JPH11328194A - Keyword retrieval method and device and storage medium storing keyword retrieval program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform appropriate retrieval even in the case that a keyword is inappropriate by adding the keyword selected by a user to retrieval conditions inputted by the user, executing retrieval, correcting a rule base based on the keyword selected by the user and displaying the result of the retrieval to the user. SOLUTION: As a keyword derivation processing, the keyword included in the retrieval conditions inputted by the user is converted to the keyword more suited to the retrieval (S110). As a derivative keyword selection processing, the keyword connected in the keyword derivation processing is presented to the user and the user himself/herself selects the keyword considered as more appropriate (S120). As a retrieval condition generation processing, new retrieval conditions are generated by using the keyword selected by the user in the derivative keyword selection processing (S130). As a retrieval execution processing, the retrieval is executed by using the retrieval conditions generated in the retrieval condition generation processing (S140). The result of the retrieval is displayed by a retrieved result display part.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a keyword search method and apparatus and a storage medium storing a keyword search program, and in particular, in a keyword search operation, converts a search condition input by a user into a more appropriate search condition. The present invention relates to a keyword search method and apparatus for improving the efficiency of an information search operation, and a storage medium storing a keyword search program.

[0002]

2. Description of the Related Art As a conventional keyword search method, there is the following method. 1. A keyword search is performed using the search conditions input by the user as they are. 2. A dictionary called a thesaurus that describes the relationship between words is created manually or automatically from a document to be searched in advance using a co-occurrence relationship in which words appearing in the same document set are related. Thereafter, when a certain keyword is input as a search condition, a related word in the thesaurus is statically added, thereby converting the first search condition into a more appropriate one and performing a keyword search.

[0003] 3. When the user's search operation is monitored, a plurality of users search using different keywords, and when one piece of information is selected as a correct search result, the keywords used in each search are related to each other. And save them as related words. Thereafter, when any one of these keywords is input as a search condition, a related word is added to convert the first search condition into a more appropriate one and perform a keyword search.

[0004]

However, the above-described conventional keyword search method has the following problems. 1. In the method of performing a keyword search using the search condition entered by the user as it is, when there is a difference between the notation of the keyword entered by the user and the keyword of the search target (eg, a computer and a computer), or when the keyword is a synonym (E.g., computers and computers) has the problem of not being able to perform a proper search.

[0005] 2. Using a manually created thesaurus,
Perform a keyword search by statically deriving a word closely related to the keyword in the input search condition and using this to convert the search condition (eg, convert a computer to "computer @ computer"). The method can solve the above-mentioned fluctuation of the notation and the problem of synonyms to some extent. In addition, a thesaurus is automatically generated using the co-occurrence relation,
In the method of converting search conditions using this, conversion of search conditions useful for the user can be performed to some extent. For example, "computer" and "graphics"
If there are many documents that include both in the search target document, a co-occurrence relationship is established between “computer” and “graphics”. By registering this relationship in the thesaurus, It can be converted to "Computer @ Graphics" to perform keyword search. However, it is almost impossible to manually describe the relationships between all words in a thesaurus. Also, for documents for which a co-occurrence relationship does not hold in a search target document, a rule for deriving a related word cannot be automatically generated. For example, even if a user wants to search for documents related to CPU exchange, if there are not many documents that include both “CPU” and “exchange” among documents to be searched,
Since a rule for deriving “Exchange” from “CPU” is not generated, the search condition “CPU” is changed to “CPU @ Exchange”
Cannot be converted to The use of a static thesaurus created in advance in this way causes a problem that appropriate related words cannot be derived.

[0006] 3. In a method in which different users input different keywords entered to search for the same information and determine that they are related to each other, and use this to convert search conditions, rules for conversion are dynamically generated, The above problems can be solved to some extent. However, if the selected information includes a plurality of contents as a result of the search, and the search conditions input by a plurality of users are respectively different, the used search conditions may not be related to each other. Despite their high propensity, they are determined to be uniformly related and stored. Therefore, when the search condition is converted by this method, there is a risk that the search condition may be different from the user's intention.

[0007] For example, the selected information is an instruction manual of a program for creating computer graphics, and a description of an operation of an operation system for starting the program and a description of a design technique using computer graphics. Suppose it was included. In this case, this method may determine that the keywords related to the operation system and the keywords related to the design are related to each other,
Thereafter, when searching for information relating to the design, there is a disadvantage that information relating to the operation system is hit. As described above, there is a problem that a conversion rule with an incorrect search condition is stored.

