Klasifikace fake news na datech z Twitteru

The spread of misinformation on social media platforms, particularly Twitter, poses a significant challenge to public debate. This paper presents a classification system for identifying misinformation tweets using Natural Language Processing (NLP) algorithms. The system integrates machine learning models such as Recurrent Neural Networks (RNN), Logistic Regression and Random Forest with NLP techniques including lemmatization and tokenization. By addressing issues such as the diversity of misinformation, the quality and quantity of training data, and the reliance on linguistic features alone, the proposed classification system demonstrates its potential for detecting misinformation tweets. Experimental evaluation showcases the system's effectiveness in identifying misinformation, suggesting its potential to mitigate the spread of misinformation on social media platforms and enhance public discourse.