Scholarlink

<p>Violence against woman is a persistent problem affecting our society. The use of Big Data represents a new challenging opportunity to integrate official statistics with more updated information. Social media constitute, in fact, a particularly useful data source for analysing gender-based violence, cyber-violence and gender stereotypes. This report describes the results of a methodological approach, aimed at studying gender stereotypes, using textual data published on social media, showing which positive or negative effects may be generated in public opinion when certain messages are spread. Sentiment and emotion analysis has been carried out to measure how the phenomenon is represented among social network users. The statistical quality of the results was assessed. The proposal of a methodology to generate a linguistic resource aimed at improving the capacity of the machine learning BERT algorithm to classify social media data on gender stereotypes integrates the specific contribution of this study.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Challenging Big Data for studying gender-based violence: a methodological proposal

  • 期刊论文

  • 发表日期：

    2025年6月20日

查看全文
  • Fiorenza Deriu,
  • Claudia Villante,
  • Maria Giuseppina Muratore,
  • Raffaella Gallo,
  • Edoardo Toppetti

摘要

Violence against woman is a persistent problem affecting our society. The use of Big Data represents a new challenging opportunity to integrate official statistics with more updated information. Social media constitute, in fact, a particularly useful data source for analysing gender-based violence, cyber-violence and gender stereotypes. This report describes the results of a methodological approach, aimed at studying gender stereotypes, using textual data published on social media, showing which positive or negative effects may be generated in public opinion when certain messages are spread. Sentiment and emotion analysis has been carried out to measure how the phenomenon is represented among social network users. The statistical quality of the results was assessed. The proposal of a methodology to generate a linguistic resource aimed at improving the capacity of the machine learning BERT algorithm to classify social media data on gender stereotypes integrates the specific contribution of this study.