As described above, in the conventional keyword search, if the keyword input by the user is not appropriate, an appropriate search cannot be executed, and the function for supporting the user search is not necessarily appropriate. There is a problem that does not work. The present invention has been made in view of the above points, and is interposed between a user and a conventional keyword search apparatus that performs a keyword search using a search condition input by a user as it is, and a search condition input by the user. To provide a keyword search method and apparatus capable of supporting a keyword search by converting a keyword into a search condition including a keyword considered to be more suitable for the user's intention, and a storage medium storing the keyword search program. Aim.

[0009]

FIG. 1 is a diagram for explaining the principle of the present invention. The present invention (Claim 1) provides a keyword search method for executing a search related to a keyword input by a user or a search condition composed of an expression obtained by combining keywords with a Boolean operator. Search for information indicating the degree of relevance to any keyword including the keyword,
A search is made for a rule base that can be saved and modified (step 1), and the degree of relevance of the keyword included in the search condition input by the user is equal to or greater than a predetermined value, and (Step 2), display the keyword obtained from the rule base to the user (Step 3), and select a keyword to be used for an actual search from the keyword (Step 4). The keyword selected by the user is added to the search condition input by the user (step 5), a search is executed (step 6), and the rule base is corrected based on the keyword selected by the user (step 7). Then, the result of the search is displayed to the user (step 8).

According to the present invention (claim 2), when a plurality of keywords are input by a user, a predetermined value is set as a degree of relevance for a combination in which information on the degree of relevancy between keywords does not exist in the rule base. , Stored in the rule base, and for an existing combination, a correction process is performed on the rule base to increase the relevance between these keywords.

According to the present invention (claim 3), for all combinations of a keyword input by a user and a keyword selected by the user, a combination in which information on the degree of relevance between the keywords exists in the rule base: A correction process is performed on the rule base to increase the degree of association between the keywords. According to the present invention (claim 4), for all combinations of a keyword input by a user and a keyword not selected by the user, information on the degree of relevance between the keywords is defined as: A correction process is performed on the rule base to reduce the degree of association between the keywords.

According to the present invention (claim 5), a related keyword is acquired and searched at the next and subsequent searches using the rule base on which the correction processing has been performed. FIG. 2 is a diagram illustrating the principle of the present invention. The present invention (claim 6) is a keyword search device for executing a search related to a keyword input by a user or a search condition composed of an expression in which keywords are combined by a Boolean operator, wherein the keyword search device is an optional device used in the search. And a rule base 290 capable of searching, storing, and modifying information indicating the degree of association between the keyword and any keyword including the keyword, and a user search condition input for receiving a search condition input from a user Means 200, a search object database 280 to be searched, a search means 250 for searching the search object database 280 based on the search condition input by the user search condition input means 200, and a rule base 290 for searching. The degree of relevance for keywords included in the search conditions entered by the user is predetermined. And when the value or more, and other keyword acquiring unit 270 that acquires a keyword that may be registered in the rule base 290
And displaying the keyword obtained from the rule base 290 to the user, and allowing the user to select a keyword to be used for an actual search from the keyword. Re-search means 255 for executing a search in addition to the search conditions input by the user, rule-base correcting means 260 for correcting the rule base 290 based on the selected keyword, and search result display for displaying the search results to the user Means 225.

According to the present invention (claim 7), the rule base 29
0 is composed of an input keyword as an input, a derived keyword derived from the input keyword, and a degree of association between the input keyword and the derived keyword. According to the present invention (claim 8), in the rule base correction unit 260, when there are a plurality of keywords input by the user, a combination in which information on the degree of association between keywords does not exist in the rule base 290 is determined in advance. There is further provided a first rule base correction unit that stores the value as the degree of relevance in the rule base 290 and performs a correction process on the existing combination to increase the degree of relevance between these keywords. .

According to the present invention (claim 9), in the rule base correcting means 260, for all combinations of a keyword input by a user and a keyword selected by the user, information on the degree of relevance between the keywords is applied to the rule base. Regarding the combination existing in the 290, there is further provided a second rule-based correction unit for performing a correction process on the rule base 290 to increase the relevance between the keywords.

According to the present invention (claim 10), in the rule base correcting means 260, for all combinations of a keyword input by a user and a keyword not selected by the user, information on the degree of association between the keywords is provided. ,
For the combinations that exist in the rule base 290,
There is further provided a third rule-base correction unit for performing a correction process on the rule base 290 to reduce the degree of association between the keywords.

