Identification and analysis of fake or real news using machine learning model

Abstract

In recent year, spreading of fake news has been one of the most highlighted things society is dealing with. We have seen in past that fake news leads to serious riots and created a fearful en- vironment for general people. Fake news refers to misleading information presented as legitimate news. This misinformation may be purposefully created and spread with a intention to distort people’s view and opinions about sensitive issues. The frequency of these fake news is too high and often have intentions to spread propaganda or accomplished some particular objective. Fake news at the time of election may influence people views about particular parties and it also have potential to affect the integrity of election and democratic processes .It can incite violence and killings, like we saw in the case of the 2020 USA Capitol attack. This shows how important it is to identify fake news and stop its transmission. Through this study, we came up with a platform that is efficient enough to suggest whether the news is authentic or misleading. For this, we have analysis different fake new detection websites such as boom live fact checker ,vishwas.net ,ndtv fact checker etc and found they have a whole big background team which works manually to verify news which is very time consuming and they need to go through lots of resources to check the authenticity of the news. We have created our own database with the help of news scrapper function. After this, we have created a machine learning and transformer which is end-to-end trainable and efficient enough to predict whether the news is misleading or not. We integrate our model with a platform that we have created to get better graphical user interface. There,we pro- vide an option to put the news and our platform lets you know the correctness of that news. We have also provided an option to check news through uploading audio. Last but not least, we have also provided the credibility of websites also.