NEWSIFY: - Article Summarization using Natural Language Processing and News Authentication using TF-IDF Vectorizer and Passive Aggressive Classifier
Rohit Shelar, Yash Gujar, Niranjan Pawal, Pratiksha Londhe, Sonali Rangdale · 2023
Reading news articles is an integral part of our daily life. It is the responsibility of journalists and editors to ascertain which articles are popular and real in order to effectively devote resources to provide a better reading experience to the users. The goal of this study is to create a model that simultaneously handles summarizing and determining whether or not a given news item is authentic. The system described here will be employed in a real-time setting. In fake news detection model, fake news by utilizing the TF-IDF vectorizer and the Passive Aggressive Classifier is implemented. Experimental results achieved an accuracy of 95.11% by combining these two methods. The system is further extended for obtaining a summary of news articles on a particular topic. This is implemented using Spacy, Genism, and NLTK methods. The time required to generate summary of news articles is found to be decrease as compare to earlier results. Experimental setup shows that the reading time is cut from “1.4 minutes” by NLTK to “0.47 minutes” by SumY, “0.47 minutes” by Genism, and “0.27 minutes” by Genism. However, the Spacy algorithm yields the best results, reducing the reading time from “1.4 minutes” to “0.48 minutes” and providing us with a high level of accuracy in the summary content, which includes all of the significant points.