How to boost scoring for whole word hits over substring hits

Elastic Stack Elasticsearch
1

Hi there,

I'd like to make it so that single-word query results that contain the entire query word by itself get higher scoring over results that contain the query word within another word.

For example in the case I am testing, when I search for 'pet', the result scoring does not seem to differentiate between cases containing 'pet' vs. 'appetite'. I would like to make it so that the phrase 'I have a pet allergy' would always return a higher score then 'I have no appetite'.

My settings look like this:

"settings": {
      "index": {
        "number_of_replicas": 2,
        "max_ngram_diff": 7,
      },
      "analysis": {
        "normalizer": {
          "my_normalizer": {
            "type": "custom",
            "char_filter": [],
            "filter": [
              "lowercase"
            ]
          }
        },
        "char_filter": {
          "my_char_filter": {
            "type": "mapping",
            "mappings": [
              "- => _"
            ]
          }
        },        
        "analyzer": {
          "my_analyzer": {
            "tokenizer": "standard",
            "char_filter": [
              "my_char_filter"
            ],
            "filter": [
              "lowercase",
              "stop",
              "asciifolding",
              "my_ngram",
              "kstem"
            ]
          },
        "filter": {
          "my_ngram": {
            "type": "ngram",
            "min_gram": 4,
            "max_gram": 9,
            "preserve_original": true
          },
        }
      }
    },

Is this behaviour something that should be set in the index or the query itself?

Thank you in advance for any help!

2

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.