Projects / Computer Science / Computer Science project 2

Fake News Detector for English and Hindi Headlines

Approx. component budget: Rs. 0 - 1,000

Objective

Build a baseline with TF-IDF and then a transformer model, and measure how much the better model improves results, including on Hindi text.

What you will build

A text classifier that reads a news headline or short article and labels it likely real or likely fake, with a confidence score and the words that influenced the result.

Results you can expect

• F1 score of TF-IDF + logistic regression vs a fine-tuned small transformer
• Confusion matrix
• Error analysis of 30 wrong predictions
• And response time per article

What you will learn

Text preprocessing, TF-IDF, embeddings, fine-tuning, evaluation on imbalanced data, and the limits of automatic fact checking.

Tools and notes

Python, scikit-learn, Hugging Face Transformers, a public fake-news dataset, Streamlit. Say clearly that the tool flags style patterns and does not prove truth.

Resume points from this project

• Built and tested Fake News Detector for English and Hindi Headlines; objective: build a baseline with TF-IDF and then a transformer model, and measure how much the better model improves results, including on Hindi text.
• Measured and analysed f1 score of TF-IDF + logistic regression vs a fine-tuned small transformer and confusion matrix; wrote the report and presented the results in a viva. Add your own numbers: [x].

Replace [x] with your own measured results. Write only what you really did.

More computer science projects

Back to all projects