elasticsearch 排序 算法
如何将tfidf映射到[1,3] 区间内
（1）建立映射x→arctanx,此为一一映射,把全体实数R映射到开区间（-π/2,π/2）
（2）开区间（-π/2,π/2）到开区间（0,1）只需要一个线性映射就可以,设为f（x）=ax+b,则有f（-π/2）=1且f（π/2）=3,解得a=2/π,b=2
综合一下,映射x→2（arctanx）/π+2,即把全体实数映射到（1,3）
所以表达式可以写成
"script_score" : {
"script" : "return atan(_score.doubleValue())*2/3.14159+2"
}
为了只使用函数值 设置 "boost_mode": "replace"，
所以整个请求如下
curl -XGET 'http://localhost:9200/alias-product/product/_search?pretty&explain' -d '{
"size" : 1,
"query" : {
"function_score" : {
"query" : {
"bool" : {
"should" : [ {
"match" : {
"_all" : {
"query" : "关键字",
"type" : "boolean",
"operator" : "AND",
"boost": 10
}
}
}, {
"match" : {
"company_name" : {
"query" : "关键字",
"type" : "boolean",
"operator" : "AND",
"boost": 10
}
}
} ],
"minimum_should_match" : "1"
}
},
"functions" : [
{
"script_score" : {
"script" : "return atan(_score.doubleValue())*2/3.14159+2"
}
},{
"filter" : {
"match" : {
"company_name" : {
"query" : "关键字",
"type" : "boolean",
"operator" : "AND"
}
}
},
"weight" : 2
}
],
"score_mode" : "sum",
"boost_mode": "replace"
}
},
"aggregations" : {
"agg" : {
"terms" : {
"field" : "member_id",
"size" : 0,
"order" : {
"top_hit" : "desc"
}
},
"aggregations" : {
"top_hit" : {
"max" : {
"script" : {
"inline" : "_score"
}
}
}
}
}
}
}'
