JP2002063203A - Information retrieving device, its method and recording medium storing information retrieving program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable a user to retrieve/acquire information corresponding to a request inputted by a natural sentence even when the user doesn't know the address of a server capable of retrieving the required information or an address indicating the existence of the required information. SOLUTION: The user describes his/her 'demand' as a natural sentence and transmits the natural sentence to an information retrieving device 12. The device 12 is provided with a natural sentence analysis means 14 for analyzing the inputted natural sentence by referring to a demand sorted conversion table and judging the sort of the demand, an address conversion table 16 previously storing the addresses of servers providing service corresponding to respective demand sorts, an address conversion means 15 for converting the demand sort obtained by the means 14 into the address of a server for serving the demand sort, and an address starting means 17 for including the information having the natural sentence as transmission information as required and then starting the address obtained by the conversion.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information retrieval apparatus and method, and more particularly, to an information retrieval apparatus and method which enables a user to search and acquire desired information by inputting a request in natural sentence. And a recording medium storing an information search program for implementing this method.

[0002]

2. Description of the Related Art A conventional information search apparatus receives a search key having contents corresponding to a search area of the apparatus, and searches and outputs data corresponding to the search key.

[0003] With the recent expansion of the use of the Internet, a large number of search engines (search engines) have been opened on the Internet, and such search engines also have their own search areas. And is included in the category of the information search device here.

[0004]

The conventional information retrieval device accepts only a retrieval key having a content corresponding to the retrieval area of the device. For this reason, a user who intends to search for information via a network or the like must select a server capable of searching for the desired information, access the server, and perform a search action. It is necessary to know which server can search the desired information. Unless the user specifically knows the server that will respond to his or her request, he or she must first search for such a server during a search, which is time-consuming. Also, even if you know a few servers that can meet your needs, but do not know which one is really the right one,
The user must try a search action for each server.

A first object of the present invention is to respond to a request entered in a natural sentence without the user knowing the address of a server for searching for desired information or the address itself indicating the location of the desired information. Another object of the present invention is to provide an information search apparatus and method that enable search and acquisition of corresponding information.

A second object of the present invention is to simplify the interface input by a user at the time of search as much as possible.

[0007]

According to an embodiment of the present invention, there is provided an information retrieval apparatus for converting a request described as a natural sentence into an address of a server which provides a service corresponding to the request. A search device, comprising: a first recording unit including a request type table for holding a request type corresponding to each keyword; and analyzing a natural sentence and referring to the request type table to determine a request type of a request included in the natural sentence. Natural sentence analyzing means for determining, a second recording means having an address conversion table for holding an address corresponding to each request type, and searching the address conversion table based on the request type obtained by the natural sentence analyzing means, Address conversion means for converting to the address of a server providing the service of the request type, and information including a natural sentence at the server providing the service of the request type are required. If that is on moistened the information as transmission information comprises an address activation means to start its address, the.

In this information retrieval apparatus, keyword extracting means for extracting a keyword corresponding to a request from the meaning and dependency relation of a word included in a natural sentence, and referring to the request type table, the extracted keywords are classified into the request type. It is preferable that the natural sentence analyzing unit is constituted by the request type converting unit to be converted.

According to a second aspect of the present invention, there is provided an information retrieval apparatus according to the above-described information retrieval apparatus, wherein an address including a request described as a natural sentence is input, and the input address is analyzed to extract a natural sentence. And a natural sentence extracted by the natural sentence extracting means is input to the natural sentence analyzing means.

In the present invention, a natural sentence is constituted by a word itself or including a plurality of phrases.

According to the present invention, a server for deriving a request type corresponding to a natural sentence input by a user with reference to a request type table and providing such a request type service using an address conversion table And activates the address. This allows the user to
Even without knowing the address of the server that retrieves the desired information or the address itself indicating the location of the desired information itself, it is possible to access such an address, perform the desired search, and obtain the desired information. It becomes possible.

