Method and system for retrieving documents via an electronic data file
AssigneeHet Babbage Instituut voor Kennis en Informatie Technologie “B.I.K.I.T.”
InventorFernand Jozef VanDamme
Life of this patent
Estimating the term…
Abstract
A method for retrieving documents via an electronic data file, characterized in that keys are used for the retrieval which find one or several relations between the textual data of the documents concerned.
Claims16
Method for retrieving documents via an electronic data file, wherein said method comprises at least a step wherein keys are used for the retrieval which find one or several relations between textual data of the documents, comprising the step of:
- indexing said documents electronically as a function of the relations occurring in the textual data of the documents in order to retrieve the documents,
- said step of indexing said documents comprising the steps of:
- filtering the textual data to obtain explicit index terms by eliminating expletives and by determining explicit index terms by means of unigrams, bigrams, trigrams, or combinations of unigrams, bigrams, and trigrams occurring in the text;
- using a thesaurus to add implicit index terms to the explicit index terms, said implicit index terms being broader than the explicit index terms;
- and
- filtering the implicit index terms based on a frequency of occurrence of the explicit index terms,
- wherein, for the relations, use is made of those relations existing between the generated index terms, which relations are automatically searched for on the basis of a list of terms containing a relation indication,
- wherein use is made of postulated frames which are composed of catchwords and of voice components, including the terms containing a relation indication, and wherein the text of the documents is examined for said frames, and
- wherein use is made of said frames with catchwords defining at least three arguments, each one of said at least three arguments representing a cause, an effect and a situation or condition respectively.
Method for retrieving documents via an electronic data file, wherein said method comprises at least a step wherein keys are used for the retrieval which find one or several relations between textual data of the documents, comprising the step of:
- indexing said documents electronically as a function of the relations occurring in the textual data of the documents in order to retrieve the documents,
- said step of indexing said documents comprising the steps of:
- filtering the textual data to obtain explicit index terms by eliminating expletives and by determining explicit index terms by means of unigrams, bigrams, trigrams, or combinations of unigrams, bigrams, and trigrams occurring in the text;
- and
- using a thesaurus to add implicit index terms to the explicit index terms, said implicit index terms being broader than the explicit index terms;
- and
- wherein said relations are causal relations between the textual data of the documents,
- wherein, for the relations, use is made of those relations existing between the generated index terms, which relations are automatically searched for on the basis of a list of terms containing a relation indication,
- wherein use is made of postulated frames which are composed of catchwords and of voice components, including the terms containing a relation indication, and wherein the text of the documents is examined for said frames, and
- wherein use is made of said frames with catchwords defining at least three arguments, each one of said at least three arguments representing a cause, an effect and a situation or condition, respectively.
Dependencies are inferred from the claim text.