I have text data for 10 years and I want to run a Word2Vec model on the text data each year and calculate the distances between words and I would like to compare them over time.

Are there any papers in which the Word2Vec model is compared to previous years data? Or if I have a matrix of the Word2Vec vectors are there ways to compare these matrices year-on-year or from period t to t-1?

I can compute the cosine distance of the matrix but cannot think how to compare these "scores" for each word over time.


1 Answer 1


Yes, you can check Diachronic Word Embeddings Reveal Statistical Laws of Semantic Change


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.