Tuning Elasticsearch for English-Language Precision
summary
Shows how a real dictionary index of about a million entries is made findable with Elasticsearch 5.5 by building and testing concrete pipelines: preprocessed keyword fields for exact matches, layered multi-fields analyzed with a whitespace tokenizer, character filters (HTML strip, quote normalization, regexp punctuation stripping), token filters (lowercase, English possessive, stemmer), and shingles. Demonstrates query-time multi_match scoring, use of explained to inspect tf‑idf math, and practical tradeoffs between precision and recall.
AI-written summary · about
chapters (5)
related talks
20mMultilingual Search System - How to Build and Improve!?
46mFull-Text Search Internals with Elasticsearch
37mBeyond the basics with Elasticsearch
55mFull-Text Search Demystified
41mBeyond the basics with Elasticsearch
37mAlgorithms & data-structures that power Lucene & ElasticSearch
26mFrom __icontains to search
38mElasticSearch in action