The problem of hoax news is something universal that can change someone perspective. The impact can generate fear, racism ideas, and lead to oppression and violence against innocent people. This problem has existed since time immemorial, even though there is no access to information throughout the world. Currently there are 2.5 trillion bytes of data and it increasing, leading to the next problem, i.e. faster at analyzing and identifying hoax news and real news. The ideal solution is using Deep Learning algorithm and the author chooses Recurrent Neural Network (RNN) method and it's particular method Long Short-Term Memory (LSTM). The accuracy obtained is 99%, greater than Machine Learning methods such as Rocchio or Multinomial Naive Bayes.
Copyrights © 2020