JP3186960B2 - Information retrieval method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention searches for unit information matching a condition designated by a user from a large number of unit information distributed and stored in each computer in a distributed network environment, and provides a result to a user. The present invention relates to an information retrieval method having an agent function for outputting and an apparatus therefor.

[0002]

2. Description of the Related Art In an agent of a conventional information retrieval apparatus, it is necessary to explicitly input a retrieval key such as a keyword in order to designate a retrieval condition to the agent. Also, the position information of the computer to be searched was explicitly input.

[0003]

In the configuration of such a conventional information retrieval apparatus, a keyboard or other character string input means for inputting a keyword is required, and a television receiver or the like generally having no character string input is required. It was difficult to apply to Further, it is not clear what keyword the unit information to be searched for corresponds to, and it has been difficult to apply it to information of a type such as television programs and travel information. Further, when information is stored in a distributed network environment in a distributed manner, it is difficult to specify the position information of the computer to be searched one by one. An object of the present invention is to solve these problems.

[0004]

According to the present invention, in order to solve the above-mentioned problems, a user specifies search conditions to an agent by inputting, as an example, unit information whose contents are similar to the unit information to be searched. It is like that.
In addition, distributed networks often included in unit information
A location information extracting means for extracting location information of related information in a work environment . This makes it possible to apply the present invention to a device having no keyword input means, and it is not necessary to explicitly specify position information of a computer to be searched. Further, the present invention can be applied to information of a type in which the corresponding keyword is not clear.

[0005] Further, a function of feeding back the selection result of the user, preferentially searching for a search position where preferable unit information is obtained, and displaying the search result obtained therefrom with priority is provided. This can reduce the problem of displaying search results derived from a search position that does not match the user's request due to not explicitly specifying the position information of the computer to be searched, and by reusing the agent repeatedly. It is possible to efficiently determine the position of the computer to be searched by adapting to the same kind of search request.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS According to the first aspect of the present invention, a user inputs unit information whose content is similar to a unit information to be searched by a user as an example, so that an agent knows a user's search request, Other related unit information that is often included in the input unit information is stored.
Label that indicates the name of the computer being
The location information consisting of the access destination and the input unit information
From the location specified by the extracted location information.
Access each computer to obtain unit information, and
Based on the similarity with the input unit information.
This is an information retrieval method for displaying to a user as a retrieval result based on the information, thereby realizing an agent that executes a retrieval in a distributed network environment without explicitly inputting a keyword or positional information of a computer to be retrieved. Can be.

According to a second aspect of the present invention, in addition to the function of the first aspect, importance is given to each of the extracted position information , and the importance is included in the position information. <br/> Retained as the knowledge of the agent, the user selects unit information that meets his / her search request from the search results displayed to the user, and the position information from which the selected unit information is derived is regarded as important. An information search method for adjusting the degree of importance held as knowledge of the agent, and when a user's search request is repeatedly input, a search result that matches the request gradually adapts to the user's preference and increases the search result. It can be displayed by priority.

The invention described in claim 3 of the present invention includes unit information input means for inputting unit information similar in content to the unit information desired by the user as an example, and the input unit information often includes the unit information. In a distributed network environment
Indicates the name of the computer where the related unit information is stored
Location information consisting of the label and the access destination to the computer
Position information extracting means for extracting from the input unit information, and each computer designated by the extracted position information
Searching means for accessing the unit to obtain unit information;
Similarity between the obtained unit information and the input unit information
Search result table displayed to the user as search results based on the
This is an information retrieval device provided with an indicating means , and has the same operation and effect as the first aspect of the present invention.

According to a fourth aspect of the present invention, in addition to the constituent elements of the third aspect , each of the extracted position information is provided.
A position information management means for assigning importance, including the importance in the position information and holding and managing the knowledge as agent knowledge; and unit information matching the user's request from the displayed search results. a unit selection input means for a user to select a, and severity adjustment means for adjusting the degree of importance which is held in the position information management means based on the selection result, the
Decide which computer gets unit information based on importance
Means for determining a search position, wherein the search means
Acquires unit information only from the computer determined by the location determination means
The same operation and effect as the invention described in claim 2 are obtained.

[0010] The invention described in claim 5 of the present invention, claim
4 and the unit selection input means further includes a related information often included in the unit information selected by the user .
The name of the computer where other unit information is stored.
Location information consisting of the bell and the access destination to the computer.
Out, an information retrieval device for holding the position information management means as knowledge of the agent, in addition to the effects of the invention described in claim 4, the unit information selected by the user in a unit selection input means, the position information extraction means By analyzing and extracting new location information from the location information and updating the storage management of the location information managed by the location information management means, the accumulation and refinement of the location information, which is the knowledge of the agent, can be performed. It has the effect of being able to realize an agent that performs a search that meets a request earlier, adapting earlier to the user's preference.

