JPH10334101A - Retrieval logical expression generating device, retrieval system, and record medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable speedy retrieval by narrowing down documents. SOLUTION: A morpheme an analyzing device 11 takes a morpheme analysis of a retrieval request sentence into words of parts of speech. A retrieval operator determination device 12 determines a retrieval key word by a retrieval key word determination part 20. A decision part 21 decides whether or not the retrieval key word constitutes a noun phrase. A process part 22 combines noun phrases with a near operator NEAR1 or NEAR2 to generate a retrieval logical expression. A document retrieval device 2 retrieves a document corresponding to the inputted retrieval logical expression.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a retrieval logical expression generation device, a retrieval system, and a recording medium, and more particularly, to a method for retrieving desired information more reliably.
The present invention relates to a search logical expression generation device, a search system, and a recording medium.

[0002]

2. Description of the Related Art FIG. 20 shows a configuration example of a conventional search system. A search request sentence is input to the search logical expression generation device 1. This search request sentence is in a language used everyday, such as “I want to know about the production of personal computers in Asia”. In this way, by enabling the user to perform a search in a natural language that is used daily, the burden on the user for creating a search request sentence can be reduced.

[0003] The search request sentence is divided into various parts of speech words by morphological analysis in the search logical formula generation device 1. Nouns are used as search keywords, and conjunctions and particles are AND or OR of search operators (search operators).
And, for example, the following search logical expression is generated. Asian AND PC AND production

[0004] The search formula thus generated is:
The document is input to the document search device 2 and a document corresponding to the search logical expression is searched.

Various methods have been proposed for generating a search logical expression from a natural language. For example, in “Japanese Query Sentence Analysis for Patent Search 1984, IPSJ Transactions Vol.25 No.3”, connectives are classified into those that are replaced with search operators AND and those that are replaced with ORs. In logical expressions, keywords are combined according to classification. Also, keywords that are not connected by a conjunction in the search request sentence are implicitly connected by AND.

Further, for example, see Japanese Patent Application Laid-Open No. 5-189492.
In the prior art section of the publication, it is disclosed that a search logical expression is generated by replacing the case particle "no" with AND, and replacing the parallel particle "to" with OR.

[0007]

However, in such a method, for example, as shown in FIG.
If the search keywords “Asia”, “PC”, and “Production” are present in relatively short sentences, they are searched as documents that match the search formula. Also, as shown in FIG. 22, when a plurality of topics are introduced in one document,
There was a problem that the search was performed as a match with the search logical expression. As a result, there is a problem in that the search target cannot be sufficiently narrowed, and even if there is no relation, even if there is a thin document, it is also searched.

[0008] The present invention has been made in view of such a situation, and it is an object of the present invention to be able to narrow down and search a desired document more reliably.

[0009]

According to a first aspect of the present invention, there is provided a search logical expression generating apparatus which determines a search keyword based on a result of morphological analysis of a character string of a search request, and wherein the search keyword forms a noun phrase. A determination unit for determining whether or not to perform the search, and a generation unit for generating a search logical expression by combining search keywords constituting the noun phrase determined by the determination unit with a nearby operator.

[0010] The neighborhood operator is an operator that performs an operation that is true when a plurality of given search keywords are within a predetermined neighborhood, and that is false otherwise.

That is, the term (K1NEAR) in which the search keyword K1 and the search keyword K2 are connected by the neighborhood operator NEAR
K2) has a value of 1 (true) when K1 and K2 have a predetermined neighborhood relation in the text to be processed, and has a value of 0 (false) when they have no predetermined neighborhood relation.

There are various types of predetermined neighborhood relations.

