using machine learning. As described in this paper, we specifically examine estimation of user gender. Our experiments demonstrated evaluation of the efficiency of gender estimation using the proposed method from a dataset obtained from Twitter and Flickr.

PUBLICATION RECORD

  • Publication year

    2018

  • Venue

    International ACM Conference on Management of Emergent Digital EcoSystems

  • Publication date

    2018-09-25

  • Fields of study

    Geography, Computer Science

  • Identifiers
    DOI 10.1145/3281375.3281383
  • External record

    Open on Semantic Scholar

  • Source metadata

    Semantic Scholar

CITATION MAP

EXTRACTION MAP

CLAIMS

  • No claims are published for this paper.

CONCEPTS

  • No concepts are published for this paper.

REFERENCES

Showing 1-25 of 25 references · Page 1 of 1

CITED BY