An Automated Sentiment-Driven News Summarization and Categorization System Using Web Scraping and NLP

This paper presents an automated system for summarizing and filtering news articles based on sentiment analysis.The system leverages Python-based tools to fetch news headlines using the Google News API, scrape article content using Beautiful Soup and Newspaper3k, and select optimal content through similarity scoring with the sentence-transformers/all-MiniLM-L6-v2 model.Summarization is performed using the Llama 3.2 3B model, while sentiment classification is achieved using the cardiffnlp/twitter-roberta-base-sentiment-latest model.The processed data is stored in Firebase and accessed via an Android app, enabling users to filter negative news and select preferred categories.The system processes 35-40 news articles in 10-11 minutes, significantly outperforming manual efforts.This approach enhances efficiency in news consumption while ensuring scalability across six categories: world, nation, business, technology, entertainment, and sports.