CODEWITHSUNDEEP
elasticsearchelasticsearch-pluginelasticsearch-analyzers
Shawn
Shawn

Reputation: 331

How to improve Elasticsearch query with ML/NLP?

I am currently using a fairly standard query with my Elasticsearch search. The only addition I am using is the metaphone analyzer. I wanted to know whether there are any in-built NLP or ML add-ons for elasticsearch. I am slightly out of my depth, but if anyone can point me to some resources, it would be of great help!

Upvotes: 1

Views: 1243

Answers (2)

Alex16237
Alex16237

Reputation: 199

You already use some of NLP by using standard mapping of Elasticsearch. Tokenization, stemming & stop words, all of those are language specific and use NLP for parsing.

Built-in analyzers

When it comes to search engine building, what you probably have in mind is 'better similarity' and more precise retrieval.

ES uses TF-IDF model for similarity by default (which is an NLP/ML method in itself)

Modules Similarity

You could definitely throw NN at it, but there isn't any specific guide for doing so. I can recommend book by Tommaso Teofili, "Deep Learning for Search" which actually describes all that you could do with ML in context of building search engine.

Upvotes: 2

Nirmal
Nirmal

Reputation: 1336

You might be interested in LearnToRank plugin - Its very interesting if you are about 'relevancy' .

Also this plugin , to enrich document during ingest using OpenNLP

Upvotes: 1

Related Questions