Further, by inputting an address including a request described in a natural sentence and adding a function of analyzing the address and extracting the written natural sentence, the user can obtain the information retrieval apparatus of the present invention. Instead of inputting a natural sentence after accessing the search screen, a request described in the natural sentence can be included in the address for the information search device. In this case, it is possible to execute the search without displaying the search screen by the information search device to the user. Further, in a terminal used by the user, an address including a request described in a natural sentence is registered in a WWW (World Wide Web) browser software (so-called Internet browsing software), an e-mail software address book, or the like. By doing so, the search can be performed more easily.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, a preferred embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram illustrating a configuration of an information search device according to an embodiment of the present invention, and FIG. 2 is a diagram illustrating a connection mode to a network.

The information retrieval device 12 is connected to a network 10 such as the Internet, and the user terminal 1 used by a user via the network 10.
1 and a service server 18 that actually performs a search for each search area. As shown in FIG. 2, a plurality of service servers 18 are connected to the network 10, and each service server 18 is identified by its network address or URL (uniform resource locator). In FIG. 1, the network 10 is not shown for the sake of explanation, and the network 1 is used instead.
0 shows the flow of information (data) among the user terminal 11, the information search device 12, and the service server 18. As the user terminal 11, a personal computer (personal computer) in which Internet browsing software or WWW software is incorporated is typically used. In the example shown here, the information search device 12 processes requests from an unspecified number of users. For this purpose, a plurality of user terminals 11 are connected to the network 10 in FIG. Of course, only a single user terminal may be connected to the information search device 12,
Further, the information search device itself may also function as a user terminal.

When an address with a natural sentence is transferred from the user terminal 11, a natural sentence extracting means 13 for extracting a natural sentence from the address with a natural sentence is provided in the information retrieval device 12, and a natural sentence extracting means 13 In the case where the natural sentence extracted in the above or the natural sentence is directly input from the user terminal 11, the natural sentence is analyzed, and the natural sentence analyzing means 14 which determines the type of the request indicated by the natural sentence; An address conversion table 16 in which the address of a service server 18 that provides a service corresponding to the request type is stored in advance, and the address conversion table 16 is searched based on the request type determined by the natural sentence analyzing unit 14 and the corresponding service is searched. Address conversion means 15 for obtaining an address of the server 18;
An address activating unit 17 is connected to the service server 18 of the address based on the address acquired by the address translating unit 15 and transfers information necessary for retrieval to the service server 18. Connecting to and transferring information to a server represented by a given address will be referred to herein as "address activation".

Here, the address with a natural sentence input by the user is, as described later, an address for connecting to the information retrieval apparatus 12 via the network 10, and the "request" (request) of the user. ) Is added. The natural sentence directly input by the user is a natural sentence representing the user's request. Natural sentences generally include a word or a plurality of phrases. The “request” of the user includes an item to be searched and an item to be requested, and in a broader sense, a question, a statement of a will and a request issued by the user in hope of providing some information and enjoying a service. .

Next, the configuration of the natural sentence analyzing means 14 will be described with reference to FIG.

The natural sentence analyzing means 14 includes a word dictionary 42,
With reference to the extraction rule table 43, the word dictionary 42 and the extraction rule table 43, a keyword extraction unit 41 that extracts keywords (product keywords, verb keywords, address keywords, etc.) from the input natural sentence, The request type conversion table 45 storing the request type corresponding to the combination (combination) and the request type conversion means 44 for searching the request type conversion table 45 and acquiring the request type for the keyword extracted by the keyword extraction means 41. Have.

Next, information retrieval using the information retrieval device 12 will be described in detail.

When the user accesses the information search apparatus 12 from his / her user terminal 11 via the network 10 by inputting the address of the information search apparatus 12 in the address input field of the Internet browsing software, the information search is performed. A screen prompting the user to input a “request” is displayed on the display of the user terminal 11 by the device 12. This screen is provided with an elongated rectangular area (so-called input window) for inputting “request”, and the user inputs a request in the area in a natural sentence. For example, "I want to learn English conversation in Shinjuku" is input. This input natural sentence is sent to the natural sentence analyzing means 14.

In the information retrieval apparatus 12, instead of displaying the above-mentioned screen and inputting the request, the request itself can be included in the address for accessing the information retrieval apparatus 12. In this case, the request described in the natural sentence is included in the address and transmitted to the information retrieval device 12 as the address with the natural sentence.

The functions of the user terminal 11 described in this specification are realized by general Internet browsing software and electronic mail software.

Here, the address with a natural sentence will be described.

