Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
1999-12-30
2002-04-02
Black, Thomas (Department: 2171)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C707S793000, C707S793000, C707S793000, C704S007000, C704S009000, C704S010000, C382S177000, C382S306000
Reexamination Certificate
active
06366908
ABSTRACT:
TECHNICAL FIELD
The present invention relates to a keyfact-based text retrieval method and a keyfact-based text index method. In particular, the methods describe the formalized concept of a document as a pair comprising an object that is the head and a property that is the modifier, and uses the information described by the pair as index information for efficient document retrieval.
BACKGROUND OF THE INVENTION
A keyfact means an important fact contained in sentences which constitute a document. The keyfact is represented by an object and property information through syntactic analysis of the sentence.
The keyword-based text retrieval method was the main stream in conventional text retrieval methods. However, the precision of the keyword-based text retrieval method was not good due to the following reasons. First, the meaning of the document is not precisely represented and the representativeness of document expression is low because the document is represented by keywords, which are nouns. This is a fundamental reason for poor retrieval precision. Second, when a query includes a natural language phrase or a natural language sentence or keywords, the intention of the user's query is not reflected precisely in a keyword-based text retrieval method because the query is expressed by keywords. Therefore, the keyword-based text retrieval method has a fundamental limitation in retrieval precision because it performs document retrieval by keywords. As a result, because the keyword-based text retrieval system provides such low level of retrieval precision, it causes a number of unnecessary retrievals and therefore precious resources, such as time and effort, are wasted.
Recently, a number of studies have been performed in the area of phrase-based text retrieval methods in order to compromise such defects of the keyword-based retrieval method. The phrase-based text retrieval methods extract a precise phrase pattern through a morphological-syntactic normalization process and perform indexing and retrieval by extracted phrase. Therefore, the phrase-based retrieval method performs more precise text retrieval than the keyword-based text retrieval method but performs less precise text retrieval than a concept-based text retrieval method, which expresses text by concept units.
A new approach to keyfact-based text retrieval methods has been proposed in order to overcome the shortcomings of the keyword-based text retrieval method and generalize phrase-based text retrieval method. In the keyfact-base text retrieval method, a part of text that represent the same meaning is described as a keyfact. Since the keyfact-based retrieval method is a sort of concept-based retrieval method, and therefore indexing and retrieval of the keyfact-based retrieval method are performed with the unit of the keyfact, precision of the retrieval is greatly improved.
In the keyfact-based retrieval method, it is desirable that phrases or words having the same meaning are indexed as the same indexing terms. For example, noun phrases including “the retrieval of information” as a subset of “the efficient retrieval of information”, “the retrieval of the distributed information”, and “the fast retrieval of the distributed information” must have common indices which can be possibly generated from “the retrieval of information” as subsets and recognize also them as different meaning with subtle conceptual different indexes at the same time.
Since the keyword-based retrieval method doesn't recognize the conceptual difference between “the retrieval of the information” and “the efficient retrieval of the information”, users are not able to retrieve the exact document that is desired.
SUMMARY OF THE INVENTION
A keyfact-based retrieval method, which extracts the precise keyfact pattern using the natural language processing techniques and indexes documents with the unit of the keyfact, is provided.
In addition, a keyfact-based retrieval method, which extracts precise keyfact patterns included in a natural query of a user using the natural language processing techniques and retrieves documents similar to the query in the keyfact-based index file, is provided.
In addition, a keyfact-based retrieval method, which retrieves and indexes documents with the unit of keyfact, is provided.
A keyfact-based text retrieval system of the present invention includes keyfact extracting means, keyfact indexing means, and keyfact retrieving means. The keyfact extracting means analyze a document collection and a user query, and extracting keywords not having part-of-speech ambiguity from the document collection and the user query, and respectively extracting keyfacts of the document collection and the user query from the keywords. The keyfact indexing means for calculating the frequency of the keyfacts of the document collection and generating a keyfact list of the document collection for a keyfact index structure. The keyfact retrieving means for receiving the keyfact of the user query and the keyfacts of the document collection and defining a keyfact retrieval model in consideration of weight factors according to a keyfact pattern and generating a retrieval result.
The keyfact extracting means includes morphology analyzing means, part-of-speech tagging means, keyfact pattern extracting means, and keyfact generating means. The morphology analyzing means analyze morphology of an input sentence and obtaining tag sequences of part-of-speech by attaching part-of-speech tags. The part-of-speech tagging means selects a tag sequence of part-of-speech out of the tag sequences of part-of-speech. The tag sequence of part-of-speech is precise. The keyfact pattern extracting means extracts a keyfact pattern by applying the tag sequences of part-of-speech to a keyfact pattern rule. The keyfact generating means applies the keyfact pattern to a keyfact pattern generation rule and generating a keyfact list, which is a set of keyfact terms.
The keyfact indexing means includes frequency calculating means, table generating means, and keyfact indexing means. The frequency calculating means calculates a frequency of various keyfacts and a document frequency of the keyfacts. The various keyfacts are included in the document collection, and the document frequency is the number of documents contained the various keyfacts. The table generating means generates a document index table, a document table, and a keyfact index table of the document collection. The keyfact indexing means forms a keyfact index structure. The keyfact index structure has information regarding document frequency, document identifier, and keyfact frequency in each corresponded documents.
The keyfact retrieving means includes following means. A means forms a document and a user query vector with an index file and the keyfact of the user query. The index file generated by the keyfact indexing means. The keyfact of the user query generated by the keyfact extracting means. A means determines keyfact weight constants in accordance with the keyfact pattern. A means calculates keyfact weights for the document and the user query by applying the keyfact weight constants to the document and the user query vector. The retrieval results displaying means displays the retrieval result by applying the keyfact weights to keyfact retrieval model. The retrieval result indicates documents with a keyfact similar to the keyfact of the user query.
A keyfact-based text retrieving method of the present invention includes keyfact extracting step, keyfact indexing step, and keyfact retrieving step. The keyfact extracting step is to analyze a document collection and a user query, and extracts keywords without part-of-speech ambiguity from the document collection and the user query, and respectively extracts keyfacts of the document collection and the user query from the keywords. The keyfact indexing step is to calculates the frequency of the keyfacts of the document collection and generates a keyfact list of the document collection for a keyfact index structure. The keyfact retrieving step is to receives the keyfact of
Chong Kyung Taek
Jang Myung-Gil
Jun MiSeon
Park Se Young
Black Thomas
Electronics and Telecommunications Research Institute
SEED IP Law Group PLLC
Tarleton E. Russell
Veillard Jacques
LandOfFree
Keyfact-based text retrieval system, keyfact-based text... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Keyfact-based text retrieval system, keyfact-based text..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Keyfact-based text retrieval system, keyfact-based text... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2873052