Method and system to index captioned objects in published literature for information discovery tasks
View Patent ↗The present invention relates to the identification, extraction, linking, storage and provisioning of data that constitute the captioned components of published or “print ready” literature for computerized information discovery activities including search, browse and data mining. These components, or objects, include the tabular presentation of data (“tables”) and graphics such as “figures”, “images” and “illustrations” typically used to supplement the textual narrative of the publication.
1. An automated method of processing at least one document containing at least one object, comprising:
extracting data from each of the objects in each of the at least one document into an object record;
linking each object record, including the extracted data, with its associated one of the at least one document including generating a document record; and
storing each linked object record in a computer readable medium;
wherein the extracting includes extracting predetermined attributes associated with the at least one object responsively to extraction rules specifying types of the predetermined attributes to be identified and extracted; and
wherein the linking provides links between object records having a same value of a respective type of the predetermined attributes, the linking further including generating an index.
2. A system for collecting data from a plurality of objects contained in a plurality of documents, the system comprising:
a computer;
means in the computer for extracting data from each of the plurality of objects into an object record;
means in the computer for linking the data extracted from each of the plurality of objects with an associated one of the plurality of documents including generating a document record; and
means in the computer for storing each linked object record, including the extracted data;
wherein the extracting includes extracting predetermined attributes associated with each of the plurality of objects responsively to extraction rules specifying types of the predetermined attributes to be identified and extracted; and
wherein the linking provides links between object records having a same value of a respective type of the predetermined attributes, the linking further including generating an index.
3. An automated method of processing at least one document containing at least one object, comprising:
providing extraction rules on a computer readable medium, the extraction rules being specific to respective object types and defining specific attributes to be extracted with the at least one object;
the at least one object including a figure or table, the figure or table having identifiers of an independent variable and a dependent variable, the identifiers being specific attributes defined by the extraction rules to be extracted with the at least one object;
the extraction rules defining the identifiers as attributes to be extracted;
extracting data from each of the objects in each of the at least one document into an object record to generate extracted data, the extracting including interpreting and applying the extraction rules by means of a processor;
linking each object record, including the extracted data, with its associated one of the at least one document including generating a document record, the linking including providing a link between object records where a same type of the specific attributes has a same value, the linking further including generating an index; and
storing each linked object record in a computer readable medium.
4. The method of claim 3 , further comprising:
implementing a display process on a computer to:
display data in the object record;
navigate between the display of the at least one object record and the associated one of the at least one document; and
navigate between object records having matching values of the identifiers.
5. The method of claim 3 , wherein the extracting further includes incorporating in the object record extracted data representing the identifiers of the independent and dependent variables and the object.
6. An automated method of processing at least one document containing at least one object, comprising:
extracting data from each of the at least one object into a corresponding object record;
linking each object record and the extracted data contained therein, with its associated one of the at least one document;
storing each linked object record in a computer readable medium;
wherein the at least one object includes a figure;
wherein the extracting includes identifying in accordance with rules specifying subject-specific predetermined user-defined object types, the at least one object from which data is to be extracted;
wherein the extracting includes extracting, from each identified object, values corresponding to predetermined attributes;
wherein the linking includes associating each object record with an abstract record, a full-text record, and a source record from which the object was extracted;
wherein the linking is effective to permit other object records, abstract records and/or full-text records to be found based on extracted attribute-specific values.