When the user terminal 11 uses Internet browsing software, as shown in FIGS. 4A and 4B, the address (UR) of the information search device 12 is used.
L) includes a natural sentence. For example, when it is desired to obtain a web page corresponding to a request, as shown in FIG.
L protocol section as "http:" (http = hypertext
transfer protocol), and the host unit is the information retrieval device 1
Use the host name of 2, and put the request for natural sentence in the path part. If the user wants to obtain a file or the like corresponding to the request by file transfer, the protocol part is set to "ftp:" (ftp = file transfer protocol) as shown in FIG.
The host part is used as the host name of the information retrieval device 12, and a request for a natural sentence is entered in the path part.

If the user terminal 11 uses e-mail software and the user wants to send an e-mail to an unspecified store or the like corresponding to the request, a mail address as shown in FIG. Include a natural sentence expressing the request. That is, in the mail address, a request for a natural sentence is entered in the destination user section (left side of @) of the mail address, and the host name of the information search device 12 is entered in the destination host section (right side of @).

The address with the natural sentence from the user terminal 11 is sent to the natural sentence extracting means 13 in the information search device 12.

Hereinafter, the processing (information search method 31) in the information search device when a natural sentence or an address with a natural sentence is received from the user terminal 11 will be described with reference to the flowchart of FIG.

First, in step 30, it is determined whether the sent from the user terminal 11 is a natural sentence or an address with a natural sentence. If the address is a natural sentence, the process proceeds to a natural sentence extraction process 32, and if the address is a natural sentence, the process proceeds to a natural sentence analysis process 33. In the natural sentence extraction process 32,
The natural sentence analysis section 13 cuts out the natural sentence portion in the address, sends the cut out natural sentence to the natural sentence analysis section 14, and then proceeds to the natural sentence analysis step 33.

In the natural sentence analyzing step 33, the given natural sentence is analyzed by the natural sentence analyzing means 14 to derive the type of request. Hereinafter, the details of the processing in the natural sentence analysis step 33 will be described with reference to the flowchart in FIG.

In the natural sentence analyzing step 33, first, a keyword is extracted by executing a keyword extracting step 51, and then, a request type is obtained from the keyword by executing a request type conversion step 52. The keyword extracting step 51 is a step of extracting a keyword from a natural sentence by the keyword extracting means 41, and performs a morphological analysis of the natural sentence with reference to the word dictionary 42 and performs a dependency analysis on the result of the morphological analysis. In the extraction rule table 43, it is specified in advance which keyword at which position in the result of the dependency analysis is to be a keyword, and the keyword extraction means 41 refers to the extraction rule table 43, A keyword is extracted from the dependency analysis result. As the types of keywords to be extracted, product keywords representing product names such as "English conversation", "file compression software", "Italian cuisine", verb keywords representing verbs such as "learn" and "eat", and "Shinjuku" , "Shibuya" and the like. Specifically, the keyword extraction processing described in Japanese Patent Application Laid-Open No. H11-85766 by the present inventors can be applied.

In the request type conversion step 52, the request type conversion means 44 converts the extracted keyword into the request type with reference to the request type conversion table 45. In the request type conversion table 45, a request type corresponding to a combination of a product keyword and a verb keyword is described in advance. FIG. 7 shows an example of the contents of the request type conversion table 45.

In the example shown here, in the request type conversion table 45, the request type “store search” corresponds to the combination of the product keyword “English conversation” and the verb keyword “learn”. In the request type conversion step 52, the request type corresponding to the combination of the product keyword and the verb keyword obtained in the keyword extraction step 51 is obtained by searching the request type conversion table 45. Here, a set of merchandise keywords or a set of verb keywords may be used as a thesaurus, and records may be set in the request type conversion table 45 only for representative combinations of merchandise keywords and verb keywords. If a combination of a certain product keyword and a verb keyword is not found in the request type conversion table 45, the product keyword or the verb keyword is set to a higher-level concept using a thesaurus, and the search is performed again. For details of the processing using this thesaurus,
For example, it is described in Japanese Patent No. 2,669,827.

The request type obtained in this manner is used together with the product keyword and the address keyword extracted in the keyword extraction step 51 together with the address conversion means 1.
Sent to 5.

