System for generating a search formula by accessing search terms on the basis of a training set of pertinent and non-pertinent objects

• US 5,442,781 A
• Filed: 11/24/1992
• Issued: 08/15/1995
• Est. Priority Date: 11/29/1991
• Status: Expired due to Term

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Yamagata, Kanako
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): AMSBURY, WAYNE P
• Application Number:
  US07/980,702
• Time in Patent Office: 994 Days
• Field of Search: 395/600, 364/419.19
• US Class Current: 707/779
• CPC Class Codes: G06F 16/3322 using system suggestions G06F16/3325 takes precedence
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

An apparatus for automatically generating a search formula uses pertinent data and non-pertinent data given by the user. Suitable search terms are selected from the pertinent data on the basis of the term appearance ratios only in the pertinent data and in the given data. The respective effectiveness values of the search terms are calculated by using the number of pieces of pertinent data containing the term and the number of pieces of given data containing the term. Among the sums of the effectiveness values of the search terms included in respective pieces of pertinent data, the smallest sum is decided to be the threshold of effectiveness. Finally, the search formula consisting of search terms combined is generated on the basis of the threshold of effectiveness and the respective effectiveness values of the search terms.

Claims

1. Apparatus for generating a search formula in an information retrieval system comprising:

input means for inputting given data consisting of pertinent data and non-pertinent data, said pertinent data and said non-pertinent data being designated by a user, said pertinent data and said non-pertinent data comprising pieces of data, each piece of said pertinent data satisfying a need of said user, each piece of said non-pertinent data not satisfying the need of said user;

search term selection means for selecting search terms among the terms included in the pertinent data on the basis of first and second term appearance ratios, said first term appearance ratio being a ratio of a number of pieces of said given data containing said particular term to a total number of pieces of said given data, said second term appearance ratio being a ratio of a number of pieces of pertinent data containing said particular term to a total number of pieces of said pertinent data;

effectiveness calculation means for calculating respective effectiveness values of the search terms selected by the search term selection means, said effectiveness calculation means calculating said effectiveness of each selected search term based on a number of pieces of said pertinent data containing the search term and the total number of pieces of said given data;

threshold determination means for determining a threshold of effectiveness by using the effectiveness values of the search terms included in respective pieces of pertinent data; and

search formula generation means for generating a search formula on the basis of the respective effectiveness values of the search terms and the threshold determined by the threshold determination means, said search formula comprising the selected search terms combined by boolean operators.

2. Apparatus according to claim 1, wherein the search term selection means selects as a search term the term whose term appearance ratio only in the pertinent data is higher than the term appearance ratio in the given data.

3. Apparatus according to claim 1, wherein the effectiveness calculation means calculates the respective effectiveness values of the search terms using the following formula:

(effectiveness value)=(1+the number of pieces of pertinent data containing the term)/(2+the number of pieces of given data containing the term).

4. Apparatus according to claim 1, wherein the threshold determination means determines the threshold of effectiveness by using the sum of the effectiveness values of the search terms included in respective pieces of pertinent data.

5. Apparatus according to claim 4, wherein the threshold determination means determines the threshold of effectiveness such that among the sums of the effectiveness values of the search terms included in respective pieces of pertinent data, the smallest sum is set as the threshold of effectiveness.

6. Apparatus according to claim 1, wherein the search formula generation means generates a search formula by using the search term having the effectiveness values equal to or greater than the threshold of effectiveness determined by the threshold determination means.

7. Apparatus according to claim 1, wherein the search formula generation means generates a search formula by using the combination of search terms having the sum of the effectiveness values of the search terms equal to or greater than the threshold of effectiveness when there is no search term having the effectiveness values equal to or greater than the threshold of effectiveness.

8. Apparatus according to claim 1, wherein the search formula generation means generates a search formula comprising a plurality of partial search formulae combined by the OR operator.

9. Apparatus according to claim 8, wherein the respective partial search formulae are comprised of the search term that has the effectiveness value equal to or greater than the threshold of effectiveness.

10. Apparatus according to claim 8, wherein the respective partial search formulae are comprised of the combination of search terms that has the sum of the effectiveness values of the search terms equal to or greater than the threshold of effectiveness, said search terms being combined by the AND operator.

11. A method of generating a search formula in an information retrieval system comprising the steps of:

receiving from a user given data containing pieces of pertinent data and pieces of non-pertinent data, each piece of pertinent data satisfying a user need, each piece of non-pertinent data not satisfying the user need, each piece of given data containing at least one term;

selecting search terms from the terms in the pertinent data on the basis of first and second term appearance ratios for each term, the first term appearance ratio being a ratio of a number of pieces of given data containing the term to a total number of pieces of given data, the second term appearance ratio being a ratio of a number of pieces of pertinent data containing the term to a total number of pieces of pertinent data;

calculating effectiveness values for each of the selected search terms based on a number of pieces of pertinent data containing the selected search term and the total number of pieces of given data;

determining an effectiveness threshold using the calculated effectiveness values; and

generating the search formula based on the effectiveness values and the threshold.

12. A method according to claim 11, wherein the selecting step selects terms for which the second term appearance ratio is higher than the first term appearance ratio as selected search terms.

13. A method according to claim 11, wherein the step of calculating effectiveness values for each of the selected search terms uses the following formula:

(effectiveness value)=(1+the number of pieces of pertinent data containing the term)/(2+the number of pieces of given data containing the term).

14. A method according to claim 11, wherein the step of determining the effectiveness threshold uses the sum of the effectiveness values of the search terms in each respective piece of pertinent data.

15. A method according to claim 14, wherein the effectiveness threshold is set to the smallest sum of effectiveness values.

16. A method according to claim 11, wherein the step of generating the search formula uses search terms having effectiveness values equal to or greater than the effectiveness threshold.

17. A method according to claim 16, wherein the step of generating the search formula uses combinations of search terms having a sum of effectiveness values equal to or greater than the effectiveness threshold when no search term has an effectiveness value equal to or greater than the effectiveness threshold.

18. A method according to claim 11, wherein the step of generating the search formula uses a plurality of partial search formulae combined by an OR operator.

19. A method according to claim 18, wherein each search term having an effectiveness value equal to or greater than the effectiveness threshold is a partial search formula.

20. A method according to claim 18, wherein combinations of search terms being combined by an AND operator and having a sum of effectiveness values equal to or greater than the effectiveness threshold are partial search formulae.