The neighborhood relation will be described in the order of the neighborhood coupling strength. (1) Strongest neighborhood coupling: when K1 and K2 exist continuously (2) Next strongest neighborhood coupling after (1): 1 between K1 and K2
When other characters are included only in the character (3) Next strongest neighbor connection after (2): N between K1 and K2
When other characters are included only in character (4) Next strongest neighbor bond after (2): 1 between K1 and K2
(5) Next strongest neighbor connection after (2): N between K1 and K2
(6) Next strongest neighbor connection after (2): K1 and K2 belong to the same sentence (7) Next strongest neighbor connection after (2): K1 and K2 are in the same paragraph If you belong

Among the above-mentioned neighborhood couplings, the strength of the neighborhood coupling is stronger in (6) than in (7).

The judgment means includes a range enclosed by quotation marks, a range in which a noun or proper noun is continuous before or after a proper noun with or without the particle "no", an adjective or an adjective verb. A continuous range or a range in which the noun includes or does not include the particle "no" can be determined as a noun phrase.

[0016] The generating means can combine the search keywords constituting the noun phrase with a first neighboring operator having a stronger binding force or a second neighboring operator having a weaker binding force in accordance with the range. .

When the noun phrase is in a range in which the noun is continuous with or without the particle "no", the generation unit causes the particle "no" to be replaced by a second neighborhood operator, Nouns can be combined with a second neighborhood operator.

When the noun phrase is a range in which the noun or proper noun is continuous before or after the proper noun including or not including the particle "no",
Can be replaced with a first neighborhood operator, and proper nouns or nouns can be combined with the first neighborhood operator.

When the noun phrase is in a range enclosed by quotation marks, the generating means can combine the search keywords constituting the noun phrase with a first neighborhood operator.

Analysis means for morphologically analyzing the character string of the search request can be further provided.

[0021] A search system according to an eighth aspect of the present invention is characterized in that the search system is provided with a search means for performing a search using the search logical expression generated by the generation means.

According to a ninth aspect of the present invention, in the recording medium, a determination step of determining a search keyword from a result of morphological analysis of a character string of a search request, and a determination step of determining whether the search keyword forms a noun phrase And a generation step of generating a search logical expression by combining search keywords constituting the noun phrase determined in the determination step by a nearby operator.

In the determining step, a range enclosed by quotation marks, a range in which the noun or proper noun is continuous before or after the proper noun with or without the particle "no",
A range in which adjectives or adjective verbs are continuous, or a range in which a noun continues with or without the particle "no" can be determined as a noun phrase.

In the generation step, a search keyword constituting a noun phrase is first matched with the range by a first keyword having a stronger binding force.
Can be connected by a nearby operator having a lower strength or a second neighbor operator having a weaker bonding force.

In the generation step, if the noun phrase is a continuous range of the noun including or not including the particle “no”, the particle “no” is replaced by a second neighborhood operator, and Can be combined by a second neighbor operator.

In the generation step, if the noun phrase is a range in which the noun or proper noun is continuous before or after the proper noun including or not including the particle "no", the particle "no" is added to the first noun. With the neighbor operator of
Proper nouns or nouns can be combined with a first neighborhood operator.

In the generation step, when the noun phrase is in a range surrounded by quotation marks, the search keywords constituting the noun phrase can be combined by the first neighborhood operator.

An analysis step of morphologically analyzing the character string of the search request can be further provided.

[0029] In the retrieval logical expression generation device according to the first aspect and the recording medium according to the ninth aspect, the retrieval keyword is determined from the result of the morphological analysis of the character string of the retrieval request. It is further determined whether or not the search keyword forms a noun phrase, and the search keywords forming the noun phrase are combined by a nearby operator to generate a search logical expression.

In the search system according to the present invention, the search is performed using the search logical expression generated in this manner.

[0031]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below. In order to clarify the correspondence between each means of the invention described in the claims and the following embodiments, each means is described. When the features of the present invention are described by adding the corresponding embodiment (however, an example) in parentheses after the parentheses, the result is as follows. However, of course, this description does not mean that each means is limited to those described.

