Implementation of basic vector space model
The query processing of VSM is quite tricky, you need of optimize every aspect of computation. The high-dimensional vector product and similarity values of query (q) and documents (d) need to optimized.
Basic Assumption for Vector Space Model (VSM) Retrieval Model
- Simple model based on linear algebra. Terms are considered as features using a weighting scheme.
- Allows partial matching of documents with the queries. Hence, able to produce good institutive scoring. Continuous scoring between queries and documents.
- Ranking of documents are possible using relevance score between document and query.