EVERLASTING RUMOR DETECTION FRAMEWORK

  • Unique Paper ID: 167294
  • Volume: 11
  • Issue: 3
  • PageNo: 883-888
  • Abstract:
  • — In this study, we address the challenge of rumor detection on Weibo and Twitter, where training data is limited and news updates rapidly. Leveraging Lifelong Machine Learning (LML), we propose a novel approach to continuously learn and accumulate knowledge for improved detection performance. We explore various models including BERT GCN, Bi-GCN, RvNN, Naive Bayes, Decision Tree, SVM-RBF, SVM-TK, and Voting Classifier. Evaluating these models on Weibo and Twitter datasets, we observe significant performance variations. BiGCN achieves 90% accuracy, while BERT GCN with LSTM and CNN, LSTM, LSTM + GRU, and Voting Classifier attain 93%, 99%, 95%, and 97% accuracy, respectively. These results underscore the effectiveness of lifelong learning paradigms in enhancing rumor detection in dynamic online environments. Our findings highlight the potential of leveraging diverse algorithms in conjunction with LML for robust and accurate rumor detection in social media platforms.

Cite This Article

  • ISSN: 2349-6002
  • Volume: 11
  • Issue: 3
  • PageNo: 883-888

EVERLASTING RUMOR DETECTION FRAMEWORK

Related Articles