[0032] The retrieval logical expression generation device according to claim 1 is
Determining means for determining a search keyword from the result of morphological analysis of the character string of the search request (for example, the search keyword determining unit 20 in FIG. 1), and determining means for determining whether the search keyword forms a noun phrase ( For example, the determination unit 2 in FIG.
1) and generating means (for example, the processing unit 22 in FIG. 1) for generating a search logical expression by combining search keywords constituting a noun phrase determined by the determination means with a nearby operator. I do.

FIG. 1 shows an example of the configuration of a search system according to the present invention. In this search system, a search request sentence composed of a natural language is input to the search logical expression generation device 1. This search request sentence is input to the morphological analysis device 11 of the search logical expression generation device 1 and subjected to morphological analysis. Then, the result of the morphological analysis is input to the search operator determination device 12 of the search logical expression generation device 1, and the search operator is determined.

The search operator determining device 12 has a search keyword determining unit 20 for determining a search keyword from the morphological analysis result supplied from the morphological analyzer 11. The search keyword output from the search keyword determination unit 20 is input to the determination unit 21, and it is determined whether or not conditions for applying various search operators are satisfied. The result determined by the determination unit 21 is supplied to the processing unit 22, where a search operator application process is performed in accordance with the determination result, and a search logical expression is generated.

The search logical expression generated by the search operator determining device 12 is input to the document search device 2, and the document search device 2 searches for a document according to the input search logical expression.

Next, referring to the flowchart of FIG.
The operation will be described. First, in step S1, the user inputs a search request sentence composed of a natural language to the search logical expression generation device 1. For example, a search request sentence “I want to know about Asian PC production” is input.
Next, in step S2, the morphological analyzer 11
The morphological analysis of the input search request sentence is performed. Thereby, the input search request sentence is decomposed into words, and the part of speech of each word is also specified.

For example, in this case, a morphological analysis result as shown in FIG. 1 is obtained. In this example, "Asia" is a proper noun, "No" is a particle, "PC" and "Production" are nouns, "Ni" and "About" are particles, "Intelligence" is a verb, "Ri"
And "tai" are auxiliary verbs.

In addition, as the part of speech to be specified, FIG.
The following can be mentioned. A known method can be used for the morphological analysis.
This morphological analysis is described in, for example, April 2, 1996.
It is disclosed on pages 118-125 of "Natural Language Processing" of Iwanami Koza Software Chemistry 15 published by Iwanami Shoten on the 6th.

As described above, the result of the morphological analysis is input from the morphological analysis device 11 to the search keyword determination section 20 of the search operator determination device 12. The search keyword determination unit 20 sets the noun, proper noun, adjective, and adjective verb set as defaults in step S3 as search keywords. The user can specify any part of speech as a search keyword as needed.

The determination unit 21 includes a search keyword determination unit 20
When the result of the morphological analysis including the search keyword is input from, it is determined in step S4 whether a conjunction or a particle corresponding to the search operator OR is included in the morphological analysis result. The conjunction equivalent to OR is, for example, "or",
"Or", "or", or "or". The particle equivalent to OR is “KA”.

If such a conjunction or particle equivalent to OR is included, the result of the morphological analysis is input to the processing unit 22, and in step S5, a process of replacing the conjunction or particle with OR is executed.

For example, as shown in FIG. 4, when a search request sentence of “new graduate worker or unemployment candidate” is input, this “or” is replaced with OR and “new graduate worker O”
A search logical expression of “R job vacant” is generated. FIG.
In the table, "new graduates employed" and "employee in employment" are set as search logical expressions of "new graduate NEAR1 employed" and "employment NEAR1 unapproved", respectively. This process is performed in step S17 described later. Done.

If it is determined in step S4 that a conjunction or particle corresponding to OR is not included in the result of the morphological analysis, or in step S5, the conjunction or particle is ORed.
After execution of the process of replacing
The determination unit 21 determines whether a quotation mark is included in the morphological analysis result. The quotes are, for example, as shown in FIG. 3 as open and closed parentheses.
In the case of this embodiment, the range enclosed by quotation marks is a noun phrase.