According to the present invention (claim 11), the search means 250
Includes means for performing a search by using the rule base 290 that has been subjected to the correction processing to acquire a related keyword at the next and subsequent searches. The present invention (Claim 12)
Is a storage medium storing a keyword search program for executing a search related to a keyword input by a user or a search condition formed by combining keywords by a Boolean operator, and receiving a search condition from the user. A search condition input process, a search process for searching the search target database based on the search condition input by the user search condition input process,
A rule base search process for searching a rule base that can search, save, and modify information indicating a degree of association between any keyword used in the search process and any keyword including the arbitrary keyword; A keyword whose degree of relevance is greater than or equal to a predetermined value and is likely to be registered in the rule base with respect to a keyword included in a search condition input by a user, by performing a search using a rule-based search process. , A keyword selection process for displaying the keyword obtained from the rule base to the user, and selecting a keyword to be used for an actual search from the keyword, and a user selection command process for selecting the keyword. The keyword is added to the search condition entered by the user to execute the search. It has a search process for a rule-based modification process to modify the rule base on the basis of the selected keyword, and a search result display process for displaying the results of the search to the user.

According to the present invention (claim 13), in the rule-based correction process, when there are a plurality of keywords input by the user, a combination in which information on the degree of association between keywords does not exist in the rule base is predetermined. The value is stored in the rule base as the relevance,
For existing combinations, the method further includes a first rule-based modification process for performing a modification process on the rule base to increase the relevance between these keywords.

According to the present invention (claim 14), in the rule-based correction process, for all combinations of a keyword input by a user and a keyword selected by the user, information on the relevance between the keywords is stored in a rule base. For the existing combinations, the method further includes a second rule-based modification process for performing a modification process on the rule base to increase the degree of association between the keywords.

According to the present invention (claim 15), in the rule-based correction process, for all combinations of a keyword input by a user and a keyword not selected by the user, information on the degree of association between the keywords is For a combination existing in the rule base, the method further includes a third rule base modification process for performing a modification process on the rule base to reduce the degree of association between the keywords.

[0020] The present invention (claim 16) includes a process of performing a search by performing a related keyword acquisition at the next and subsequent searches by using the rule base having been subjected to the correction processing in the search process. As described above, the present invention differs from the conventional art in that keywords are converted manually or as a conversion rule, instead of using a static thesaurus created in advance by using co-occurrence relations. The rule base has a function to search, save, and modify information indicating the degree of association between any keyword used in the search and any keyword. A keyword whose degree is equal to or more than a predetermined value and which is not included in the search condition input by the user can be acquired and presented to the user.

Further, when there are a plurality of input keywords, for a combination in which the information on the keyword relevance does not exist in the rule base, a predetermined value is stored in the rule base as the relevance and exists. For the combination, a correction process is performed on the rule base to increase the relevance between these keywords. Also, for all combinations of the keyword input by the user and the keyword selected by the user, if the information on the relevance between these keywords is a combination existing in the above rule base, Then, a correction process for increasing the degree of association between these keywords is performed.

In addition, for all combinations of keywords entered by the user and keywords not selected by the user, combinations for which information on the degree of relevance between these keywords exists in the database are defined in the rule base. Is performed to lower the degree of association between keywords. This makes it possible to perform dynamic conversion of search conditions in subsequent searches.

[0023]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS In the following description, an expression in which keywords are combined by a Boolean operator is referred to as a "search condition", and a process for converting a keyword into a keyword more suitable for a search is referred to as a "keyword derivation process". The conversion rule for this is “keyword derivation rule”, the process of dynamically modifying the keyword derivation rule is “derivation rule modification process”, and the process of generating search conditions using derived keywords is “search condition generation process”. Write.

First, an outline of the retrieval according to the present invention will be described.
FIG. 3 is a flowchart for explaining the outline of the search according to the present invention. Step 110) As a keyword deriving process, a keyword included in the search condition input by the user is converted into a keyword more suitable for the search. Step 120) As a derived keyword selection process, the keyword converted in the keyword derivation process (step 110) is presented to the user, and the user himself / herself is made to select a keyword that he / she thinks appropriate.

Step 130) As a search condition generation process, a new search condition is generated using a keyword selected by the user in a derived keyword selection process (step 120). Step 140) As a search execution process, a search is performed using the search condition generated in the search condition generation process (step 130).

The above processes will be described in detail using the following symbols. U _t : a set composed of keywords included in the search conditions input by the user: R _t : a set composed of keywords included in the search conditions input to the keyword search and retrieval device: A _t : a set composed of derived keywords displayed on the screen: S _t: in a _t, consisting keyword selected by the user set: D _t: in a _t, the set consisting of the user has not selected keyword: in all _t, a t = S t ∨
D _t, it is a S _t ∧D _t = φ (empty set). Moreover, the search for the first time _{(t = 1), A t} , S t, D t is phi,
R _t = U _t .

