Attribute extraction apparatus and attribute extraction method
Among pieces of attribute information contained in a document, user-designated attribute information and other attribute information that may give additional value to the user-designated attribute information are accurately and efficiently presented. An attribute extraction apparatus includes a storage section that stores a document; and a computing section that executes a process for calculating a predetermined index on the basis of the document's structure with respect to a relationship between a plurality of pieces of attribute information described in the document, and a process for calculating an importance degree of each piece of the attribute information other than predetermined attribute information for the predetermined attribute information on the basis of at least one of the index and a predetermined statistical property with respect to a relationship between the predetermined attribute information and the attribute information other than the predetermined attribute information among the plurality of pieces of the attribute information.
1. An attribute extraction apparatus comprising:
one or more processors and one or more non-transitory computer-readable storage media, the non-transitory computer-readable storage media having stored thereon at least:
instructions for executing a process for calculating a predetermined index on a basis of a document structure of a document with respect to a relationship between a plurality of pieces of attribute information described in a stored document,
a process for calculating an importance degree of each piece of the attribute information other than predetermined attribute information with respect to predetermined attribute information on a basis of at least one of the index and a predetermined statistical property with respect to a relationship between the predetermined attribute information and the attribute information other than the predetermined attribute information among the plurality of pieces of the attribute information, and extracting portions of the document based on the importance degree of each piece of the attribute information other than predetermined attribute information, wherein calculating the importance degree is made at least on a basis of an absolute appearance position of each piece of the attribute information in the document, and wherein the attribute information comprises an attribute name and an attribute value, and wherein the attribute name and the attribute value are terms in the document.
2. The attribute extraction apparatus according to claim 1 , the non-transitory computer-readable storage media having stored thereon instructions for
calculating an association degree between the pieces of attribute information and the importance degree of each piece of the attribute information on a basis of a feature of the document structure at a time of calculating the index.
3. The attribute extraction apparatus according to claim 2 , wherein
the non-transitory computer-readable storage media having stored thereon instructions for
calculating the association degree on a basis of co-occurrency and an appearance distance between the pieces of attribute information.
4. The attribute extraction apparatus according to claim 2 , the non-transitory computer-readable storage media having stored thereon instructions for
calculating the importance degree on a further basis of at least one of an appearance frequency of each piece of the attribute information in the document, and a position relationship with a preset specific word.
5. The attribute extraction apparatus according to claim 1 , the non-transitory computer-readable storage media having stored thereon instructions for
narrowing down attribute information for which the importance degree is to be calculated, on a basis of the index.
6. The attribute extraction apparatus according to claim 1 , the non-transitory computer-readable storage media having stored thereon instructions for
setting user-designated attribute information as the predetermined attribute information, and
calculating the importance degree with respect to the relationship between the user-designated attribute information and the attribute information other than the user-designated attribute information.
7. The attribute extraction apparatus according to claim 1 , the non-transitory computer-readable storage media having stored thereon instructions for eliminating overlaps of same attribute information for which the index is to be calculated in a case in which a predetermined number or more of pieces of the same attribute information are present in the document.
8. An attribute extraction method comprising:
calculating a predetermined index on a basis of a document structure of a stored document with respect to a relationship between a plurality of pieces of attribute information described in the document;
executing a process for calculating an importance degree of each piece of the attribute information other than predetermined attribute information for the predetermined attribute information on a basis of at least one of the index and a predetermined statistical property with respect to a relationship between the predetermined attribute information and the attribute information other than the predetermined attribute information among the plurality of pieces of the attribute information; and
extracting portions of the document based on the importance degree of each piece of the attribute information other than predetermined attribute information,
wherein calculating the importance degree is made at least on a basis of an absolute appearance position of each piece of the attribute information in the document, and wherein the attribute information comprises an attribute name and an attribute value, and wherein the attribute name and the attribute value are terms in the document.
9. A non-transitory computer readable storage medium having a computer program recorded thereon an attribute extraction program comprising:
instructions for storing a document,
instructions for executing a process for calculating a predetermined index on a basis of a document structure of the document with respect to a relationship between a plurality of pieces of attribute information described in the document; and, instructions for executing a process for calculating an importance degree of each piece of the attribute information other than predetermined attribute information with respect to the predetermined attribute information on a basis of at least one of the index and a predetermined statistical property with respect to a relationship between the predetermined attribute information and the attribute information other than the predetermined attribute information among the plurality of pieces of the attribute information, and instructions for extracting portions of the document based on the importance degree of each piece of the attribute information other than predetermined attribute information, wherein calculating the importance degree is made at least on a basis of an absolute appearance position of each piece of the attribute information in the document, and wherein the attribute information comprises an attribute name and an attribute value, and wherein the attribute name and the attribute value are terms in the document.