Eureka! Institutional Repository

This thesis aims at presenting a solution to the problem of "Fake news" using machine learning. The problem of "Fake news" and ways of dealing with it both with conventional ways and methods of mechanical learning is presented. In addition to that, ways of representing text in vector spaces with word2vec and GloVe are presented. Finally, the experiments performed using SVM for the classification on two sets of data and solutions for the above problem are presented.