Inspiration

We noticed that many students submit similar project ideas, and checking originality manually is time-consuming. This led us to build PRISM to automate idea validation using AI.

What it does

PRISM checks project ideas for similarity with past work and gives decisions like Accept, Modify, or Reject, helping reduce duplication.

How we built it

Built using React.js, Node.js, Python (Flask), and MySQL. Uses NLP (TF-IDF + cosine similarity) to compare project ideas.

Challenges we ran into Tuning similarity thresholds Handling different writing styles Limits of TF-IDF in understanding meaning Accomplishments Built a working AI system Achieved ~88% accuracy Reduced manual checking effort What we learned NLP and similarity detection Full-stack development Importance of data preprocessing What’s next Add advanced models like BERT/SBERT Improve semantic understanding Support multiple languages

Built With

  • an
  • and
  • architecture
  • authentication.
  • cosine
  • follows
  • for
  • idea
  • jwt
  • mvc
  • prism-is-developed-using-react.js-for-the-frontend-and-node.js-with-express.js-for-the-backend.-the-nlp-engine-is-implemented-in-python-using-flask-and-scikit-learn.-it-uses-mysql-8.0-as-the-database.-communication-is-done-through-rest-apis
  • similarity
  • system
  • tf-idf
  • the
  • uses
  • with
Share this project:

Updates

Submission history