Information retrieval system with a neuro-fuzzy structure

• US 6,845,354 B1
• Filed: 09/09/1999
• Issued: 01/18/2005
• Est. Priority Date: 09/09/1999
• Status: Expired due to Term

• Current Assignee: Institute For Information Industry
• Original Assignee: Institute For Information Industry
• Inventors: Kuo, Yau-Hwang, Hsu, Jang-Pong
• Primary Examiner(s): Knepper, David D.
• Application Number:
  US09/392,572
• Time in Patent Office: 1,958 Days
• Field of Search: 704/1, 704/9, 704/10, 707 2- 3, 707/4, 707 5- 6, 707/104, 707/102, 715/531, 715532-534
• US Class Current: 704/9
• CPC Class Codes: G06F 40/284 Lexical analysis, e.g. tokenisation or collocates
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

An intelligent information retrieval system for finding information components corresponding to an input query word. The system includes a synonym block for finding synonymous indexed keywords of the query word; an information indexing block for finding corresponding information components based on the indexed keyword; a component ranking-filtering block for ranking and filtering the found information components and outputting the desired information components being selected; a synonym adjusting block for adjusting the fuzzy mechanism of the synonym block based on the found information components; and a filtering adjusting block for adjusting the fuzzy mechanism of the ranking-filtering block based on the found information components. The aforementioned synonym block and information-indexing block are implemented by neuro-fuzzy networks for accelerating parallel processing and automatic learning. Further, the synonym block can tolerate input errors by way of query word encoding and position shift compensation.

Claims

1. An information retrieval system for finding corresponding information components according to an entered query word, comprising:

a synonym block for finding indexed keywords according to the query word;

an information indexing block for finding a corresponding information component based on the indexed keywords; and

a component ranking-filtering block for ranking and filtering the found information components and for outputting the information components for selection;

wherein the synonym block comprises;

encoding means for encoding the query word and generating an encoded pattern of the query word, the encoded pattern of the query word corresponding to a feature of the query word;

similarity means for comparing the encoded pattern of the query word with the encoded patterns of system synonyms, the latter generated by encoding a plurality of system synonyms, and generating corresponding similarity values, the similarity values representing the similarity between the query word and the system synonyms, each of the system synonyms corresponding to at least one of the system keywords, respectively; and

correlation means for selecting the indexed keywords from the system keywords according to correlation values between the system synonyms and the system keywords, the similarity values generated by the similarity means and a predetermined threshold value for comparison.

2. The information retrieval system as claimed in claim 1, wherein the information components are software components classified by types, functions, associated storage media and operating platforms.

3. The information retrieval system as claimed in claim 1, wherein the similarity means comprises:

a plurality of similarity-calculating units for respectively storing the encoded patterns of the system synonyms corresponding to the system synonyms, receiving the encoded pattern of the query word, and calculating the corresponding similarity values according to a similarity-calculating function.

4. The information retrieval system as claimed in claim 1, wherein the correlation means comprises:

a plurality of correlation units, corresponding to the system keywords, respectively, for receiving the similarity values generated by the similarity means and comparing the predetermined threshold value with the combination of the correlation values and the similarity values, and determining whether the corresponding system keywords serve as the indexed keywords.

5. The information retrieval system as claimed in claim 1, wherein the encoding means encodes the query word into the corresponding encoded pattern of the query word by recording the occurrences of the alphabetical letters in the query word , the location of the first occurrence for each of the alphabetical letters and a portion of the query word.

6. The information retrieval system as claimed in claim 5, wherein the synonym block further comprises a shifting means for calculating the position shift amount of each alphabetical letter by the location of the first occurrence of each alphabetical letter to determine whether there is any input error, and compensating the query word when there is an input error.

7. An information retrieval system for finding corresponding information components according to an entered query word, comprising:

a synonym block for finding indexed keywords corresponding to the query word;

an information indexing block for finding a corresponding information component based on the indexed keywords; and

a component ranking-filtering block for ranking and filtering the found information components and for outputting the information components for selection;

wherein the synonym block comprises;

encoding means for encoding the query word and a keyword for the found information components and generating an encoded pattern of the query word and an encoded pattern of the keyword for the found information components, the encoded pattern of the query word corresponding to a feature of the query word, the encoded pattern of the keyword for the found information components corresponding to a feature of the keyword for the found information components;

similarity means for respectively comparing the encoded pattern of the query word and the encoded pattern of the keyword for the found information components with encoded patterns of system synonyms, the latter generated by encoding a plurality of system synonyms, and generating a first set of similarity values and a second set of similarity values, the first set of the similarity values representing the similarity between the query word and the system synonyms, the second set of the similarity values representing the similarity between the keyword for the found information components and the system synonyms, each of the system synonyms corresponding to at least one of the system keywords, respectively;

correlation means for calculating an output value according to the first set of the similarity values, the second set of the similarity values and original correlation values between the system synonyms and the system keywords; and

determining means for determining whether the similarity means and the correlation means need adjustment according to the first set of the similarity values, the second set of the similarity values and the output value.