The format of the keyword derivation rule will be described as a premise of each process. The keyword derivation rule has the following format. (Key1, key2, value) Here, key1 is a keyword as an input, key
2 is a keyword derived from key1 and va
lue indicates the degree of association between key1 and key2. One
For one key1, there can be a plurality of derived rules with different key2.

First, keyword derivation in step 110
The processing will be described. The keyword derivation process is t-1
All keywords R included in the search condition for the second time _t-1Or
A_tThis is the process of generating. R_t-1In the individual elements of
A certain keyword key₁, Key_Two, ..., key_mNoso
For each, describe the corresponding keyword derivation rule
And a list of pairs of keywords to be derived and their relevance (a₁₁, V₁₁) (A₁₂, V₁₂)… (A_{1n '}, V_{1n '}) (A_{twenty one}, V_{twenty one}) (A_{twenty two}, V_{twenty two})… (A_{2n "}, V_{2n "}). . (A_m1, V_m1) (A_m2, V_m2)… (A_mn ^('m), V_mn ^('m)) Is generated. Here, a_pqIs the key_pDerived from
The q-th keyword, v_pqIs the key_pAnd a_pqNoseki
It is continuous. Also, (n^('r)) Is the key_pDerived by
Number of rules.

Next, this list is scanned, and a_ij= A_pqWhen
If an element exists, (a_ij, V _ij) To (a_ij, V_ij
+ V_pq) As (a)_pq, V_pq) From the list
You. From the list generated in this way v_ijDescending order of the value of
Then, a predetermined number of k elements are extracted, and
-By extracting words, A_tGenerate Next
In the derived keyword selection process of step 120 described above,
explain about.

[0030] This process is a process which selectively presents keywords derived through the list A _t of the generated keywords (step 110) to the user, the keyword that matches the purpose of searching for user to user. Thereby, _St is generated. Also produce D _t from A _t and S _t. Next, the search condition generation processing in step 130 described above will be described.

In the t-th search, since the number of hits in the (t-1) -th search is too large, the search condition is added to the (t-1) -th search condition so that the search condition becomes more severe. Search ”and“ re-search ”in which a search condition is added to the t−1 search condition so that the number of hits in the t−1 search is too small, so that the search condition becomes gentler. Search exists.

According to the present invention, the two are distinguished, and a search generation process is prepared for each of them.・ Refinement search: (t = 1) R ' _t = U' _t (1) (t ≧ 2) R ' _t = R _t-1 ∧ (U' _t ∨S ' _t ) (2) ・ Re-search: (T = 1) R ' _t = U' _t (3) (t ≧ 2) R ' _t = R _t-1 ∨ (U' _t ∨S ' _t ) (4) R' _t is the t-th search Is a search condition input to the keyword search device. R ' _t-1 is the value of the previous (t-
This is the search condition used in the (first) search. _U't is U
_This is an expression in which "," which is an element of _t , is connected by a Boolean operator.
S _'t is an expression bound in ∨ all of the keywords that are included in the S _t.

Next, the process of modifying the keyword derivation rule in step 140 will be described. First, the terms used are explained. Add rules (k ₁ , k ₂ , def_value, v
alue): If there is an item (k ₁ , k ₂ , v) in the rule base, change this item to (k ₁ , k ₂ , v +
value).

Correction of rules (k ₁ , k ₂ , value
e): If there is an item (k ₁ , k ₂ , v) in the rule base, replace this item with (k ₁ , k ₂ , v + val)
ue). If there is no item (k ₁ , k ₂ , v), nothing is performed. Num (G) The number of elements included in the set (G).

The system modifies the keyword derivation rule using R _t−1 , U _t , S _t , and D _t . 1. For all combinations of any two of the elements in U _t , the rule tuika, u _j , def_va
lue1, value1). (num (U _t ) C ₂ rules added) For all combinations of elements in the R _t-1 and U _t, additional rules (r _i, u _j, def_valu
e2, value2).

(Num (R _t-1 ) C ₁ × num (U ₁ ) C ₁ rule added) For all combinations of elements in the R _t-1 and S _t, modification of rules _{(r i, s j, +} value3)
I do. (Modification of num (R _t-1 ) C ₁ × num (s _i ) C ₁ rule) For all combinations of elements in the R _t-1 and D _t, modify the rule _{(r i, d j, -value4} )
I do.

