ABSTRACT
Navigating through large volumes of information has become increasingly challenging, highlighting the need for effective document summarisation techniques. This project aims to seamlessly navigate and summarize text documents using a combination of advanced methods, including BERT, Spacy and LexRank. The summaries generated by these models will be evaluated using the Rouge score to determine the most effective summarization approach. This paper proposes a supervised machine learning technique, Passive aggressive classifiers that uses Count Vectorizer and TermFrequency Inverse Utilizing Document Frequency Vectorizer for feature extraction to identify propaganda news based on the polarity of the associated article. Furthermore, this project incorporates a Passive Aggressive Classifier to distinguish between genuine and fake documents, enhancing the summarised document analysis and decision-making process. The integration of diverse summarisation techniques and a classifier contributes to a comprehensive solution for efficient information processing and verification. The aim of this research endeavor is to create a model that simultaneously handles summarising and assessing the authenticity of a provided news article.
