OpenSearch - Hybrid search in OpenSearch Vector Search Edition
Jan 08 2024
OpenSearchContent
Intended customers: 1. Customers who need to search for text by using semantic search 2. Customers who need to search for text by using keyword search 3. Customers who need to improve the performance of text search 4. Customers in scenarios such as the content community, media, conversational search in a knowledge base, and e-commerce shopping guide New features or specifications: The hybrid search feature combines semantic search with keyword search to improve the performance of text search. You can generate sparse vectors and dense vectors based on text data and import the generated vector data into OpenSearch Vector Search Edition. OpenSearch combines a dense vector and a sparse vector into a single vector. Dense vectors are used for semantic search, whereas sparse vectors are used to match keywords and express the frequency of words. This way, you can use sparse vectors and dense vectors to improve the performance of text search.