Twitter Bot Detection Using Neural Networks and Linguistic Embeddings

Twitter is a web application playing the dual role of online social networking and micro-blogging. The popularity open structure have attracted large number automated programs, known as bots. In this article, we propose bot detection model using recurrent neural networks, specifically bidirectional lightweight gated unit (BiLGRU), linguistic embeddings. To best our knowledge, first that does not require any handcrafted features, or prior knowledge assumptions about account profiles, friendship networks historical behavior. proposed uses only textual content tweets embeddings to classify human accounts on Twitter. Experimental results show performs better comparably state-of-the-art models while requiring no feature engineering, making it faster easier train deploy in real network. We also present experimental performance computational costs different types recurrence network variants for task detection. will potentially help researchers design high-performance deep-learning similar tasks.