Book Image

ElasticSearch Blueprints

Book Image

ElasticSearch Blueprints

Overview of this book

Table of Contents (15 chapters)
Elasticsearch Blueprints
About the Author
About the Reviewer

How scoring works

Scoring is the process where we determine how well a query matches against a document. This can be based on the relevance of the query against the text it matches or based on a logic that we are interested in. For example, during a restaurant search on a site like Zomato, the closer the restaurant is, the better the match. In another instance, in a news search, a mix of recent news and a specific topic would give a better match.


Scoring can be based on the query or on a custom logic.

We will look at how custom logic can be introduced in scoring in the later part of this chapter. In this section, we will discuss how scoring is done on text based on the query text.

There are three main factors that determine the score of a query against a text field, which are as follows:

  1. Term frequency: Let's assume that we are searching for the word elephant, we get the following results:

    • Document-1: There are two types of elephants, African elephant and Indian elephant.

    • Document-2: Elephants...