When the natural sentence analyzing step 33 is completed as described above, in the next address converting step 34, the address converting means 15 converts the acquired keywords and the like into addresses by referring to the address converting table 16. The address conversion table 16 includes http, ftp, e-mail
It is prepared in advance for each communication type such as (e-mail). The internal record of the address conversion table 16 includes a set of a request type, a constraint condition such as an address and a product, and an address corresponding thereto.

FIG. 8 shows an example of the contents of the address conversion table 16 corresponding to the communication type http. For the combination of the request type “restaurant search” and the constraint condition “address keyword: Tokyo”, the URL of “Tokyo restaurant guide page” corresponds as the address. “Tokyo Restaurant Information Page” is one of search engines on the network 10. FIG. 9 shows an example of the contents of the address conversion table 16 corresponding to the communication type ftp. For the combination of the request type “download” and the constraint condition “product keyword: file compression software”, the URL of the “utility software distribution FTP server” corresponds as the address. FIG. 10 shows an example of the contents of the address conversion table 16 corresponding to the communication type e-mail. The mail address of each Italian restaurant corresponds to the combination of the request type “restaurant search” and the constraint condition “product keyword: Italian cuisine, address keyword: Shibuya”.

The address obtained in the address conversion step 34 may be the address of the server of the search engine as described above, or may be the address of the target information itself. In the address conversion step 34, the product keyword and the address keyword sent from the natural sentence analysis step 33 are used as constraint conditions, and the address conversion table 16 is searched by combining the keyword and the request type to obtain the corresponding address. Here, a set of product keywords and a set of address keywords may be made into a thesaurus, and only the representative product keywords and address keywords may be set in the constraint conditions of the address conversion table 16. If there is an inclusive relation in the thesaurus between a product keyword or an address keyword as a search key and a product keyword or an address keyword in the constraint condition of the address conversion table 16, it is determined that they match.

The address obtained in the address conversion step 34 is sent to the address activating means 17.

In an address activation step 35 following the address translation step 34, the address activation means 17 activates the address obtained in the address translation step 34,
Connect to the server at that address. At this time, depending on the type of service of the server, a natural sentence or other information input by the user (such as a title and a text when the communication type is e-mail) is also transmitted to the server. Hereinafter, as a specific example of the processing, a case where the user transmits the addresses illustrated in A) to C) of FIG. 4 will be described.

Process flow in the case of A) in FIG. 4: Assuming that the address with a natural sentence shown in A) in FIG. 4 is transmitted,
In the natural sentence extraction process 32, the natural sentence “I want to learn English conversation in Shinjuku” is extracted. Next, in the keyword extraction process 51 of the natural sentence analysis process 33, the product keyword “English conversation”, the verb keyword “learn”, the address keyword "Shinjuku" is extracted. Then, in the request type conversion process 52, the request type conversion table 4 is set using the combination of the product keyword “English conversation” and the verb keyword “learn” as a search key.
5 (FIG. 7) is searched, and as a result, "shop search" is obtained as the request type.

Next, in the address conversion process 34, the request type: "store search", the constraint condition: "commodity keyword:
The address conversion table 16 (FIG. 8) is searched for "English conversation, address keyword: Shinjuku", and the URL of the yellow page as a search engine is obtained. In the address activating process 35, the URL of the yellow page is activated after including the input natural sentence "I want to learn English conversation in Shinjuku" as transmission information. The Yellow Page server searches based on the natural sentence "I want to learn English conversation in Shinjuku".
The search result is displayed on the user terminal 11.

Process flow in case of FIG. 4B: Assuming that the address with a natural sentence shown in FIG. 4B) is transmitted,
In a natural sentence extraction process 32, a natural sentence “file compression software” is extracted.
The product keyword “file compression software” is extracted.
Next, in the request type conversion process 52, the request type conversion table 45 (FIG. 7) is searched using the product keyword "file compression software" as a search key, and the request type "download" is obtained.

Next, in the address conversion step 34, the address conversion table 16 (FIG. 9) is searched for the request type: "download" and the constraint condition: "commodity keyword: file compression software".
The URL of the TP server is obtained. Address activation process 3
In 5, the input natural sentence “file compression software” is included as transmission information, and the utility software distribution F
Activate the URL of the TP server. The utility software distribution FTP server sends an inquiry about which file compression software to download to the user terminal 11 based on the input natural sentence "file compression software".

