Retrieval device and method

• US 9,659,004 B2
• Filed: 09/08/2014
• Issued: 05/23/2017
• Est. Priority Date: 09/17/2013
• Status: Active Grant

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Okura, Seiji, Ushioda, Akira
• Primary Examiner(s): Woo, Isaac M
• Application Number:
  US14/480,237
• Publication Number:
  US 20150081715A1
• Time in Patent Office: 988 Days
• Field of Search: 707600-899
• CPC Class Codes: G06F 16/313 Selection or weighting of terms for indexing
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A processor performs semantic analysis on a query and generates one or more semantic structures where each structure is expressed by a graph. The processor generates retrieval keys corresponding to combinations of nodes connected directly or indirectly in the semantic structures, in addition to retrieval keys corresponding to minimum units of semantic connections between nodes in the generated semantic structures. The processor retrieves relevant documents whose sentences are matched to combinations of nodes, by using the generated retrieval keys, in the semantic structures stored in an index for retrieval on a database storing the documents.

Claims

1. A retrieval device comprising:

a processor that

performs semantic analysis on a query and generates one or more semantic structures where each structure is expressed by a graph;

generates retrieval keys corresponding to combinations of nodes connected directly or indirectly in the semantic structures and corresponding to minimum units of semantic connections between nodes in the generated semantic structures; and

retrieves relevant documents whose sentences are matched to combinations of nodes, by using the generated retrieval keys, in the semantic structures stored in an index for retrieval on a database storing the documents.

2. The retrieval device according to claim 1, wherein

the processor generates retrieval keys in which arcs existing in the combinations of the nodes connected indirectly are OR-connected.

3. The retrieval device according to claim 2, wherein

the processor performs matching with the combinations of the nodes in the index for retrieval including each of the OR-connected arcs corresponding to the combinations of the nodes in the retrieval keys.

4. The retrieval device according to claim 1, wherein

the processor generates retrieval keys in which condition designation is performed by a wildcard to one node of the combinations of the nodes connected indirectly and a node existing in the combinations.

8. The retrieval device according to claim 1, wherein

the processor defines an indirect connection by an attribute of the node and an attribute of an arc that starts from the node.

5. A retrieval device comprising:

a processor that

performs semantic analysis on a query and generates one or more semantic structures where each structure is expressed by a graph;

generates retrieval keys corresponding to combinations of nodes connected directly or indirectly in the semantic structures and corresponding to minimum units of semantic connections between nodes in the generated semantic structures; and

retrieves relevant documents whose sentences are matched to combinations of nodes, by using the generated retrieval keys, in the semantic structures stored in an index for retrieval on a database storing the documents, wherein

the processor groups arcs representing a similar relationship from among arcs corresponding to the combinations of the nodes in the retrieval keys and generates the retrieval keys.

6. The retrieval device according to claim 5, wherein

the processor performs matching with the combinations of the nodes in the index for retrieval including each arc belonging to the grouped arcs corresponding to the combinations of the nodes in the retrieval keys.

7. The retrieval device according to claim 5, wherein

the combinations of the nodes in the index for retrieval is associated with the grouped arcs, and

the processor performs matching with the combinations of the nodes in the index for retrieval including the grouped arcs corresponding to the combinations of the nodes in the retrieval keys.

9. A retrieval system comprising:

a first computer; and

a second computer connected to the first computer via a network, wherein

the first computer includes a first processor that

performs semantic analysis on a query and generates one or more semantic structures where each structure is expressed by a graph, and the second computer includes a second processor that

generates retrieval keys corresponding to combinations of nodes connected directly or indirectly in the semantic structures and corresponding to minimum units of semantic connections between nodes in the semantic structures generated by the first computer, and

retrieves relevant documents whose sentences are matched to combinations of nodes, by using the generated retrieval keys, in the semantic structures stored in an index for retrieval on a database storing the documents.

10. A retrieval method comprising:

performing, by a processor, semantic analysis on a query and generating one or more semantic structures where each structure is expressed by a graph;

generating, by the processor, retrieval keys corresponding to combinations of nodes connected directly or indirectly in the semantic structures generated by the semantic analysis and corresponding to minimum units of a semantic connection between nodes in the generated semantic structures; and

retrieving, by the processor, relevant documents whose sentences are matched to combinations of nodes, by using the generated retrieval keys, in the semantic structures stored in an index for retrieval on a database storing the documents.

11. A computer-readable recording medium having stored therein a program for causing a computer to execute a process comprising:

performing semantic analysis on a query and generating one or more semantic structures where each structure is expressed by a graph;

generating retrieval keys corresponding to combinations of nodes connected directly or indirectly in the semantic structures generated by the semantic analysis and corresponding to minimum units of a semantic connection between nodes in the generated semantic structures; and

retrieving relevant documents whose sentences are matched to combinations of nodes, by using the generated retrieval keys, in the semantic structures stored in an index for retrieval on a database storing the documents.