Scholarlink

<p>The rapid rise of social media platforms has raised significant concerns regarding data security and the spread of harmful or misleading content. To address these issues, this research presents an innovative model that combines topic mining and security analysis on social media through the use of Deep Learning (DL) techniques. Topic extraction is conducted using Bidirectional Encoder Representations from Transformers (BERT), enabling the extraction of meaningful insights from large-scale social media data. For security analysis, a Deep Belief Network (DBN) is employed to detect potential cyber threats and malicious behavior in online discussions. To improve the model’s accuracy, the Aquila Optimization Algorithm (AOA) is used to fine-tune the hyperparameters. The integration of BERT, DBN, and AOA results in a more effective solution for both topic identification and security threat detection in social networks, demonstrated using Weibo data. This work provides a comprehensive framework for social media surveillance, cybersecurity, and misinformation control, contributing to the development of safer and smarter online environments.</p> Graphical Abstract <p></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

A BERT-Driven Optimized Deep Belief Network for Topic Mining and Security Analysis on Social Media

  • 期刊论文

  • 发表日期：

    2025年9月13日

查看全文
  • K. Parameswari,
  • T. Kamalakannan

摘要

The rapid rise of social media platforms has raised significant concerns regarding data security and the spread of harmful or misleading content. To address these issues, this research presents an innovative model that combines topic mining and security analysis on social media through the use of Deep Learning (DL) techniques. Topic extraction is conducted using Bidirectional Encoder Representations from Transformers (BERT), enabling the extraction of meaningful insights from large-scale social media data. For security analysis, a Deep Belief Network (DBN) is employed to detect potential cyber threats and malicious behavior in online discussions. To improve the model’s accuracy, the Aquila Optimization Algorithm (AOA) is used to fine-tune the hyperparameters. The integration of BERT, DBN, and AOA results in a more effective solution for both topic identification and security threat detection in social networks, demonstrated using Weibo data. This work provides a comprehensive framework for social media surveillance, cybersecurity, and misinformation control, contributing to the development of safer and smarter online environments.

Graphical Abstract