vector space model
In plain wordsAn early, still-influential idea for search: treat every document and query as an arrow in a space with one direction per word, and rank documents by how closely their arrows point the same way as the query's.
A model of information retrievalinformation retrieval, n.: The study and engineering of search: how to organize huge piles of documents so that the few that matter to someone's question can be found and ranked quickly. AI systems rely on it to look things up. that represents documents and queries as vectors in a space with one dimension per index term, weighted for example by TF-IDFTF-IDF, abbr.: A simple score for picking out the words that characterize a document: a word counts for more the more often it appears there, and for less the more of the collection's other documents also use it., and ranks documents by the similarity of their vectors to the query vector, most often the cosine similaritycosine similarity, n.: A score from −1 to 1 for how closely two lists of numbers point the same way, ignoring their size. It is the standard test of whether two texts mean similar things.. It replaced yes-or-no Boolean matching with graded, ranked results and supported relevance feedback, in which the query vector is moved toward documents a user marks as relevant. Its geometric framing — meaning as position, relevance as proximity — carries over to latent semantic analysislatent semantic analysis, n.: An early technique for capturing word meaning by counting which words appear in which documents, then compressing those counts so that words used in similar contexts end up close together., word embeddingsword embedding, n.: A list of numbers standing for a word, learned so that words used in similar ways get similar lists. It lets a computer measure that 'cat' is closer to 'dog' than to 'car'. and modern dense retrievaldense retrieval, n.: Finding relevant documents by turning the question and every document into lists of numbers that capture meaning, then picking the documents whose numbers are closest to the question's., which replace term dimensions with learned ones.
Embedding search is the vector space model with learned dimensions instead of words.
Origin
Developed by Gerard Salton and colleagues for the SMART retrieval system at Harvard and Cornell from the 1960s; the classic statement is Salton, A. Wong and C. S. Yang, 'A vector space model for automatic indexing' (Communications of the ACM, November 1975).
Usage note
A frequently cited 1975 Salton paper on a 'vector space model for information retrieval' does not exist, as David Dubin documented in 2004; cite the Communications of the ACM paper instead.
Related
- information retrievalinformation retrieval, n.: The study and engineering of search: how to organize huge piles of documents so that the few that matter to someone's question can be found and ranked quickly. AI systems rely on it to look things up.
- TF-IDFTF-IDF, abbr.: A simple score for picking out the words that characterize a document: a word counts for more the more often it appears there, and for less the more of the collection's other documents also use it.
- cosine similaritycosine similarity, n.: A score from −1 to 1 for how closely two lists of numbers point the same way, ignoring their size. It is the standard test of whether two texts mean similar things.
- latent semantic analysislatent semantic analysis, n.: An early technique for capturing word meaning by counting which words appear in which documents, then compressing those counts so that words used in similar contexts end up close together.
- dense retrievaldense retrieval, n.: Finding relevant documents by turning the question and every document into lists of numbers that capture meaning, then picking the documents whose numbers are closest to the question's.
- sparse retrievalsparse retrieval, n.: Finding documents by the words they contain, giving more weight to rare, telling words — the classic way search engines work.
Sources
- Salton, G., Wong, A. & Yang, C. S. (1975). A vector space model for automatic indexing. Communications of the ACM 18(11): 613–620.
- Dubin, D. (2004). The Most Influential Paper Gerard Salton Never Wrote. Library Trends 52(4): 748–764.
Mentioned in
bag of wordsbag of words, n.: A simple way of turning text into numbers by counting how often each word appears and ignoring word order — like describing a recipe only by its shopping list., information retrievalinformation retrieval, n.: The study and engineering of search: how to organize huge piles of documents so that the few that matter to someone's question can be found and ranked quickly. AI systems rely on it to look things up., TF-IDFTF-IDF, abbr.: A simple score for picking out the words that characterize a document: a word counts for more the more often it appears there, and for less the more of the collection's other documents also use it.