The Explore of Using Deep Learning Models for Fake News Classification

摘要

Recently, due to the booming Internet, the Internet is full of fake news headlines, and people who are interested can manipulate the public or influence other people's ideas through the swaying fake news headlines. Therefore, it is an important task to identify fake news. In the Kaggle platform, the WSDM-Fake News Classification aims to identify fake news, but it does not perform well in terms of accuracy. Therefore, this study proposes a neural network architecture to solve this fake news task. For the semantic vector, Word2vec and ELMo are respectively proposed to compare and improve the accuracy.