Analysing language for preventing women from gender violence: NLP and machine learning techniques to classify tweet messages

Gender violence is often overlooked, with few women seeking help from anti-violence shelters. This study explores how to effectively recognise women at risk of violence through the combined use of multidimensional statistical techniques such as text mining, Natural Language Processing, and machine learning to detect signs of violence in social media posts. Analysing the characteristics of the language used by social media users, this research aims to develop an automated system for the early detection of women potentially exposed to gender violence. This information could be employed to provide these selected social media users with details on available healthcare, psychological, and legal support services.