Document retrieval-assisting method and system for the same and document retrieval service using the same with document frequency and term frequency

• US 5,987,460 A
• Filed: 07/03/1997
• Issued: 11/16/1999
• Est. Priority Date: 07/05/1996
• Status: Expired due to Term

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Niwa, Yoshiki, Sakurai, Hirofumi
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): RONES, CHARLES
• Application Number:
  US08/888,017
• Time in Patent Office: 866 Days
• Field of Search: 707/5, 707/2, 707/6, 707/3, 707/4, 707/10
• US Class Current: 1/1
• CPC Class Codes: G06F 16/338 Presentation of query results
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

Because the whole image of a retrieved document group cannot be viewed, next retrieval request after one retrieval attempt has been determined only perceptually. Therefore, displaying a means for displaying topic words on a display means, extracting a group of words characteristically appearing in a retrieved document group on request from a user, further examining the relation between topic words, and preparing a graph using the topic words as nodes, the whole image of the retrieval results is displayed on the means for displaying topic words. Additionally, by selecting an interesting word or a word with no interest on a displayed graph of topic words, the user can design a subsequent retrieval strategy effectively.

Claims

1. A method for assisting document retrieval, comprising,

detecting documents each including one or a plurality of keywords corresponding to a query as documents of retrieval results from a retrieval subject document group,

detecting the document frequency of a word, representing the number of documents where the word appears in the group of the documents of retrieval results,

detecting the total document frequency of a word representing the number of documents where the word appears in the whole retrieval subject document group,

introducing a frequency ratio representing the ratio of the document frequency of a word to the total document frequency of the word,

classifying the document frequency in a set of frequency classes as based on a given relation and assigning each word to a corresponding frequency class, depending on the document frequency of the word,

extracting an appropriate number of words from each of the frequency classes in the decreasing order of frequency ratio of word as topic words, and

displaying the extracted topic words in the form of a graph or a list.

2. A method for assisting document retrieval, according to claim 1, wherein the extracted topic words are displayed in the form of a list per frequency class or the form of a graph representing the relation between topic words.

3. A method for document retrieval service, comprising,

detecting documents each including a keyword transmitted from a retriever as documents of retrieval results from a retrieval subject document group,

detecting the document frequency of a word, representing the number of documents where the word appears in the group of documents as retrieval results,

detecting the total document frequency of a word representing the number of documents where the word appears in the whole retrieval subject document group,

introducing a frequency ratio representing the ratio of the document frequency of a word to the total document frequency of the word,

classifying the document frequency in a set of frequency classes based on a given relation and assigning each word to a corresponding frequency class, depending on the document frequency of the word,

extracting an appropriate number of words from each of the frequency classes in the decreasing order of frequency ratio of word as topic words,

composing the extracted topic words as a data displayable in the form of a list per frequency class or in the form of a graph representing the relation between topic words, and

transmitting the topic words and said composed data for displaying to the retriever.

4. A method for document retrieval service, according to claim 3, wherein the retriever is equipped with a means for transmitting keywords for retrieval to the retrieval system, and receiving the composed data concerning topic words transmitted from the retrieval system, and displaying the topic words according to the transmitted composed data, so as to receive retrieval service.

5. A method for document retrieval service, according to claim 4, wherein the retriever receives, from a document retrieval service system before acting document retrieval for transmitting keywords for retrieval to the retrieval service system, and receiving the composed data concerning topic words transmitted from the retrieval service system, and displaying the topic words according to the transmitted composed data.

19. A document retrieval system, according to claim 3, wherein the relation between topic words is determined on the basis of the co-occurrence relation between the topic words, and wherein the form of a graph is a graph composed by linking together a word pair of topic words with higher relation, using topic words as nodes.

20. A document retrieval system, according to claim 19, wherein the direction of the longitudinal axis in the graphic representation of topic words represents the document frequency of a topic word in the retrieved document group.

6. A machine readable data storing media on which the word frequency data for selecting topic words are recorded, wherein the frequency data of each word comprises (a) character sequence, (b) the document frequency of the word, representing the number of documents where the word appears in the group of documents of retrieved results, (c) the total document frequency of the word representing the number of documents where the word appears in the whole retrieval subject document group, (d) the frequency ratio representing the ratio of the document frequency of the word to the total document frequency of the word, and (e) the frequency class of the word assigned to the word depending on its document frequency, and wherein topic words are extracted from each of the frequency classes in the decreasing order of frequency ratio of word.

7. A machine readable data storing media on which the co-occurrence data for calculating the relatedness among topic words are recorded, wherein the co-occurrence data of each pair of topic words comprises (a) the co-occurrence frequency of the word pair, that is, the number of documents in the retrieved document set where both words of the word pair appear, and (b) the co-occurrence intensity of the word pair, such as their co-occurrence frequency divided by document frequency of the second word of the word pair, and wherein the links of the graphical display of topic words are generated for word pairs with strong relation.

8. A machine readable data storing media on which the coordinate data for displaying a graphical display of topic words are recorded, wherein the data comprises data for displaying nodes of the topic word graph and data for displaying links of the graph representing strong relation between topic words, and the data for displaying nodes comprises the coordinate center, character sequence, and the character number in the crosswise and lengthwise directions of a region displaying the characters, and the size of the displaying region, and the data for displaying links comprises the initiation coordinate and the termination coordinate of each link, and wherein the graphic display of a word graph are displayed following the data.

