In CBF, the minimum distance between 2 vectors in the vector space model denotes the vector pair with maximum similarity. Distance between 2 vectors is denoted by
Cosine of the angle between 2 or more vectors. Both user tastes and item descriptions can be modeled as vectors in a vector space model. In the
Vector space model (also see
Dot product), each keyword is one dimension. Each dimension is weighted using TFIDF/ BM25, etc.
Keyword weighting: TF- IDF &
BM25 (
link1 +
wikipedia definition)