[0011] The invention described in claim 6 of the present invention, claim
4 and 5 comprising the same components as those described above,
Means further for each of said extracted location information
Service form or meaning of information at each access destination
And the importance of each sub-label
By the unit selection input means
Unit information or sub-label information selected by the user
An information retrieval apparatus that adjusts the degree of importance added to the pair of the label and the sub-label based on a user's search request. It can be displayed by priority. That is, the information search device in the present embodiment has the effect of the invention described in claim 4 or 5 ,
When a search unit searches for a computer based on certain location information, and the search result is divided into categories, the location information is retained in a detailed form with a sub-label attached, and the unit selected by the unit input means The position information corresponding to the information is weighted in a refined form by the importance adjustment means, and is managed in a refined form by the position information management means. This has the effect of being able to implement an agent that adapts more quickly to the user's preference and performs a search that meets the requirements.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 shows a configuration of an information search device for explaining an embodiment of the present invention. FIG.
1 is an information search device, 2 is a unit information input means for inputting unit information similar in content to unit information desired by a user as an example, and 3 is a distributed network often included in the input unit information. The location information extracting means 4 for extracting the location information of the related information under the environment accesses each computer specified by the location information based on the similarity with the unit information input by the user as an example in the distributed network environment. Search means 5 for obtaining a search result is a search result display means for displaying the search result to the user. 6 is a position information management means for storing and managing the extracted position information as the knowledge of the agent, 7 is a unit selection input means for the user to select unit information matching the user's request from the displayed search results, and 8 is a unit selection input means. Importance adjusting means for adjusting the importance added to each of the position information stored in the position information managing means 6 based on the selection result, and a distributed network 9 for performing a search based on the position information managing means 6. Search position determining means for determining the position of the computer under the environment.

Next, the operation of the above embodiment will be described. In the information search device 1, the unit information input means 2 receives input of unit information specified by a user,
The position information extracting means 3 extracts position information in a distributed network environment in which related information is stored from the unit information, and the search means 4 accesses a computer designated by the position information to obtain a search result. , The search result display means 5
The search result is displayed to the user.

In the information search device 1, the unit selection input means 7 accepts a user's selection of unit information matching the user's search request from the search results, and the importance adjusting means 8 controls the unit information. Is weighted according to the frequency of selection, and the position information management means 6
Manages and accumulates the weights and the position information, and when the search request is made again, the search position determining means 9 determines the position of the search destination in a priority order according to the weights and searches. As a result, position information serving as knowledge of the agent is accumulated, and the knowledge is refined in a form in which the position information from which the unit information selected by the user is regarded as important, and gradually adapted to the user's preference. The search results that match the request will be displayed with high priority.

(Embodiment 1) Next, claim 1 of the present invention will be described.
A first embodiment corresponding to FIGS.

In the present embodiment, as an example, unit information relating to travel is distributed and stored in a distributed network environment. By inputting unit information relating to a program as an example, the unit information according to the user's preference can be selected from the information. An operation in the case of searching for unit information that has been performed will be described. In addition to the unit information, events, companies, people, place names, etc.
Things related to various information can be considered. There are various storage formats of unit information, such as a WWW homepage on the Internet, database items, and files. Further, from the aspect of the contents, there are a type having a characteristic as supplementary information of other information such as a program, such as a program introduction, and a type having a single unit of information, such as a travel catalog. For the search using the information search device 1, an agent that receives a search request from a user, performs overall processing related to the search, and outputs a search result is used.

FIG. 2 shows an example of a display screen 10 for the user to access the agent.
Titles 17 to 19 of one unit information are displayed. As described above, the agent is a function for searching for and presenting information such as a program according to the user's preference. Although it is easy to understand that the agent is implemented by using a personified user interface as shown in FIG. 2, this is not a necessary condition for realizing the agent function.

FIG. 3 shows an example of unit information relating to a program to be input in the present embodiment. The unit information related to a program includes a program title, a character string describing a synopsis of the program, and information for viewing the program. The unit information often includes position information of the related information. In the case of unit information related to a program, the position information of the related information is directly expressed as a network address as in a), and a label that co-occurs with a specific format such as “provide:” as in b) (b). It is stored in various formats, for example, when it is obtained by retrieval from "xx tourist"), when a label is implicitly included in a sentence in natural language as in c).

