Database aliasing in information access system
AssigneeInternational Business Machines Corporation
InventorsJane Wen Chang, Raymond Lau, Michael Kyle McCandless
Life of this patent
Estimating the term…
Abstract
A method of database aliasing is provided. A method for aliasing a collection of data includes extracting names and associated terms from the collection of data, generating normalized text from the names and associated terms, parsing the normalized text in conjunction with a grammar to produce meaning fragments and mapping the meaning fragments to a collection of concepts.
Claims19
A computer-implemented method for aliasing a collection of data, comprising:
- extracting, by a computing device, names and associated terms from a collection of data;
- normalizing, by the computing device, the extracted names and associated terms;
- parsing, by the computing device, the normalized names and associated terms in conjunction with a grammar to produce meaning fragments, the grammar representing how words and their component parts combine to form sentences;
- and
- mapping, by the computing device, the meaning fragments to a collection of concepts stored in an aliasing file, the concepts representing general ideas inferred or derived from specific instances, wherein alias entries are stored in the aliasing file when a determination is made that multiple normalized names and associated terms map to a single concept, and the aliasing file is used for resolving variations in the extracted names and associated terms during the normalizing step.
An article comprising:
- a storage medium having stored thereon instructions that when executed by a computer cause the computer to:
- extract names and associated terms for a collection of data;
- normalize the extracted names and associated terms;
- parse the normalized names and associated terms in conjunction with a grammar to produce meaning fragments, the grammar representing how words and their component parts combine to form sentences;
- and
- map the fragments to a collection of concepts stored in an aliasing file, the concepts representing general ideas inferred or derived from specific instances, wherein alias entries are stored in the aliasing file when a determination is made that multiple normalized names and associated terms map to a single concept, and the aliasing file is used for resolving variations in the extracted names and associated terms during the normalize step.
A method of information processing comprising:
- generating, by a computing device, a database file;
- generating, by the computing device, an aliasing file in conjunction with a grammar for the database file, the aliasing file containing mappings of multiple normalized content to a single concept, the grammar representing how words and their component parts combine to form sentences;
- and
- applying, by the computing device, the aliasing file to a query directed to the database file for resolving variations in the query during normalization.
A method of information processing comprising:
- generating, by a computing device, a database file;
- extracting, by the computing device, names from the database file;
- normalizing, by the computing device, the extracted names;
- applying a language lexicon to the normalized extracted names, the language lexicon specifying a list of words along with their normalized forms;
- storing, by the computing device, multiple normalized extracted names that map to a single concept in an aliasing file, the single concept representing a general idea inferred or derived from specific instances;
- and
- applying, by the computing device, the aliasing file to a query directed to the database file for resolving variations in the query during the normalizing step.
An article comprising:
- a storage medium having stored thereon instructions that when executed by a machine result in the following:
- generate a database file;
- generate an aliasing file in conjunction with a grammar for the database file, the aliasing file containing mappings of multiple normalized content to a single concept, the grammar representing how words and their component parts combine to form sentences;
- and
- apply the aliasing file to a query directed to the database file for resolving variations in the query during normalization.
An article comprising:
- a storage medium having stored thereon instructions that when executed by a machine result in the following:
- generate a database file;
- extract names from the database file;
- normalize the extracted names;
- apply a language lexicon to the normalized extracted names, the language lexicon specifying a list of words along with their normalized forms;
- store multiple normalized extracted names that map to a single concept in an aliasing file, the single concept representing a general idea inferred or derived from specific instances;
- and
- apply the aliasing file to a query directed to the database file for resolving variations in the query during the normalize step.
Dependencies are inferred from the claim text.