(Num (R _t-1 ) C ₁ × num (D _t ) C ₁ Rule Modification) Finally, the search execution processing in step 150 will be described. To search for the search target database using the search condition generation processing filters R _'t generated in (step 130). Note that the search process can use a normal keyword search device as it is.

[0038]

FIG. 4 shows the configuration of an information retrieval apparatus according to one embodiment of the present invention. In the following example, an information search device is realized using “WWW” widely used on the Internet. The user inputs search conditions using a WWW browser and realizes the input. The user inputs search conditions using a WWW browser, and the input search conditions are transmitted to the information search device of the present invention using WWW communication means. The transmitted search condition is processed according to the procedure described below. The processing result is transmitted to the user's WWW browser using the WWW communication procedure, and is displayed to the user.

The information retrieval apparatus shown in FIG. 1 includes a morphological analysis unit 210, a WWW processing unit 220, a retrieval condition generation unit 230,
Information search control unit 240, search processing unit / result conversion unit 25
0, a derived rule modification processing unit 260, a keyword derivation processing unit 270, a search target database 280, and a rule base 290. The morphological analysis unit 210 receives the additional search condition b from the search condition generation unit 230, analyzes it, converts it into a set of words, extracts keywords, and combines the results using an appropriate Boolean operator. Then, this part is transmitted to the search condition generation unit 230 as a search condition analysis result d.

The WWW processing unit 220 displays the information on the user's WWW browser in the format shown in FIG. 5 and receives the previous search condition a, the additional search condition b, and the keyword selection information c by operating the user. This is a part that transmits the information to the information search control unit 240 and, as a result, receives the search result f, the derived keyword g, and the search condition e, and transmits to the user's WWW browser for display.

The search condition generation unit 230 receives the previous search condition a, the additional search condition b, and the keyword selection information c from the information search control unit 240, and actually performs a search using the morphological analysis unit 210. This is a part for generating the search condition e and transmitting it to the information search control unit 240. Information search control unit 2
40 is a WWW processing unit 220, a search condition generation unit 230,
This section communicates with the search processing section / result conversion section 250, the derivation rule correction processing section 260, and the keyword derivation processing section 270, and performs processing for generating a search condition e, executing a search, and deriving a keyword.

The search processor / result converter 250 receives the search condition e from the information search controller 240, transmits it to the search target database 280, executes the search, and displays the result on the WWW browser. This is a part which is converted to a possible format and transmitted to the information search control unit 240. The derivation rule correction processing unit 260 is a part that receives the previous search condition a, the additional search condition b, and the keyword selection information c from the information search control unit 240, corrects the keyword derivation rule, and stores it in the rule base 290. .

The keyword derivation processing unit 270 is a part that receives the search condition e from the information search control unit 240, executes a keyword derivation process, and transmits the derived keyword g to the information search control unit 240. The search target database 280 receives a search condition from the search processing unit / result conversion unit 250, executes a search, and stores the search result in the search processing unit /
This is a part to be transmitted to the result converter 250.

The rule base 290 is a database for storing, retrieving, and modifying keyword derivation rules. The keyword derivation rule has the form (key1, key2, value). key1 is a keyword as input,
key2 is a keyword derived from key1, and value indicates the degree of association between key1 and key2. The rule base 290 can output all the derived rules related to key1 by inputting key1, and can output the relevance value of these words by inputting key1 and key2.

Next, information exchanged in each block will be described. a: The previous search condition is the search condition used in the immediately preceding search, and is obtained by combining words serving as keywords with a Boolean operator. In the first search, an empty character string is given. b: Additional search condition is a search condition newly input by the user in the current search. Information is added to the format of the natural sentence, indicating whether this is a “refine search” or a “re-search”.

C: The keyword selection information is the user's WW
This is a set of derived keywords displayed on the W browser and information on whether or not each of the keywords has been selected by the user. d: The search condition analysis result is obtained by performing a morphological analysis on the additional search condition b input by the user in the form of a natural sentence, extracting a keyword word, and combining the extracted words with a Boolean operator.

E: The search condition is a search condition for actually executing a search, and is obtained by combining words serving as keywords with a Boolean operator. f: The search result is the result of the search. User W
It is expressed in a format that can be displayed on a WW browser. g: Derived keywords are a set of keywords derived using the keyword derivation rules in the rule base 290.

FIG. 5 shows an example of a screen of the user interface of the information retrieval apparatus according to one embodiment of the present invention. This is generated by the WWW processing unit 220 in FIG.
This is displayed on a WW browser. The additional search condition input unit 310 is a part where the user inputs a new search condition in the form of a natural sentence.

