BIASVAL
The BIASVAL
field specifier (case sensitive) allows you to bias the score of result documents by a specified percentage, based on whether they include a specific value in the specified field.
NOTE: You can optimize the speed of this field specifier by restricting the field to one of the following configuration options:
-
, when the configuration parameter is set to
True
.
Format
FieldText=BIASVAL{fieldValue,percentage,fuzzydistance}:fieldName
fieldValue
|
The value that the specified field must contain to increase or decrease the weight by the |
percentage
|
A percentage in the range -100 to 100. The percentage (positive or negative) by which to boost the relevance of documents that have the value When you set to |
fuzzydistance
|
(Optional) An integer that determines a fuzzy matching distance. If you set a fuzzy distance, IDOL applies a scaled weighting boost to documents where the field value for the hit document is within this fuzzy distance of the query value. The fuzzy distance corresponds to the |
fieldName
|
The name of the field that must be present (and must contain the proper value) for the document relevance to be boosted. |
NOTE: BIASVAL
requires two arguments. BIASVAL{percentBoost}:fieldName
is not valid syntax.
Examples
FieldText=BIASVAL{Cambridge,10}:TOWN
A document whose TOWN
field value is Cambridge
has its weighting increased by 10% (although the weighting cannot go over 100%).
FieldText=BIASVAL{Cambridge,10}:TOWN+BIASVAL{London,-10}:TOWN
A document whose TOWN
field value is Cambridge
has its weighting increased by 10%, and a document whose TOWN
field value is London
has its weighting decreased by 10%. This example shows that biases can be negative as well as positive, and that you can include more than one set of parameters with the specifier.
FieldText=BIASVAL{,-10}:COLOUR
A document that does not have the COLOUR
field, or that has a COLOUR
field with an empty value, has its weighting decreased by 10%.
FieldText=BIASVAL{Cambridge,10,2}:TOWN
A document whose TOWN
field value is Cambridge
has its weighting increased by 10% (although the weighting cannot go over 100%). If the TOWN
field value contains a value that is similar to Cambridge
(within a fuzzy distance of 2), a reduced weighting boost is applied.