zeeshanlakhani/proof_docsimilarity

This project is a proof-of-concept tool that figures out how similar two pieces of text are to each other. It was built as a class assignment for a math course at NYU, and it lets you feed it two documents or text passages to get back a measure of how closely they relate. Under the hood, it works by turning text into numbers and then comparing those numbers using basic linear algebra. The approach ignores "stop words", common words like "the" or "and" that don't carry much meaning, and focuses on the more significant words in each document. By converting the remaining words into a mathematical representation, the tool can calculate how much the two documents overlap in content. The main use case is academic: someone learning about document similarity models, vector spaces, or the math behind text comparison. You would run it in MATLAB by calling a function with two text inputs, and it returns a similarity score. The README notes that a PDF with more detail about the problem is included in the repository. It's worth noting that this is a straightforward academic exercise rather than a production-ready library. The author even apologizes in advance for potentially messy code, so it's best treated as a learning reference rather than something to build on directly.