The previous search condition display section 320 is a part for displaying the search conditions used in the previous search. The derived keyword display / selection unit 330 is a part for displaying keywords derived from the keywords included in the previous search condition, and the user clicks a check box given to these keywords to execute the current search. Can select whether or not to select these keywords.

The search type selection section 340 is a part for instructing whether the current search is “narrow search” or “re-search”. The search result display section 350 is a part for displaying the result of the previous search. This is displayed as WWW link information. Next, the operation of this embodiment will be described in order.

(1) The WWW processing unit 220 receives the previous search condition a, the additional search condition b, and the keyword selection information c from the user, and transmits them to the information search control unit 240. (2) The information search control unit 240 transmits the received information to the search condition generation unit 230. (3) The search condition generation unit 230 converts the information in the form of a natural sentence of the additional search condition b in the received information into the morphological analysis unit 2
Send to 10.

(4) The morphological analysis unit 210 analyzes the received natural sentence, extracts words, sets the words as a set of keywords, combines them with a Boolean operator, and
Send to 0. (5) The search condition generation unit 230 includes the morphological analysis unit 210
In the analysis result d of the search condition received from the above, the previous search condition a received in the above (3), the information of “refined search” or “re-search” in the additional search condition, and the keyword Using the set of keywords selected by the user in the selection information c, the above-described search condition generation processing (step 1 in FIG. 3)
30) is performed to generate a search condition e and transmit it to the information search control unit 240.

(6) The information search control unit 240 transmits the received search condition e to the search processing unit / result conversion unit 250. (7) The search processing unit / result conversion unit 250 executes a search on the search target database 280 using the received search condition e, receives the search result, and performs a search having a format that can be displayed on a WWW browser. The result is converted to the result f and transmitted to the information search control unit 240.

(8) The information search control unit 240 converts the search condition e used in the current search into the keyword derivation processing unit 2
70. (9) The keyword derivation processing unit 270 accesses the rule base 290 for each keyword included in the received search condition e, performs the above-described keyword derivation processing (step 110 in FIG. 3), and derives the keyword. A keyword g is generated and transmitted to the information search control unit 240.

(10) The information search control unit 240 adds the search condition e used this time and the derived keyword g to the search result f and transmits the result to the WWW processing unit 220 to generate the user interface screen of FIG. 5 again. And the user's WWW
Display it in the browser. (11) The information search control unit 240 transmits the previous search condition a, the additional search condition b, and the keyword selection information c to the derived rule correction processing unit 260. The derivation rule correction processing unit 260 performs the above-described keyword derivation correction processing (step 140 in FIG. 3) using these, and stores the correction result in the rule base 290.

As described above, according to the present invention, it is possible to solve the problem of notation fluctuation and synonyms. In detail, using the examples of “computer”, “computer”, and “computer” illustrated above, it will be described that the present invention has the effect of solving the problem of the notation and synonyms. It is assumed that (computer, computer, degree of relevance “large”) (computer, computer, degree of relevance “large”) is prepared as a keyword derivation rule.

The user entered "computer" as the first search condition and executed a search. However, since the keyword contained in the information being searched for was "computer" or "calculator", a search was made. It is assumed that there is no information. When the second search is performed using the present invention, the following processing is executed. (1) Keyword derivation processing (FIG. 3, step 110)
Thus, a “computer” and a “computer” are generated.

(2) In the derived keyword selection process (step 120 in FIG. 3), the user is given a "computer"
And "Calculator" are displayed. Here, it is assumed that the user has selected both keywords. (3) Since no information was hit in the first search, the next search is “re-search”. Therefore, in the search condition generation process (FIG. 3, step 130), 2
"Computer @ (computer @ computer)" is generated as the search condition for the second time.

(4) Search execution processing (step 1 in FIG. 3)
At 50), a search for "computer @ (computer @ computer)" is performed. As a result, information including "computer" or "computer" can be obtained as a search result, and the problem of notational fluctuation and synonyms can be solved. Further, according to the present invention, it is possible to solve the problem that an appropriate related word cannot be derived. In other words, using the example of CPU exchange, the present invention states that "appropriate related words cannot be derived by using a manual or a static thesaurus created using co-occurrence relations." It is possible to solve the problem.

