System and method for identifying semantically relevant documents

A system for identifying semantically relevant documents, comprising:

• an index comprising semantic index key terms from one or more documents;
• a query module to obtain a query and to identify semantic key terms of the query, wherein each of the query key terms is associated with information comprising use of the word and a grammatical role of the word;
• and
• a search module to look up the each of the query key terms and variants of the query key terms in the index based on the information;
• an identification module to identify one or more of the documents in the index as possible relevant candidates based on the look up;
• a candidate set generator to generate a potential match candidate set by filtering the possible relevant candidates;
• and
• a document matching module to identify at least one of the documents that match the query by comparing a semantic representation for each filtered potential match candidate in the set with a semantic representation for the query, wherein each of the semantic representation for the query and the semantic representation for each potential match candidate in the set comprises one or more semantic substructures, comprising:
• a comparison module to align the substructures for the query semantic representation and one or more of the semantic representations for the potential match candidates;
• and
• a scoring module to assign a score to the one or more potential match candidates by identifying those substructures that are shared with the query semantic representation.