Use TF-IDF to compute query and document similarity
- The idea is to:
- compute term frequency for each term in each document
- compute inverse document frequency for each term in all documents and query as a weight for each term
- compute tf-idf for each document and the query
- use cosine similarity as a metric to compute query-document similarity