Detection of Fake News Relate to CoViD-19

The purpose of the paper is an approach to automatic recognition of alarming and misleading news, either of which are colloquially known under more commonly used term “Fake news”. The contribution describes the field of misinformation and particularly the methods suitable for generation of models for fake news detection. The detection models are learned on short text data, so they can classify a new unknown text into the classes labeled as “True” or “Fake”. The following supervised machine learning methods were used: K Nearest Neighbors, Naïve Bayes Classifier, Logistic Regression, Random Forests, Support Vector Machine, and Neural Networks. The results of experiments showed that the best method for this task are neural networks. They achieved the best results in Recall, Fl and Accuracy. Support Vector Machine achieved best results in Precision. The conclusion of the paper is that it is very important to prepare well-processed data as much as possible to achieve the best possible results. Another conclusion is that the overall performance of the models is highly dependent on a set of model parameters.