Fake News Detection using NLP

• 

  roc curve

• 

  scikit learn perfomance metrics

• 

  validation accuracy of 100 percent

• 

  auc score

Inspiration

Fake news is something we see very often, especially on social media platforms. It’s not always easy to identify whether a piece of news is real or misleading. This inspired me to build a system that can automatically classify news articles and help in identifying fake content.

What it does

This project classifies news articles as real (TRUE) or fake (FALSE) using NLP techniques. It takes the title and full text of an article, processes it, and predicts whether the news is genuine or not.

How we built it

I combined the title and text of each article to capture more context. Then I cleaned the text by removing unnecessary characters and converting it to lowercase. For feature extraction, I used TF-IDF with unigrams and bigrams. Finally, I trained a Logistic Regression model to perform the classification.

Challenges we ran into

Handling dataset issues was one of the main challenges. There were encoding errors while loading the data, and the labels needed cleaning because of inconsistent formats. Making sure the final output file matched the required submission format was also important.

Accomplishments that we're proud of

The model achieved around 100% validation accuracy, which shows that even simple techniques can perform really well when applied properly. Completing the full pipeline within the given time was also a big achievement.

What we learned

I learned how important preprocessing is in NLP tasks. Feature engineering using TF-IDF made a big difference in performance. I also gained better understanding of building an end-to-end machine learning workflow.

What's next for Fake News Detection using NLP

In the future, this project can be improved by using advanced models like BERT or other transformer-based approaches. It can also be extended to support multiple languages and deployed as a real-time web application.

Built With

• collab
• logistic-regression
• pandas
• python
• scikit-learn
• tf-idf