Process flow in the case of FIG. 4C) Assuming that the address with a natural sentence shown in FIG. 4C) is transmitted,
In a natural sentence extraction process 32, a natural sentence “I want to eat Italian food in Shibuya” is extracted, and a keyword extraction process 51 is performed.
In, the product keyword “Italian food”, the verb keyword “eat”, and the address keyword “Shibuya” are extracted. Next, in the request type conversion process 52, the request type conversion table 45 (FIG. 7) is searched using the combination of the product keyword “Italian food” and the verb keyword “eat” as a search key, and the request type “restaurant search” is obtained. Can be

Next, in the address conversion process 34, the address conversion table 16 (FIG. 10) is searched as the request type: "restaurant search" and the constraint condition: "product keyword: Italian cuisine, address keyword: Shibuya". A set of email addresses for Italian restaurants in the district is obtained. In the address activation process 35, after including the input natural sentence "I want to eat Italian food in Shibuya" and the information of the title and the text input by the user as transmission information, the mail address of the Italian restaurant is activated, and An email is sent to the Italian restaurant.

The Italian restaurant receiving this e-mail
Input natural sentence "I want to eat Italian food in Shibuya" and title,
Upon receiving the information in the text, it is possible to appropriately perform some action to the user based on the information, for example, send a reply mail.

The information search apparatus described above can also be realized by causing a computer such as a server computer to read a computer program for executing the above-described information search and executing the program. The program for executing the above-described information search is a magnetic tape or a CD.
-Read into a computer by a recording medium such as a ROM. FIG. 11 is a block diagram illustrating a configuration of a computer that functions as an information search device by executing the above-described information search.

This computer has a central processing unit (CPU) 2
1, a hard disk device 22 for storing programs and data, a main memory 23, an input device 24 such as a keyboard and a mouse, and a display device 25 such as a CRT.
It comprises a reading device 26 for reading a recording medium 27 such as a magnetic tape or a CD-ROM, and a communication interface 28 serving as an interface to the network 10. The hard disk device 22, main memory 23, input device 24, display device 25, reading device 26, and communication interface 28 are all connected to the central processing unit 21 via an internal bus or the like.

In this computer, a recording medium 27 storing a program for executing the above-described information search is mounted on a reading device 26, the program is read from the recording medium 27, stored in the hard disk device 22, and stored in the hard disk device 22. When the stored program is executed by the central processing unit 21, it functions as the information search device 12. In this case, the address conversion table 16, the word dictionary 42, the extraction rule table 43, and the request type conversion table 45 are generally constructed as files in the hard disk device 22 as storage means. Therefore, prior to executing the above-described information search, the address conversion table 1 as a file
6. Necessary information must be stored in the word dictionary 42, the extraction rule table 43, and the request type conversion table 45, respectively. Further, the natural sentence extracting means 13, the keyword extracting means 41 and the request type converting means 44, the address converting means 15 and the address activating means 17 are substantially constituted by the central processing unit 21 which executes a program.

As described above, the present invention has been specifically described based on the preferred embodiments. However, the present invention is not limited to the above embodiments, and various modifications may be made within the scope of the claims. And extensions are possible.

[0050]

As described above, according to the present invention, a user inputs a natural sentence expressing a request without knowing the address of a server for searching for desired information or the address itself directly indicating desired information. Alone, it is possible to access such an address.

Further, since it is possible to transmit the address indicating the information retrieval device including the request of the natural sentence, a simple retrieval interface can be realized.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an information search device according to an embodiment of the present invention.

FIG. 2 is a diagram showing a connection form to a network.

FIG. 3 is a block diagram showing a configuration of a natural sentence analyzing unit.

FIGS. 4A to 4C are diagrams illustrating examples of addresses with natural sentences;

FIG. 5 is a flowchart illustrating an example of processing of an information search method according to the present invention.

FIG. 6 is a flowchart showing a process in a natural sentence analysis process.

FIG. 7 is a diagram illustrating an example of a configuration of a request type conversion table.

FIG. 8 is a diagram showing an example of the contents of an address conversion table corresponding to “communication type: http”.