8. The information retrieval system as claimed in claim 7, further comprising an adjusting means for adjusting the similarity means and the correlation means according to the output value generated by the determining means.

9. The information retrieval system as claimed in claim 7, wherein the information components are software components classified by types, functions, associated storage media and operating platforms.

10. The information retrieval system as claimed in claim 7, wherein the similarity means comprises:

a plurality of similarity-calculating units for respectively storing the encoded patterns of the system synonyms corresponding to the system synonyms and receiving the encoded pattern of the query word and the encoded pattern of the keyword for the found information components, to calculate the first set of the similarity values and the second set of the similarity values.

11. The information retrieval system as claimed in claim 8, wherein the determining means comprises a first selection means for comparing a first threshold value with the maximal of the first set of the similarity values and the maximal of the second set to decide an adjustment procedure pertaining to the similarity means and the correlation means.

12. The information retrieval system as claimed in claim 8, wherein the determining means comprises a second selection means for comparing a second threshold value with the maximal of the output values to decide an adjustment procedure pertaining to the similarity means and the correlation means.

13. The information retrieval system as claimed in claim 7, wherein the encoding means encodes the query word and the keyword for the found information components into the corresponding encoded patterns by recording the occurrences of the alphabetical letters in the query word and the keyword for the found information components, the location of the first occurrence for each of the alphabetical letters and a portion of the query word and the keyword for the found information components.

14. The information retrieval system as claimed in claim 13, wherein the synonym block further comprises a shifting means for calculating the position shift amount of each alphabetical letter by the location of the first occurrence of each alphabetical letter to determine whether there is any input error, and compensating the query word when there is an input error.

15. An information retrieval system for finding corresponding information components according to an entered query word, comprising:

a synonym block for finding indexed keywords according to the query word;

an information indexing block for finding a corresponding information component based on the indexed keywords; and

a component ranking-filtering block for ranking and filtering the found information components and for outputting the information components for selection;

wherein the information indexing block comprises;

storing means for sequentially storing the indexed keywords; and

a plurality of indexing units, respectively corresponding to the information components and storing a component keyword set and identification data pertaining to the corresponding information components, for calculating a similarity value between the component keyword set and the received indexed keywords and outputting the identification data pertaining to the corresponding information component when the similarity value exceeds a predetermined threshold value.

16. The information retrieval system as claimed in claim 15, wherein the information components are software components classified by types, functions, associated storage media and operating platforms.

17. The information retrieval system as claimed in claim 15, wherein the predetermined threshold value is adjustable and the adjusted threshold value is used to decide the number of the outputted identification data pertaining to the corresponding information components.

18. An information retrieval system for finding corresponding information components according to an entered query word, comprising:

a synonym block for finding indexed keywords responsive to the query word according to a fuzzy synonym procedure;

an information-indexing block for finding a corresponding information component based on the indexed keywords;

a component ranking-filtering block for ranking and filtering the found information components and for outputting the information components for selection according to a fuzzy ranking-filtering procedure;

a synonym adjusting block for adjusting the fuzzy synonym procedure of the synonym block according to the found information components; and

a ranking-filtering adjusting block for adjusting the fuzzy ranking-filtering procedure of the component ranking-filtering block according to the found information components.

19. The information retrieval system as claimed in claim 18, wherein the information components are software components classified by types, functions, associated storage media and operating platforms.

20. The information retrieval system as claimed in claim 18, wherein the synonym block is implemented by a fuzzy neural network for accelerating parallel processing.

21. The information retrieval system as claimed in claim 18, wherein the information-indexing block is implemented by a fuzzy neural network for accelerating parallel processing.

22. The information retrieval system as claimed in claim 18, wherein the synonym block comprises encoding means for encoding the query word into an encoded pattern by recording the occurrences of the alphabetical letters in the query word, the location of the first occurrence for each of the alphabetical letters and a portion of the query word;

and shifting means for calculating the position shift amount of each alphabetical letter by the location of the first occurrence of each alphabetical letter to determine whether there is any input error and compensating the query word when there is an input error.