Survey of fake news detection using machine intelligence approach

With the extensive spreading of all information through digital platforms, it is of maximal importance that each people get to differentiate between them. Fake news is a vast problem in our society we cannot predict which news is fake or real without having knowledge or proof of that particular news. This has become a supreme problem, so we decided to create a solution to this problem. Thus, we built a small model which helps in detecting fake news, where we are dealing with some articles which have been collected from the internet. We have labeled each of them as either fake or true. We have trained our dataset using these articles and have used different machine learning algorithms like Passive Aggressive Classifier, Naive Bayes, Logistic Regression, Decision Tree, Long short term memory (LSTM), and Bidirectional Encoder Representations from Transformers (BERT) to compare the results. Our experimental result has achieved 99.6% accuracy from Decision Tree algorithm and obtained 99.8% recall from LSTM for detection of fake news. Passive Aggressive Classifier performs excellent on a large data set.