Random_score , how to moderate its impact?

Elastic Stack Elasticsearch
1

Hi,

I use the following query, in this one I have 2 functions:

  • distance_feature : compute a score based on distance
  • random_score : compute a value between 0 and 1

The final score = distance_feature * random_score

The problem is that I have such situation:

distance_feature = 0.9
random_score = 0.001

=> score = 0.0009

and so, something that is near of the origin point, will be put at the end of the results !
The random_score 'absord' the distance_feature !

I'm looking for a way to compute something like this, in order to moderate its impact:

distance_feature  + (distance_feature * 0.2) * random_score

The result will be something that vary between 100% and 120%:

0.9 + (0.9 * 0.2) * 0.0009 = 0.9009

PS: noticed that I wish to use it with pagination (So, using script_score does seems possible without having some duplicates).

Here is my query, for example:

GET /_search
{
"explain": true,
"query": {
"function_score": {
"query": {
"bool": {
"should": [
{
"distance_feature": {
"field": "geoLocations",
"pivot": "20km",
"origin": [-71.3,41.15]
}
}
]
}
},
"random_score": {},
"boost": "1",
"boost_mode": "multiply"
}
}
}

2

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