Method and apparatus for constructing a query based upon concepts associated with one or more search terms

摘要

A method and apparatus are provided to efficiently generate a fulsome query in order to increase the recall and/or precision provided by the search. A method may construct a query by receiving the one or more initial search terms and then defining a concept for each search term. In order to define a concept, the method may determine if a concept associated with a respective search term has been previously defined. In an instance in which a concept associated with a respective search term has been previously defined, the method at least initially utilizes the previously defined concept. However, in an instance in which a concept associated with a respective search term has not been previously defined, the method constructs the concept based on terms related to the respective search term. The method may then combine the concepts defined for the one or more search terms to generate the query.

主权项

1. A method for constructing a query comprising one or more initial search terms, the method comprising:
receiving the one or more initial search terms; defining a concept for each initial search term with processing circuitry, defining a concept comprising:
determining if a concept associated with a respective search term has been previously defined;when a concept associated with a respective search term has been previously defined, at least initially utilizing the previously defined concept; when a concept associated with a respective search term has not been previously defined, constructing the concept based on terms related to the respective search term, wherein constructing the concept comprises, prior to conducting a search of one or more data sources, identifying a plurality of terms related to the respective search term based upon string matching within sample data and constructing the concept associated with the respective search term incorporating at least some of the terms related to the respective search term that were identified based upon string matching, wherein the string matching is performed in accordance with a predefined string matching algorithm that identifies a term to be related to the respective search term comprising a plurality of letters when the term includes a plurality of the same letters or letter sequences as the search term even though the term also exhibits one or more differences including transposed letters relative to the search term; and combining the concepts defined for the one or more initial search terms to generate the query; and following construction and combining of the concepts, conducting the search of one or more data sources in accordance with the query in order to identify one or more records that satisfy the query.