Scholarlink

Social media is an interactive and reliable tool used for communication among people around the world. It provides a facility for people to make accounts of it and to share their personal ideas, photos, interests, and other expressions. Through social media, people can easily interact with each other. Everyone is in contact with their loved ones. In social media mostly used applications are Facebook, Instagram, WhatsApp, Twitter, LinkedIn, and YouTube. Due to the ever-increasing popularity of social media, it has become an ideal platform for the creation of fake accounts by spammers for undesirable activities. Spammers make fake accounts on social media to create panic among people. In this research, we aimed to detect fake profiles in one of the most used content-sharing applications Twitter. We have used the MIB Twitter dataset. Based on user-based attributes we have implemented deep learning techniques LSTM and GRU, and we proposed a hybrid LSTM-GRU. Through experimental results, we gain better accuracies and small losses with deep learning techniques than machine learning algorithms for efficient detection of fake Twitter profiles.

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

Efficient Detection of Fake Twitter Profiles on Social Media Through Hybrid Deep Learning Model

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Haleema Bibi,
  • Humaira Ashraf,
  • N. Z. Jhanjhi

摘要

Social media is an interactive and reliable tool used for communication among people around the world. It provides a facility for people to make accounts of it and to share their personal ideas, photos, interests, and other expressions. Through social media, people can easily interact with each other. Everyone is in contact with their loved ones. In social media mostly used applications are Facebook, Instagram, WhatsApp, Twitter, LinkedIn, and YouTube. Due to the ever-increasing popularity of social media, it has become an ideal platform for the creation of fake accounts by spammers for undesirable activities. Spammers make fake accounts on social media to create panic among people. In this research, we aimed to detect fake profiles in one of the most used content-sharing applications Twitter. We have used the MIB Twitter dataset. Based on user-based attributes we have implemented deep learning techniques LSTM and GRU, and we proposed a hybrid LSTM-GRU. Through experimental results, we gain better accuracies and small losses with deep learning techniques than machine learning algorithms for efficient detection of fake Twitter profiles.