|
What is matching? - functional
Matching is a type of searching, where in addition to the data, knowledge about the data is used
to find the best search result. A dating-site example is used to explain all the technologies.
Search criteria
The candidates indicate their own features (age, sex, eye color, etc) and also indicate the
features that they are looking for (age category, hobbies, minimum education level, etc).
| Criteria |
Value |
Type |
| Gender |
Male |
Offered |
| Age |
26 |
Offered |
| Education |
College |
Offered |
| Gender |
Female |
Requested |
| Age category |
20-30 |
Requested |
| Education |
College |
Requested |
Match rules
The candidates can be compared by match rules.
A match rule can match one or more features of a
candidate with one or more features of all other candidates.
| Match rule name |
Offered criterion |
Requested criterion |
| Age match rule |
Age |
Age category |
| Gender match rule |
Gender |
Gender |
| Education match rule |
Education |
Education |
A detailed description of all available match rules is available in
the Match Rules tutorial that is available on the documentation page.
Weight per match rule
To every match rule a weight can be attached that indicates how much the
match rule counts in the end result. Someone can indicate that it is very
important that the age is within the age category range.
| Match rule |
Weight |
| Age match rule |
80% |
| Education match rule |
20% |
Gliding scale
In real time situations it often happens that a match rule is not exactly matched.
For example some one is looking for a partner with the age 25. Is a person with
the age of 24 or 26 not suited ? With gliding scale match rules it can be determined that
the ideal age is 25, but that 26 matches for 90%.
This is a very powerful method because it will also find results that do not match for 100%
but are still acceptable results. Without gliding scale technology a lot of potential results will be missed.
Matrix matching
For a lot of discrete domains, matrices can be used.
For example: education. College level is requested, but high school may match
for 75%.
Set matching
Set matching can be used for example to match sets of hobbies or interests.
Running
Skating
Gardening
Sailing
Reading
|
 |
Running
Skating
Cooking
Reading
|
60% match (3 out of 5 match)
Minimal score
This indicates that a match result is excluded from the result list
when the match percentage is less then the minimal required match score.
This applies to both the total result score as the match percentage of
an individual matchrule.
| Match rule |
Minimal score |
| Gender match rule |
100% |
| Minimal total score |
80% |
The example indicates that only candidates are added to the result list that
match for 100% with the gender rule (must be 100% female) and that have
a minimal total score of 80%.
Two-way matching
By default a matchrule compares how well a candidate matches with the request.
Two-way matching also determines how well the request
matches to the requirements of the candidates.
For example a 26 year old man is looking for a woman in the
20-30 age category. Two-way matching also matches the requested age
of candidate with the offered age of the request.
So a woman that is in the 20-30 age category does not match if she
is looking for a man with a minimal age of 30.
Custom match rules
The match rules determine the quality of the search result.
The power of Match4J exists in you being able to configure the default
match rules or add custom match rules.
These match rules can be developed by yourself or by Match4J experts.
|