If it is determined in step S6 that the morphological analysis result includes a quotation mark, the process proceeds to step S7.
Then, it is determined whether or not the quotes include an auxiliary verb or a particle other than “no”. If the quotation mark contains no auxiliary verb or a particle other than "no", the process proceeds to step S8. If the particle "no" is included, it is replaced with the strongest neighbor NEAR2. . Further, nouns or proper nouns included in the quotation marks are combined by the neighborhood operator NEAR2.

The neighborhood operator is described on page 416 of the above-mentioned document.

For example, as shown in FIG. 5, when a search request sentence of “Company that sells“ Golf Competition Insurance ”” is entered, “Golf Competition Insurance” is enclosed in quotation marks. As a result, a search logical expression of “Golf NEAR2 Competition NEAR2 Insurance” is generated. In FIG. 5, a search logical expression of “insurance AND release AND company” is also generated.
8 is generated.

In addition, for example, as shown in FIG.
If a search request sentence of “Process Development Center of Company A” is entered, the search formula of “Process NEAR2 Development NEAR2 Center” is generated because “Process Development Center” is enclosed in quotation marks. Is done. Note that “A” in FIG.
The search logical expression of "company AND process" is generated in step S18 described later.

Further, as shown in FIG. 7, even when a search request sentence "about" Semiconductor supply center for personal computer production in Asia "is input," Semiconductor supply center "is enclosed in quotation marks. Has been described, "Semiconductor NEAR2
A search formula for "supply NEAR2 center" is generated. Note that the logical expression of “Asia AND personal computer” and “Production AN”
The search logical expression of "D semiconductor" is generated in step S18, and the search logical expression of "production of personal computer NEAR1" is generated in step S17 described later.

On the other hand, as shown in FIG. 8, for example, when a search request sentence of “Process for preparing a crisis management response by the government” is input, this search request sentence is enclosed in quotation marks. Particles other than "no" are included. Therefore, a determination of Yes is made in step S7, and the process of step S8 is skipped. That is, the connection by NEAR2 is not performed. The search logical formula of “government AND crisis” and “correspondence AND process” in FIG. 8 is generated in step S18, and the search logical formula of “crisis NEAR1 management NEAR1 correspondence” is generated in step S17.

If it is determined in step S6 that the result of the morphological analysis does not include a quotation mark, step S7
When it is determined that the quotation mark contains an auxiliary verb or a particle other than "no", or after the generation of the search logical expression by NEAR2 in step S8, Proceeding to S9, it is determined whether the morphological analysis result includes a proper noun. When the proper noun is included, the process proceeds to step S10, and before and after the proper noun, a range in which the noun or proper noun is continuous without including the particle, or continuously including the particle "no". Is determined. In the case of this embodiment, this range is also a noun phrase. Then, in step S11, the particle "no" is changed to NEAR2 in the range determined in step S10.
Is performed. Also, nouns and proper nouns
The combining process is performed in NEAR2.

For example, as shown in FIG. 9, when a search request sentence of “about Lake Biwa Science Award” is entered, since the proper noun “Lake Biwa” is included, the “Lake Biwa Science Award” is entered. It is extracted as a continuous range (noun phrase) including proper nouns, and these are combined by the neighborhood operator NEAR2 to generate a search logical expression as shown below. Lake Biwa NEAR2 Science NEAR2 Award

In addition, for example, as shown in FIG.
When a search request sentence of “product sold by Yamada Shoji” is input, “Yamada” is a proper noun, so “Yamada Shoji” is extracted as a range of nouns following the proper noun, and “Yamada” "Trading" is joined by nearby operator NEAR2,
A search logical expression of “Yamada NEAR2 Shoji” is generated. In addition,
The search logical expression of “commercial AND product” is generated in step S18.

