The filtering process consists of translating documents to their vector space representations, finding documents which are similar to the profiles and selecting the top-scoring articles for presentation to the user.
