(23-12-2025, 12:21 PM)DG97EEB wrote: You are not allowed to view links. Register or Login to view.Fascinating analysis.. Have you looked at Cosine similarity based purely on token commonality?
I'll try it with word frequency vectors, it should be interesting. I'm trying to figure out what makes You are not allowed to view links. Register or Login to view. similarity special. The usefulness of the SVD is totally unclear to me, maybe all this linear algebra is overkill for the purpose of comparing pages. Note: frequencies may need to be adjusted with tf-idf because the most frequent words are usually less meaningful than rare words.