FIG. 4 shows an example of the display screen 10 in which the user inputs a search request to the agent based on unit information similar to the unit information to be searched. Here, the unit information “travel to New Zealand” is selected with the mouse from the unit information related to the program displayed by the “program” agent 11, and dragged into the “search request” area of another agent as an example of the unit information to be searched. You are doing input by doing.

FIG. 5 shows a mechanism by which the position information extracting means 3 extracts the position information of the computer to be searched from the unit information. The mechanism for extracting the position information differs depending on the format in which the position information is stored in the unit information. When the network address is explicitly included in the unit information 21 as shown in FIG. 3A), as shown in FIG.
Retrieves it directly and obtains a label / network address pair as location information 32. As shown in FIG. 3B), when a label co-occurring with a specific format is explicitly included in the unit information 22, as shown in FIG.
4, the label “xx tourist” 35 is extracted. Using this label as a keyword, the name server inquiry unit 36 in the location information extraction unit 3 inquires the name server 37 in a distributed network environment to obtain a network address. Thus, the position information 32
As a label and network address pair. Also, as shown in FIG. 3C), when a label is implicitly included in a sentence in a natural language, the input sentence is parsed using a natural language parsing technique to extract a label “xx tourist” 35. Then, as in the case of FIG. 5B), the name server is queried to search for the network address, and the position information 32 is obtained.

The core words of template matching and syntax analysis include "provide" for program information, but "host", which can co-occur with company names and personal names,
Many words are retained, such as "support", "maker", and "artist".

FIG. 6 is a conceptual diagram showing a distributed network environment in the present embodiment. On the network, information retrieval device 1 of the present invention used by the user, computers 41, 42, 43 storing unit information relating to travel,
A computer 44 storing unit information relating to programs and a name server 37 for searching a network address of a corresponding computer from a label of position information (for example, “xx tourist”) are connected. FIG. 6 shows a minimum configuration for describing the present embodiment, and is greatly different from an actual distributed network, for example, the Internet in terms of scale and complexity. Note that the name server 37 itself is not a component of the present invention, but a server mounted under a distributed network as a service for searching for a network address from a label. For example, in the case of the Internet, "Yahoo" and "Lycos" are known.

The search means 4 accesses the computer 41 indicated by the position information 32 extracted by the position information extraction means 3 based on the above method, and performs a search based on the similarity with the text which is the content of the unit information. In general, a plurality of unit information is obtained as a search result. In the present embodiment, based on unit information on a program input by a user as an example, unit information on a trip is searched. The search here is performed based on the similarity between texts (here, “synopsis”) included in the two unit information. As a method of determining the similarity between texts, a set of keywords extracted from each text is treated as a vector,
A method of calculating the similarity and a method of calculating the similarity of the meaning by syntactic analysis and semantic analysis of each text are known, and any method may be used here.

In the search method, a program having an agent function moves to a computer that actually performs a search, is executed there, and returns a search result. A plurality of agents handle computers at a large number of search positions. Search
A method of merging search results, a computer at the search position also has an agent that specializes in searching, a method of passing search conditions and search results by communication between agents, a method of locally executing a program in the computer used by the user There are various methods, such as a method of executing and thereby simulating the function of the agent. Regardless of the method, the functions of the present invention can be applied effectively.

The agent displays a plurality of unit information obtained as a result of the search to the user through the search result display means 5. FIG. 7 shows an example of the display screen 10 on which the agent displays the search result to the user. Here, three unit information titles 51 to 53 relating to travel are displayed. As a result, an agent that accesses a related computer and searches for unit information having contents similar to the input unit information as an example has been realized. The user can select and use the most suitable unit information from the unit information presented by the agent.

(Embodiment 2) Next, claim 2 of the present invention
A second embodiment corresponding to claim 4 will be described.

The user selects the unit information which seems to match his or her request from the displayed plurality of unit information by the unit selection input means 7. In FIG. 7, the user has selected the unit information 52 having the label "shot over jet tour".

The result of the selection made here is used to adjust the importance of the position information stored in the position information management means 6. FIG. 8 shows the format of the position information.
The location information corresponding to the location of each computer in the distributed network environment includes three fields of importance, label, and address. The importance is an index indicating the degree to which the unit information stored at the position is important from the viewpoint of the preference of the user managed by this agent. The higher the value, the higher the possibility that the unit information obtained from the position is selected by the user. In other words, it can be said that there is more information that better suits the user's preference. Conversely, if this value is small, it is unlikely that the user will select the unit information obtained from that position, indicating that there is less information that matches the user's preference.

The storage of new position information in the position information management means 6 is performed based on unit information input by the user as an example. FIG. 8A shows the contents of the position information management means 6 before the search is performed for the first time and the unit selection is input. By the above-mentioned position information extracting means 3, a pair of a label and an address in the unit information inputted as an example is obtained. By giving it an initial value of importance, for example, 1.0, a new item of location information is created.