9. A machine readable data storing media on which are recorded the data of each word for selecting topic words, data of each pair of topic words for calculating the relatedness among topic words, and data of each topic word for displaying a graphical display of topic words, wherein the data of each word comprises (a) character sequence, (b) the document frequency of the word, representing the number of documents where the word appears in the group of documents of retrieved results, (c) the total document frequency of the word representing the number of documents where the word appears in the whole retrieval subject document group, (d) the frequency ratio representing the ratio of the document frequency of the word to the total document frequency of the word, and (e) the frequency class of the word assigned to the word depending on its document frequency, and topic words are extracted from each of the frequency classes in the decreasing order of frequency ratio of word, wherein the co-occurrence data of each pair of topic words comprises(a) the co-occurrence frequency of the word pair, that is, the number of documents in the retrieved document set where both words of the word pair appear, and (b) the co-occurrence intensity of the word pair, such as their co-occurrence frequency divided by document frequency of the second word of the word pair, and the links of the graphical display of topic words are generated for word pairs with strong relation, wherein the data comprises data for displaying nodes of the topic word graph and data for displaying links of the graph representing strong relation between topic words, and the data for displaying nodes comprises the coordinate center, character sequence, and the character number in the crosswise and lengthwise directions of a region displaying the characters, and the size of the displaying region, and the data for displaying links comprises the initiation coordinate and the termination coordinate of each link, and the graphic display of a word graph can be displayed following the data.

10. A document retrieval system comprising

a means for detecting documents each including one or a plurality of keywords corresponding to a query as documents of retrieval results from a retrieval subject document group,

a means for detecting the document frequency of a word, representing the number of documents where the word appears in the group of the documents of retrieval results,

a means for detecting the total document frequency of a word representing the number of documents where the word appears in the whole retrieval subject document group,

a means for introducing a frequency ratio representing the ratio of the document frequency of a word to the total document frequency of the word,

a means for classifying the document frequency in a set of frequency classes as based on a given relation and assigning each word to a corresponding frequency class, depending on the document frequency of the word,

a means for extracting an appropriate number of words from each of the frequency classes in the decreasing order of frequency ratio of word as topic words, and

a means for displaying the extracted topic words in the form of a graph or a list.

11. A document retrieval system, according to claim 10, wherein said display means displays the extracted topic words in the form of a list per frequency class or in the form of a graph representing the relation between topic words and said retrieval system includes means for selecting the display form.

12. A document retrieval system, according to claim 11, wherein the query comprises three types of keywords of requisite keyword, positive keyword and negative keyword, and the retrieval by requisite keyword is carried out under AND conditions with each requisite keyword, and wherein individual documents in the group of documents as retrieval results can be given a higher score depending on the number of positive keywords if the documents include the positive keywords or are reduced of their scores depending on the number of negative keywords if the documents include the negative keywords, to extract topic words from a document group with a higher score given.

13. A document retrieval system, according to claim 12, wherein the type of a keyword can be changed among three types of keywords, namely the one or a plurality of keywords of the query, positive keywords and negative keywords, and the topic words displayed can be copied as any of these three types of keywords.

14. A document retrieval system, according to claim 12, wherein retrieval with positive keywords is run if no requisite keyword is set and retrieval with each positive keyword is run under OR conditions.

15. A document retrieval system, according to claim 14, wherein the type of a keyword can be changed among three types of keywords, namely the one or a plurality of keywords of the query, positive keywords and negative keywords, and the topic words displayed can be copied as any of these three types of keywords.

16. A document retrieval system, according to claim 11, wherein the relation between topic words is determined on the basis of the co-occurrence relation between the topic words, and wherein the form of a graph is a graph composed by linking together a word pair of topic words with higher relation, using topic words as nodes.

17. A document retrieval system, according to claim 16, wherein the direction of the longitudinal axis in the graphic representation of topic words represents the document frequency of a topic word in the retrieved document group.

18. A document retrieval system, according to claim 11, wherein the direction of the longitudinal axis in the graphic representation of topic words represents the document frequency of a topic word in the retrieved document group.

21. A document retrieval system, according to claim 10, wherein the query comprises three types of keywords of requisite keyword, positive keyword and negative keyword, and the retrieval by requisite keyword is carried out under AND conditions with each requisite keyword, and wherein individual documents in the group of documents as retrieval results can be given a higher score depending on the number of positive keywords if the documents include the positive keywords or are reduced of their scores depending on the number of negative keywords if the documents include the negative keywords, to extract topic words from a document group with a higher score given.

22. A document retrieval system, according to claim 21, wherein the type of a keyword can be changed among three types of keywords, namely the one or a plurality of keywords of the query, positive keywords and negative keywords, and the topic words displayed can be copied as any of these three types of keywords.

23. A document retrieval system, according to claim 21, wherein retrieval with positive keywords is run if no requisite keyword is set and retrieval with each positive keyword is run under OR conditions.

24. A document retrieval system, according to claim 23, wherein the type of a keyword can be changed among three types of keywords, namely the one or a plurality of keywords of the query, positive keywords and negative keywords, and the topic words displayed can be copied as any of these three types of keywords.