Projects per year
Abstract
Online misogyny presents a significant societal challenge, as it perpetuates gender inequalities and deters women from fully engaging in digital spaces. Traditional moderation methods often struggle to address the evolving and context-dependent nature of misogynistic language, necessitating more robust and adaptive solutions. To address this gap, this study introduces a comprehensive framework that leverages advanced natural language processing (NLP) techniques and strategic data augmentation to enhance the detection of misogynistic content. Key NLP contributions include the use of emoji decoding, GloVe embeddings, and LDA-based topic modelling, which collectively enhance contextual understanding and data richness. By using different machine learning (ML), deep learning (DL), and transformer-based models, the framework demonstrates versatility and adaptability in handling complex and contextually nuanced language. The performance analysis of the proposed approach highlights the effectiveness of the selected models. A comparative analysis of the results underscores the transformative role of data augmentation, which significantly enhanced the robustness and generalisation of all model categories and improved misogyny detection.
Original language | English |
---|---|
Article number | 856 |
Number of pages | 23 |
Journal | Applied Sciences |
Volume | 15 |
Issue number | 2 |
DOIs | |
Publication status | Published - 16 Jan 2025 |
Keywords
- online misogyny
- deep learning
- machine learning
- transformer models
- predication models
- NLP
- data augmentation
Fingerprint
Dive into the research topics of 'Leveraging advanced NLP techniques and data augmentation to enhance online misogyny detection'. Together they form a unique fingerprint.Projects
- 1 Active
-
AI-Driven Analysis of Online Misogyny (Development Project Funding)
Mohasseb, A. (PI) & Tranchese, A. (CoI)
1/10/24 → 1/05/25
Project: Research