In step S9, when it is determined that the proper noun is not included in the morphological analysis result, or after the processing by the nearby operator NEAR2 in step S11 is completed, the process proceeds to step S12, and the adjective or adjective is included in the morphological analysis result. It is determined whether a verb is included. If an adjective or an adjective verb is included, step S
Proceeding to 13, the adjective is standardized to the final form, and the adjective verb is standardized only to the stem.

For example, as shown in FIG. 11, when a search request sentence “Company-wide deployment of personal computers” is input, this search request sentence includes an adjective “company-wide”. I have. This adjective verb is standardized only to the "company-wide" stem.

Next, the process proceeds to step S14, in which a range including nouns including the adjective or the adjective verb is determined.
In the case of this embodiment, this range is also a noun phrase. In the case of the example of FIG. 11, “company-wide deployment of personal computers” is determined as a range in which nouns are continuous. Next, step S15
In the determined range, the noun, adjective, and adjective verb are changed to the neighborhood operator NE weaker than the neighborhood operator NEAR2.
The process of combining with AR1 is performed. As a result, in the case of the example shown in FIG. 11, a search logical expression of “PC NEAR1 company-wide NEAR1 development” is generated.

As described above, after the processing in step S15 is completed, or when it is determined in step S12 that the result of the morphological analysis does not include an adjective or an adjective verb, the process proceeds to step S16, and the noun is changed. A continuous range including no particles or a continuous range including nouns “no” is determined. In the case of this embodiment, this range is also a noun phrase. Then, the process proceeds to step S17,
If the particle "no" exists in the determined range, it is replaced by the nearby operator NEAR1. The noun is NEAR1
Are combined. For example, as shown in FIG. 12, when a search request sentence "strengthening the main business of the leasing industry" is input, "strengthening the main business of the leasing industry" is determined as a continuous range (noun phrase) including the particle "no". Is done. And the particle "no"
It is replaced by NEAR1 and each noun is also joined by NEAR1, so
As a result, the following search formula is generated. Leasing NEAR1 industry NEAR1 main business NEAR1 reinforcement

In addition, for example, as shown in FIG.
When a search request sentence of "PC production in Asia" is entered, "PC production" is extracted as a continuous range across the particle "NO", and a search formula for "PC NEAR1 production" is generated. You.

Further, as shown in FIG. 14, when a search request sentence of "PC production in Asia" is input, "PC production" is extracted as a range in which nouns are continuous without interposing particles. The search logical expression of "PC NEAR1 production" is generated.

Next, the process proceeds to step S18, in which processing is performed to replace words of parts of speech (words not used as search keywords) other than nouns, proper nouns, adjectives, and adjective verbs with the search operator AND. For example, of the parts of speech shown in FIG.
Parts of speech that are not search keywords, such as adverbs, adverbs, conjunctions, and verbs, are replaced with AND. As a result, FIG.
Alternatively, in the search request sentence of FIG.
D is substituted.

Next, the process proceeds to step S19, and if two or more ANDs are consecutive in the search logical expression generated as a result of the processing up to that point, a process of reducing the number to one is executed. If AND is at the beginning or end of the search logical expression, it is deleted.

The search logical expression generated as described above is output from the processing unit 22 to the document search device 2 in step S20. If the search request sentence "I want to know about PC production in Asia" is entered, "Asian AN
The search logical expression of "D personal computer NEAR1 production" is output.

In the document search device 2, a document is searched according to the input search logical expression. The search operator included in the search logical expression performs the following search.

X OR Y Search keyword X or Y
Documents containing any one of the above are searched. X AND Y Documents containing both keywords X and Y are searched. X NEAR1 Y Documents in which keyword X and keyword Y are included in one paragraph are searched. This neighborhood operator is (7) of the neighborhood operators (1) to (7) described in the section of "Means for Solving the Problem" described above.
Corresponds to a nearby operator. X NEAR2 Y Search keyword X and search keyword Y
Are sequentially searched in this order. This neighborhood operator corresponds to the neighborhood operator (1) among the neighborhood operators (1) to (7) described in the above “Means for Solving the Problem”.

