The Use of Computational Text-Mining Methods to Understand and Detect Intimate Partner Abuse

该项目寻求使用计算文本挖掘和信息检索方法来研究大型数据集中的亲密伴侣虐待。这项工作分为三个部分：1)应用计算文本挖掘来识别社交媒体论坛上的心理虐待报告，并检查这些报告的普遍性和内容；2)利用计算文本分析在警察案件摘要的大型数据集中调查非致命勒死和技术辅助跟踪和骚扰的流行程度；3)收集和分析亲密伴侣之间的辱骂短信的假名语料库，作为训练机器学习模型以自动识别辱骂短信的第一步。

The project seeks to use computational text mining and information retrieval methods to research intimate partner abuse in large datasets. The work is split into three parts: 1) applying computational text mining to recognise reports of psychological abuse on social media forums, and examine the prevalence and content of these reports; 2) using computational text analysis to investigate the prevalence of Non-Fatal Strangulation and tech-facilitated stalking and harassment in a large dataset of police case summaries; and 3) the collection and analysis of a pseudonymised corpus of abusive text messages between intimate partners, as a preliminary step towards training a machine learning model to automatically recognise abusive text messages.