CourseFuse

Python, Neo4j based project for graph-based similarity using BERT and Cosine Similarity.
Case Study - Syllabus Matching

Supervised by:- Dr Ujwala Bharambe
Project by:- Yash Dalwani and Etash Dorwani (Computer Engineering, Thadomal Shahani Engineering College)

Subjects

Currently works for the following Subjects:-

1. Artificial Intelligence and Machine Learning (AI and ML)
2. Software Engineering and Project Management (SE and PM)
   More subjects to be added in the near future.

Project Components

1. PreProcessing:-
   OCR implemented using Tesseract OCR from PyTesseract.
   Keyword Extraction done using Gensim Library.
2. GraphQL:-
   Used Neo4j for creating Graphs using Cypher Queries.
3. Similarity Prediction:-
   Made a fine tuned BERT model to tokenise the nodes of the graph and predict the final global similarity by normalising the Similarity Matrix created using Cosine Similarity.
4. Output:-
   Created a Website to display all the studies found out by implementing the above steps.

Output