The priorities of these search operators are set as follows.

NEAR2> NEAR1> AND> OR

FIGS. 15 to 19 show the search order of the search logical expressions generated as described above in numerical order from the one with the highest priority. For example, in FIG. 15, the search logical expression of “PC NEAR1 production” indicated by number 1 is searched first, and those that include the search keyword of “Asia” are further searched from those satisfying the search logical expression. .

Further, for example, in the example shown in FIG. 19, a search logic that satisfies the search logical expression of “semiconductor NEAR2 supply” is searched, and a search logic of “semiconductor NEAR2 supply NEAR2 center” is further searched from those satisfying this. Those satisfying the formula are searched, and those satisfying "production of personal computer NEAR1" are searched from those satisfying the search logical formula. In the following, those that satisfy "Asian AND personal computer NEAR1 production" are searched.
Those that satisfy "D personal computer NEAR1 production AND semiconductor NEAR2 supply NEAR2 center" are searched.

As a result, for example, when a search request sentence “I want to know about Asian PC production” is input,
As shown in FIG. 20, since a search logical expression of “Asian AND personal computer AND production” was generated, FIG. 21 and FIG.
Although both documents shown in FIG. 2 are searched, in the case of the present embodiment, as shown in FIG. 1, a search logical expression of “Production of Asian AND personal computer NEAR1” is generated. Documents such as those shown in FIG.
As shown in (1), documents that include the search keywords of “PC” and “production” but are not included in the text in one paragraph are not searched. As a result, it is possible to narrow down the search and search.

In the above description, the case of searching for Japanese sentences has been described as an example. However, the present invention can be applied to the case of searching for sentences other than Japanese.

The program for performing the above processing can be recorded on a recording medium such as a magnetic disk or a CD-ROM disk and provided to a user or provided via a medium such as a network. Of course, it is also possible to record in advance on a recording medium such as a hard disk or a solid-state memory built in the search logical expression generation device 1.

In the above embodiment, the near operator NEAR2 is given priority over NEAR1, but the latter may be given priority.

[0072]

As described above, according to the retrieval logical expression generation apparatus of the first aspect and the recording medium of the ninth aspect,
Since the noun phrase is determined and the search keywords constituting the noun phrase are combined by the neighborhood operator, it is possible to narrow down a desired sentence and generate a search logical expression that can be searched quickly. .

According to the search system of the present invention, since the search is performed using the search logical expression generated in this manner, desired sentences can be narrowed down, and quickly and reliably. Can be searched.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration example of a search system according to the present invention.

FIG. 2 is a flowchart illustrating an operation of the search system in FIG. 1;

FIG. 3 is a diagram illustrating parts of speech that are subjected to elementary analysis by the morphological analyzer of FIG. 1;

FIG. 4 is a diagram illustrating an example of a search request sentence and a search logical expression.

FIG. 5 is a diagram showing an example of a search request sentence and a search logical expression.

FIG. 6 is a diagram illustrating an example of a search request statement and a search logical expression.

FIG. 7 is a diagram showing an example of a search request sentence and a search logical expression.

FIG. 8 is a diagram showing an example of a search request statement and a search logical expression.

FIG. 9 is a diagram showing an example of a search request statement and a search logical expression.

FIG. 10 is a diagram showing an example of a search request sentence and a search logical expression.

FIG. 11 is a diagram showing an example of a search request sentence and a search logical expression.

FIG. 12 is a diagram illustrating an example of a search request statement and a search logical expression.

FIG. 13 is a diagram showing an example of a search request statement and a search logical expression.

FIG. 14 is a diagram showing an example of a search request sentence and a search logical expression.

FIG. 15 is a diagram illustrating a search order.

FIG. 16 is a diagram illustrating a search order.

FIG. 17 is a diagram illustrating a search order.

FIG. 18 is a diagram illustrating a search order.

FIG. 19 is a diagram illustrating a search order.

