Using of n-grams from morphological tags for fake news classification

Research of the techniques for effective fake news detection has become very needed and attractive. These have a background in many research disciplines, including morphological analysis. Several researchers stated that simple content-related n-grams POS tagging had been proven insufficient classification. However, they did not realise any empirical results, which could confirm these statements experimentally last decade. Considering this contradiction, main aim paper is to evaluate potential common use tags correct classification true news. The dataset published or real about current Covid-19 pandemic was pre-processed using As result, were prepared further analysed. Three based on proposed applied different groups pre-processing phase detection. n-gram size examined as first. Subsequently, most suitable depth decision trees sufficient generalization scoped. Finally, performance measures models compared with standardised reference TF-IDF technique. model like accuracy, precision, recall f1-score are considered, together 10-fold cross-validation Simultaneously, question, whether technique can be improved researched detail. results showed newly comparable traditional At same time, it analysis improve baseline model, precision news, statistically significantly improved.