The reason will be described below. Here, it is assumed that no keyword derivation rule has been registered yet. When a certain user sets "CP
Suppose that a search is executed by inputting "U", but there are too many documents containing "CPU" and it is not known which is related to "replacement of CPU". At this time, in the present invention, it is not possible to derive an appropriate related word as in the case of using a static thesaurus created manually or using a co-occurrence relationship. However, if "exchange" is added as a search condition as a result of the user's thinking, (CPU, exchange, relevance "large") is accumulated as a keyword derivation rule. Therefore, after that, a user who has the same intention (want to obtain a document relating to the replacement of the CPU) similarly receives “CPU” as a search condition.
Then, "exchange" is presented to the user as a related word, and the search execution process (FIG. 3, step 15) is performed.
In 0), a search for "CPU @ exchange" is performed. As described above, in the present invention, the thesaurus is dynamically configured, so that it is possible to derive a more appropriate degree of association than in the case of using a static thesaurus created manually or using a co-occurrence relationship. it can.

Further, according to the present invention, it is possible to solve the problem that an incorrect conversion rule of a search condition is stored. Specifically, the keyword derivation rule of the present invention is dynamically modified using information as to whether or not the user has actually used the rule, and does not depend on the result of the search. Therefore, even if a semantically incorrect keyword derivation rule is stored, the related word is not selected by the user, so that the degree of relevance is gradually reduced and its influence is gradually reduced. Therefore, it is possible to solve a problem that an erroneous conversion rule of a search condition is stored.

Although the above embodiment has been described with reference to FIGS. 3, 4, and 5, the present invention is not limited to this example, and the configuration shown in FIGS. 4, 5 and the series shown in FIG. The present invention is constructed as a program and stored in a portable storage medium such as a disk device, a floppy disk, or a CD-ROM connected to a device used on the user side or a computer used as an information search device. The present invention can be easily realized by installing it at the time of implementation.

It should be noted that the present invention is not limited to the above-described embodiment, but can be variously modified and applied within the scope of the claims.

[0064]

As described above, according to the present invention, an appropriate search can be executed even when a keyword input by a user is not appropriate. In addition, it is possible to appropriately operate the function for supporting the user's search.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining the principle of the present invention.

FIG. 2 is a principle configuration diagram of the present invention.

FIG. 3 is a flowchart illustrating an outline of a search according to the present invention.

FIG. 4 is a configuration diagram of an information search device according to an embodiment of the present invention.

FIG. 5 is an example of a screen of a user interface of the information search device according to one embodiment of the present invention.

[Explanation of symbols]

 200 User search condition input unit 210 Morphological analysis unit 220 WWW processing unit, user selection instruction unit 225 Search result display unit 230 Search condition generation unit 240 Information search control unit 250 Search processing unit / result conversion unit, search unit 255 Re-search unit 260 Derived rule correction processing unit, rule base correction unit 270 Keyword derivation processing unit, other keyword acquisition unit 280 Search target database 290 Rule base 310 Additional search condition input unit 320 Previous search condition display unit 330 Derived keyword display / selection unit 340 Search type selection Part 350 Search result display part

Claims (16)