FIG. 20 is a block diagram illustrating a configuration example of a conventional search system.

FIG. 21 is a diagram illustrating an example of a document to be searched.

FIG. 22 is a diagram illustrating an example of another document to be searched.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 search logical formula generation device 2 document search device 11 morphological analysis device 12 search operator determination device 20 search keyword determination unit 21 determination unit 22 processing unit

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Shinichi Mukakawa 10 Hanazono Todocho, Ukyo-ku, Kyoto, Kyoto Prefecture Inside of Omron Co., Ltd. OMRON Corporation

Claims (15)

[Claims] 1. A determination unit for determining a search keyword from a result of morphological analysis of a character string of a search request, a determination unit for determining whether the search keyword forms a noun phrase, and a determination by the determination unit Generating means for generating a search logical expression by combining said search keywords constituting said noun phrase by a nearby operator. 2. The method according to claim 1, wherein the determining unit includes: a range enclosed by quotation marks; a range in which the noun or proper noun is continuous before or after the proper noun including or not including the particle “no”;
The range in which an adjective or an adjective verb is continuous, or a range in which a noun is continuous with or without the particle "no" is determined as the noun phrase. Search formula generator. 3. The method according to claim 1, wherein the generating unit includes:
3. The search logical expression generation according to claim 2, wherein the search keywords forming the noun phrase are combined by a first neighbor operator having a stronger binding force or a second neighbor operator having a weaker binding force. 4. apparatus. 4. The method according to claim 1, wherein the generating unit is configured to, when the noun phrase is in a range in which the noun is continuous including or not including the particle "no", the particle "no" to the second neighborhood operator. The search logical expression generation apparatus according to claim 3, wherein the noun is replaced by the second neighborhood operator. 5. The method according to claim 1, wherein the noun phrase is a range in which the noun or proper noun is continuous before or after the proper noun including or not including the particle “no”. ”Is replaced by the first neighborhood operator, and the proper noun or noun is combined by the first neighborhood operator. 6. The method according to claim 1, wherein, when the noun phrase is in a range enclosed by quotation marks, the generation unit combines the search keywords constituting the noun phrase with the first neighborhood operator. The search logical expression generation device according to claim 3. 7. The search logical expression generation device according to claim 1, further comprising an analysis unit configured to perform a morphological analysis on the character string of the search request. 8. A search system comprising a search unit for performing a search using a search logical expression generated by the generation unit according to claim 1. Description: 9. A determination step of determining a search keyword from a result of morphological analysis of a character string of a search request, a determination step of determining whether the search keyword forms a noun phrase, and a determination in the determination step Generating a search logical expression by combining the search keywords forming the noun phrase with a nearby operator. 10. In the determining step, a range enclosed by quotation marks, a range where nouns or proper nouns are continuous before or after proper nouns with or without the particle "no", adjectives or adjectives 10. The recording medium according to claim 9, wherein a range in which a verb is continuous or a range in which a noun includes or does not include the particle "no" is determined as the noun phrase. 11. In the generating step, the search keyword forming the noun phrase is defined according to the range,
11. The recording medium according to claim 10, wherein the connection is performed by a first neighboring operator having a stronger binding force or a second neighboring operator having a weaker binding force. 12. In the generating step, when the noun phrase is a range in which the noun is continuous including or not including the particle “no”, the noun phrase is replaced with the second neighboring operator. The recording medium according to claim 10, wherein the noun is replaced by the second neighbor operator. 13. In the generating step, when the noun phrase is a range in which the noun or proper noun is continuous before or after the proper noun including or not including the particle “no”, "Is replaced by the first neighborhood operator, and the proper noun or noun is combined by the first neighborhood operator. 14. In the generating step, when the noun phrase is in a range enclosed by quotation marks, search keywords constituting the noun phrase are combined by the first neighborhood operator. The recording medium according to claim 10, wherein 15. The recording medium according to claim 9, further comprising an analysis step of morphologically analyzing the character string of the search request.