Scholarlink

The rise of social media networks has become a critical component of modern human living due to recent developments and advances in computer technology. The present context has evolved into a widely recognized medium for the distribution of information and news pertaining to a wide range of topics, including daily reports. This timeframe serves as the main period for data collecting and information transmission. This environment has a multitude of advantages; yet, it also entails a significant amount of misinformation and material that may potentially perplex both readers and users seeking accurate information. A significant concern associated with this system is to the limited availability of reliable information and real-world news pertaining to social media. An integrated system has been built that leverages several components of chain type of technology and natural language processing (NLP) to use machine learning techniques for the purpose of detecting false news and enhancing the accuracy of identifying fraudulent user accounts and postings. Various methodologies have been used to extract and choose features in order to provide effective training rules and appropriately assess the test classifier. A comprehensive empirical investigation demonstrates that the Twitter dataset is a useful resource for assessing the precision of fake news detection in classification tasks. The deep learning based RNN system demonstrates a high accuracy rate of 98.80% throughout both training and testing phases. This accuracy surpasses that attained by traditional machine learning methods such as SVM, ANN, and NB. Furthermore, the proposed methodology has the capability to detect superfluous information, a phenomenon that may be corroborated via several sources or by referencing current events. The prime purpose of this approach is to enhance the eminence of the material, identify any problematic sections within the text, and generate the original versions of those sections, all while preserving the original content of the text. The proposed methodology presents a testing approach aimed at identifying and discerning incorrect information within a given subject matter. It involves evaluating the veracity of assertions across a range of test scenarios.

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

Fake News Detection and Classification Using Deep Learning Techniques for Social Media Text

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • A. Kumari Shalini,
  • Sameer Saxena,
  • B. Suresh Kumar

摘要

The rise of social media networks has become a critical component of modern human living due to recent developments and advances in computer technology. The present context has evolved into a widely recognized medium for the distribution of information and news pertaining to a wide range of topics, including daily reports. This timeframe serves as the main period for data collecting and information transmission. This environment has a multitude of advantages; yet, it also entails a significant amount of misinformation and material that may potentially perplex both readers and users seeking accurate information. A significant concern associated with this system is to the limited availability of reliable information and real-world news pertaining to social media. An integrated system has been built that leverages several components of chain type of technology and natural language processing (NLP) to use machine learning techniques for the purpose of detecting false news and enhancing the accuracy of identifying fraudulent user accounts and postings. Various methodologies have been used to extract and choose features in order to provide effective training rules and appropriately assess the test classifier. A comprehensive empirical investigation demonstrates that the Twitter dataset is a useful resource for assessing the precision of fake news detection in classification tasks. The deep learning based RNN system demonstrates a high accuracy rate of 98.80% throughout both training and testing phases. This accuracy surpasses that attained by traditional machine learning methods such as SVM, ANN, and NB. Furthermore, the proposed methodology has the capability to detect superfluous information, a phenomenon that may be corroborated via several sources or by referencing current events. The prime purpose of this approach is to enhance the eminence of the material, identify any problematic sections within the text, and generate the original versions of those sections, all while preserving the original content of the text. The proposed methodology presents a testing approach aimed at identifying and discerning incorrect information within a given subject matter. It involves evaluating the veracity of assertions across a range of test scenarios.