Factify-AI:Detecting Deception in Digital News Using RoBERTa and NLP Techniques

We are living in a time when fake news is spreading fast. This is happening because of all the digital media outlets that are around. Many people are getting a lot of information. This is because new technologies are allowing news to spread quickly across platforms. People might not even realise that the news they saw is not real. This can happen after it has already been shared with others. A lot of information is being spread online. This can change how people think about things.The goal of this project is to help solve this problem. We want to make a program that uses machine learning and natural language processing. This program will help find news.The program will use machine learning. Machine learning will help the program learn what fake news looks like.The program will use natural language processing. Natural language processing will help the program understand language.By using these two technologies the program will be able to find news.Fake news is a problem. We hope that this program will help make the internet a trustworthy place.This project will help people to be more careful when they read news online.It will also help people to know what is real and what is notThis project uses two data sets from Kaggle. These are lists of news articles and fake news articles. The people working on this project removed any articles, punctuation and common words, like "the" and "and" from these data sets. After that they broke down the text into words and changed them into their basic form. Then they used something called TF-IDF to give numbers to these words. The Fake News Detection Program uses these numbers to create embeddings. The program created three machine learning models to detect news. These models are Multinomial Bayes, Logistic Regression and RoBERTa. The project uses these models Multinomial Naive Bayes, Logistic Regression and RoBERTa to figure out if a news article's real or fake.