Scholarlink

With social media becoming a dominant channel for communication, the proliferation of misinformation and rumors presents significant challenges. This paper presents a machine learning-based approach tailored for real-time identification of rumors on Twitter. Utilizing a Convolutional Neural Network (CNN), the system excels in classifying tweets with an impressive accuracy of 88%. This performance is attributed to the CNN’s proficiency in deciphering complex textual patterns and nuances. Additionally, an ensemble model comprising Random Forest, Support Vector Machine (SVM), and Logistic Regression also shows considerable efficacy, achieving 87% accuracy. This framework is embedded within a user-friendly, Streamlit-powered web interface that provides instant feedback on tweet classifications, thus enhancing usability. This system adeptly handles linguistic challenges such as sarcasm and ambiguous language, offering a scalable and reliable tool for combating misinformation and bolstering the integrity of social media platforms.

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

Advancing Social Media Integrity: Rumor Detection on Twitter Using Convolutional Neural Networks

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • K. Uthej Kumar,
  • Kit Denis Anaparthi,
  • Manukumar Shanthi Thangam

摘要

With social media becoming a dominant channel for communication, the proliferation of misinformation and rumors presents significant challenges. This paper presents a machine learning-based approach tailored for real-time identification of rumors on Twitter. Utilizing a Convolutional Neural Network (CNN), the system excels in classifying tweets with an impressive accuracy of 88%. This performance is attributed to the CNN’s proficiency in deciphering complex textual patterns and nuances. Additionally, an ensemble model comprising Random Forest, Support Vector Machine (SVM), and Logistic Regression also shows considerable efficacy, achieving 87% accuracy. This framework is embedded within a user-friendly, Streamlit-powered web interface that provides instant feedback on tweet classifications, thus enhancing usability. This system adeptly handles linguistic challenges such as sarcasm and ambiguous language, offering a scalable and reliable tool for combating misinformation and bolstering the integrity of social media platforms.