[Claims] 1. A keyword search method for executing a search related to a keyword input by a user or a search condition composed of an expression obtained by combining keywords with a Boolean operator. Searches a rule base that can search, save, and modify information indicating the degree of relevance to any keyword, including the keyword, and searches for a keyword included in the search condition entered by the user. The keyword that is more than a predetermined value and is likely to be registered in the rule base, displays the keyword acquired from the rule base to the user, and performs an actual search from the keyword. The user selects a keyword to be used, and sets the keyword selected by the user to a search condition input by the user. Additionally performing a search, modifying the rule base based on a keyword selected by the user, and displaying a result of the search to the user. 2. When there are a plurality of keywords input by the user, for a combination in which information on the relevance of the keyword does not exist in the rule base, a predetermined value is used as the relevance in the rule base. 2. A method for modifying a rule base for storing and existing combinations to increase the relevance of these keywords.
The described keyword search method. 3. For all combinations of the keyword input by the user and the keyword selected by the user, information on the degree of association between the keywords exists in the rule base. 2. The keyword search method according to claim 1, wherein a correction process for increasing the degree of relevance between the keywords is performed. 4. For all combinations of the keyword input by the user and the keyword not selected by the user, information on the degree of association between the keywords is
The keyword search method according to claim 1, wherein, for a combination existing in the rule base, a correction process is performed on the rule base to reduce the degree of association between the keywords. 5. The keyword search method according to claim 1, wherein a related keyword is acquired and searched at the next and subsequent searches using the rule base on which the correction processing has been performed. 6. A keyword search device for performing a search related to a search condition composed of a keyword input by a user or an expression obtained by combining keywords with a Boolean operator, wherein: an arbitrary keyword used in the search; A rule base capable of searching, storing, and modifying information indicating a degree of association with an arbitrary keyword including the arbitrary keyword; a user search condition input means for receiving an input of a search condition from a user; A search target database to be searched; a search unit that searches the search target database based on the search condition input by the user search condition input unit; and a search condition that performs the rule-based search and is input by the user. The degree of relevance is greater than or equal to a predetermined value for the keyword contained in Other keyword acquiring means for acquiring a keyword that may be registered in the rule base, a user selection instructing means for displaying a keyword acquired from the rule base to a user, and selecting a keyword to be used for an actual search from the keyword. Re-search means for adding a keyword selected by the user to a search condition input by the user by the user selection instructing means and executing a search; and a rule for correcting the rule base based on the selected keyword A keyword search device comprising: a base correction unit; and a search result display unit that displays a result of the search to the user. 7. The keyword search device according to claim 6, wherein the rule base includes an input keyword as an input, a derived keyword derived from the input keyword, and a degree of association between the input keyword and the derived keyword. 8. The rule-base modifying unit, when there are a plurality of keywords input by the user, sets a predetermined value for a combination in which information on the degree of association between the keywords does not exist in the rule base. As the degree of relevance, the first combination is stored in the rule base, and for the existing combinations, a first modification process is performed on the rule base to increase the degree of relevance between these keywords.
7. The keyword search device according to claim 6, further comprising a rule base correction unit. 9. The combination in which information on the degree of relevance between keywords exists in the rule base for all combinations of the keyword input by the user and the keyword selected by the user. With respect to the rule base,
7. The keyword search device according to claim 6, further comprising a second rule-based correction unit that performs a correction process for increasing the degree of association between the keywords. 10. The rule-base modifying means, for all combinations of the keyword input by the user and the keyword not selected by the user, information on the relevance between the keywords is stored in the rule base. 7. The keyword search device according to claim 6, further comprising a third rule base correction unit that performs a correction process on the existing rule base to reduce the degree of association between the keywords. 11. The keyword search method according to claim 8, wherein the search unit includes a search unit that obtains a related keyword at the next and subsequent searches using the rule base on which the correction processing has been performed, and performs the search. Search device. 12. A keyword input by a user, or
A storage medium storing a keyword search program for executing a search related to a search condition composed of an expression in which keywords are combined by a Boolean operator, wherein a user search condition input process for receiving an input of a search condition from a user; A search process for searching the search target database based on the search condition input by a condition input process; an arbitrary keyword used in the search process; and a degree of association between any keyword including the arbitrary keyword. A rule-based search process for searching for a rule base that can search, save, and modify the information indicating, and a search is performed by the rule-based search process, and for a keyword included in the search condition input by the user, , The degree of association is greater than or equal to a predetermined value, A keyword acquiring process for acquiring a keyword that may be registered in the rule base; and a user displaying a keyword acquired from the rule base to a user and selecting a keyword to be used for an actual search from the keyword. A selection instructing process; a re-searching process of adding a keyword selected by the user to a search condition input by the user by the user selecting instructing process to execute a search of a search target database; A storage medium storing a keyword search program, comprising: a rule base correction process for correcting the rule base by using a search result display process for displaying the result of the search to the user. 13. The rule-based modification process, wherein, when there are a plurality of keywords input by the user, a predetermined value is set for a combination in which information on the degree of association between the keywords does not exist in the rule base. 13. The method according to claim 12, further comprising: a first rule-based correction process for storing a relevancy in the rule base and performing a correction process for the existing combination to increase the relevancy between these keywords. A storage medium storing the described keyword search program. 14. The rule-based modification process may include, for all combinations of the keyword input by the user and the keyword selected by the user, a combination in which information on the relevance between the keywords exists in the rule base. With respect to the rule base,
13. The storage medium storing the keyword search program according to claim 12, further comprising a second rule-based correction process for performing a correction process for increasing the degree of association between the keywords. 15. The rule-based modification process includes, for all combinations of the keyword input by the user and the keyword not selected by the user, information on the relevance between the keywords in the rule base. The method according to claim 1, further comprising a third rule-based modification process for performing a modification process on the existing combination to reduce the degree of association between the keywords.
A storage medium storing the keyword search program according to Item 2. 16. The search process uses the rule base on which the correction processing has been performed, at the time of the next and subsequent searches.
16. The storage medium storing the keyword search program according to claim 12, further comprising a process of obtaining a related keyword and performing a search.