SNOMED CT Spell Checker

1. Search with Dynamically Estimated Parameters.

   OR    Click on a term below to try one of the suggested examples:
Bankreatitis
Punkreatitis
Verteprate
Vertepratee
Sodeum floride
Sodeum Florida
Ischemia
Fracture of the Femur
Broken Leg
Ammonium
Haemophilia

Parameter alpha is estimated by the formula
The search is limited to the SNOMED CT terms meeting the following criterion:
The search returns SNOMED CT terms below the threshold: LACP(S,T) < 0.1

2. Search with Static Parameters.

   OR    Click on a term below to try one of the suggested examples:
Bankreatitis
Punkreatitis
Verteprate
Vertepratee
Sodeum floride
Sodeum Florida
Ischemia
Fracture of the Femur
Broken Leg
Ammonium
Haemophilia

Parameter alpha is set to 1
The search is limited to the SNOMED CT terms meeting the following criterion:
The search returns SNOMED CT terms below the threshold LACP(S,T) < 0.1

3. Search with User-Defined Parameters.

Input Query:
Parameter alpha:
Parameter a:
Parameter b:
LACP Threshold:
   OR    Click on a term below to try one of the suggested examples
with the following parameter values:
alpha=2, a=3, b=3, LACP_Treshold=0.1:
Bankreatitis
Punkreatitis
Verteprate
Vertepratee
Sodeum floride
Sodeum Florida
Ischemia
Fracture of the Femur
Broken Leg
Ammonium
Haemophilia

Parameter alpha defines the number of allowed mismatches in histogram difference:
Parameters a and b limit the allowed length of SNOMED CT terms:
The LACP treshold narrows the interval of the allowed LACP values: LACP(S,T) < LACP_Treshold

About the LACP Function.

The LACP method is based on an approximate histogram match of string prefixes. It determines the similarity value of a pair of strings. The method compares the histogram difference of the prefixes of two strings to a threshold parameter alpha. This search begins from the first character of the strings. The prefix length is returned when the histogram difference is equal to alpha or the last character of the shorter string is reached. Then the obtained value of the prefix length is divided by the average length of the pair of strings. The division is implemented to take into consideration string lengths, since strings that have lengths that vary significantly have a higher degree of dissimilarity compared to strings that do not. The division by the average string lengths also assures that the value of the LACP function stays in the [0, 1] interval.

About the SNOMED CT Spell Checker.

The interactive online spell checker for SNOMED CT terms is based on the LACP method (see (1) above). The spell checker is a PHP script, which connects to a MySQL database containing the full set of unduplicated SNOMED CT terms extracted from the 2009AB edition of the UMLS knowledge source. The spell checker accepts an input query and interactively outputs the SNOMED CT terms satisfying the specified initial parameters. The spell checker returns a record set containing a Concept Unique Identifier (CUI) of a term, a term itself, a value of the parameter alpha (formula (2) above), and a value of LACP(S,T) defined in formula (1).