FIG. 9 is a diagram illustrating an example of the contents of an address conversion table corresponding to “communication type: ftp”;

FIG. 10 is a diagram illustrating an example of the contents of an address conversion table corresponding to “communication type: e-mail”;

FIG. 11 is a block diagram illustrating an example of a configuration of a computer system for configuring the information search device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Network 11 User terminal 12 Information search device 13 Natural sentence extraction means 14 Natural sentence analysis means 15 Address conversion means 16 Address conversion table 17 Address activation means 18 Service server 31 Information search method 32 Natural sentence extraction process 33 Natural sentence analysis process 34 Address Conversion process 35 Address activation process 41 Keyword extraction means 42 Word dictionary 43 Extraction rule table 44 Request type conversion means 45 Request type conversion table 51 Keyword extraction process 52 Request type conversion process

Claims (9)

[Claims] 1. An information retrieval apparatus for converting a request described as a natural sentence into an address of a server providing a service corresponding to the request, comprising a request type table holding a request type corresponding to each keyword. First recording means; natural sentence analyzing means for analyzing the natural sentence and referring to the request type table to determine a request type of a request included in the natural sentence; and holding an address corresponding to each request type. A second recording unit having an address conversion table, and an address conversion unit for searching the address conversion table based on the request type obtained by the natural sentence analysis unit, and converting the address conversion table into an address of a server performing a service of the request type. When information including the natural sentence is required in a server performing the service of the request type, the information is included as transmission information. After having allowed, information retrieval device having an address start means for starting the address. 2. A natural sentence analyzing means, comprising: keyword extracting means for extracting a keyword corresponding to a request from a meaning and a dependency relationship of a word included in a natural sentence; and referring to a request type table, extracting the extracted keyword. The information search device according to claim 1, further comprising: request type conversion means for converting into a request type. 3. An apparatus according to claim 1, further comprising: a natural sentence extracting unit that receives an address including a request described as a natural sentence and analyzes the input address to extract the natural sentence. 3. The information retrieval apparatus according to claim 1, wherein a natural sentence is input to the natural sentence analyzing means. 4. A first recording unit having a request type table for holding a request type corresponding to each keyword, and a second recording unit having an address conversion table for holding an address corresponding to each request type. An information search method for converting a request described as a natural sentence into an address of a server that provides a service corresponding to the request, wherein the natural sentence is analyzed and included in the natural sentence with reference to the request type table A natural sentence analysis step of determining a request type of the request to be received; and an address for searching the address conversion table based on the request type obtained in the natural sentence analysis step, and converting the address conversion table into an address of a server that performs a service of the request type. A conversion process, and when information including the natural sentence is required in a server that performs the service of the request type, the information is used as transmission information. And an address activation step of activating the address after including the information. 5. The natural sentence analyzing step includes: a keyword extracting step of extracting a keyword corresponding to a request from a meaning and a dependency relationship of a word included in the natural sentence; and the keyword extracted by referring to a request type table. 5. The information search method according to claim 4, comprising: converting a request type into a request type. 6. A natural sentence extracting step of inputting an address including a request described as a natural sentence and analyzing the input address to extract the natural sentence, wherein the natural sentence extracting step extracts the natural sentence. 6. The information retrieval method according to claim 4, wherein a natural sentence is used as an input in a natural sentence analysis process. 7. A computer-readable recording medium, which is an information search process for converting a request described as a natural sentence into an address of a server that provides a service corresponding to the request. A natural sentence analysis process for analyzing the natural sentence and determining a request type of a request included in the natural sentence; and a server for providing a service of the request type based on the request type obtained in the natural sentence analysis process. An address translation process for translating to an address, and, if the information including the natural sentence is required in a server that performs the service of the request type, including the information as transmission information, and then activating the address A recording medium storing a program for executing a process having an activation process. 8. A natural sentence analyzing process includes a keyword extracting process for extracting a keyword corresponding to a request from a meaning and a dependency relationship of a word included in a natural sentence, and a request type conversion for converting the extracted keyword into a request type. The recording medium according to claim 7, comprising a process. 9. The information retrieval process further includes a natural sentence extraction process of inputting an address including a request described as a natural sentence and analyzing the input address to extract the natural sentence. 9. The recording medium according to claim 7, wherein the natural sentence extracted by the sentence extraction process is used as an input in the natural sentence analysis process.