Fuzzy search in the elasticsearch gives matches with incorrect order - elasticsearch

Question

I am trying to build an engine where we can match areas mentioned in the address with the list available in elasticsearch.
I am using this query to search areas similar to "iit".
My query is :
{
"query": {
"fuzzy": {
"locality": {
"value": "iit",
"fuzziness": 1
}
}
},
"highlight": {
"fields": {
"locality": {}
}
}
}
I am getting below results :
{
"took": 4,
"timed_out": false,
"_shards": {
"total": 5,
"successful": 5,
"failed": 0
},
"hits": {
"total": 2,
"max_score": 2.1290483,
"hits": [
{
"_index": "geocoding_1",
"_type": "localities",
"_id": "AVuzRiZ04pEQsZFpK6A_",
"_score": 2.1290483,
"_source": {
"locality": [
"emerald isle ii"
]
},
"highlight": {
"locality": [
"emerald isle <em>ii</em>"
]
}
},
{
"_index": "geocoding_1",
"_type": "localities",
"_id": "AVuzRfof4pEQsZFpK59H",
"_score": 1.877402,
"_source": {
"locality": [
"iit - bombay",
"iitb",
"indian institute of technology - bombay"
]
},
"highlight": {
"locality": [
"<em>iit</em> - bombay",
"<em>iitb</em>"
]
}
}
]
}
}
Because "iit" is directly available in the 2nd document and hence I was expecting that to be returned as best match with highest score. What is the change that I should make so that I get 2nd document with highest score.
I am using ES 2.3.4 .

Answer 1

If you also are interested in exact matching to score better, I always suggest a bool with should statements and adding a match or term query in there. In this way the combined scores will favor the exact matching:
{
"query": {
"bool": {
"should": [
{
"fuzzy": {
"locality": {
"value": "iit",
"fuzziness": 1
}
}
},
{
"match": {
"locality": "iit"
}
}
]
}
},
"highlight": {
"fields": {
"locality": {}
}
}
}