The position information of the position storing the unit information selected by the user is always stored in the position information management means 6. When the user selects unit information of a certain position from the viewpoint of whether or not the user's request is met, the importance adjusting means 8 increases the importance of the corresponding position information. Conversely, the importance of the position information corresponding to the unselected unit information is reduced.

When a similar search request for the same agent, for example, a request for searching for unit information relating to travel is repeated, as shown in FIG. Information is stored. The search position determining means 9 determines the position of the computer to be searched on the basis of the importance. For example, a method of not searching for a position corresponding to position information whose importance is smaller than a certain value can be applied. In this way, the agent adapts to the same type of search request and efficiently controls the position of the computer to be searched. In addition, the search result display means 5 displays the unit information searched from the position having a high importance with a high priority. For example, a method of displaying the position near the head of the list or numerically expressing the priority and displaying it explicitly can be applied. With these two functions, it is possible to realize an agent that gradually adapts to the user's preference and displays useful information with high priority.

(Embodiment 3) Next, claim 5 of the present invention will be described.
A third embodiment corresponding to FIG. In the functions of claims 1 to 4, the position information stored in the position information management means 6 is only the one derived from the unit information input by the user as an example. According to the fifth aspect, in addition, the unit information selected by the user in the unit selection input means 7 is also treated as an input of the position information extraction means 3, and the position information contained therein is already stored in the position information management means 6. If not, it is newly stored and managed. In the present embodiment, for example, position information from which information on a local tour company can be searched can be extracted from the text in the unit information of “shot over jet tour”. As a result, it is possible to collect more location information with a smaller number of search requests than when only the functions of claims 1 to 4 are implemented,
A wide range of unit information can be searched.

(Embodiment 4) Next, claim 6 of the present invention will be described.
A fourth embodiment corresponding to FIG. In the functions of the first to fifth aspects, the position information stored in the position information management means 6 is only the position of the computer to be searched and the position of the directory in the file system of the computer. However, in a search on the Internet or the like, information at each search position has an internal structure, and various types of information may be available on the same computer. For example, different service forms such as a WWW homepage, a database search service, an FTP service, and download using a broadcasting network, and information having different meanings such as tour information and guidebook information can be discriminated by a label on the same search position. May be mixed.

When the function of claim 6 is implemented, a sub-label is provided as a new field in the position information stored in the position information management means 6, and the service form and the meaning of the information at the same search position are added. Be able to express. The sublabel is acquired from the search position when the search unit 4 performs a search, and is stored in the position information management unit 6 when the unit is selected by the user.

FIG. 9 shows the contents of the location information management means 6 after the same search request has been repeated several times for the same agent. Even if the search location is the same as the "xx tourist", if the tour information is selected and used more frequently than the guidebook information in the search location, the importance of the location information corresponding to the tour information as shown in FIG. The degree is high, and this is searched with priority. As described above, even at the search position of the same label, each importance can be assigned according to the service form and the meaning of the information. Adaptation can be realized.

[0036]

As described above, according to the present invention, a search is performed based on the similarity with the unit information inputted by the user as an example, so that the present invention can be applied to a television receiver having no keyword input means. Therefore, the present invention can be applied to a type of information such as a television program or travel information in which it is not clear what keyword the unit information to be searched for corresponds to.

According to the present invention, the position information of the related information is automatically extracted and used from the unit information input by the user as an example or the unit information selected from the search result. It is not necessary to explicitly specify the position information each time, and the user can use the information without being aware that the unit information is stored in a distributed manner.

Further, according to the present invention, information that can be used in various forms such as a WWW homepage, a database search service, an FTP service, and download using a broadcast network in a distributed network environment is transmitted in response to a user's search request. A unified mechanism for use is feasible. With the unified mechanism, cross-reference such as searching for information in another form using information in a certain form as a search key becomes easy.

Further, according to the present invention, an agent having a learning ability adapted to the same kind of search request and capable of efficiently determining the position of the computer to be searched can be realized. For example, they can be used efficiently according to various personal preferences such as a preference for a television program and a preference for traveling abroad. This function can be provided to the user in an intuitive and easy-to-understand manner by using a user interface in which an agent is personified.

[Brief description of the drawings]

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

FIG. 2 is a display screen view for a user to access an agent according to the first embodiment of the present invention.

FIG. 3 is a schematic diagram showing an example of unit information input by a user as an example or displayed as a search result according to the first embodiment;

FIG. 4 is a display screen diagram in which a user inputs a search request to an agent by inputting unit information similar to the unit information to be searched in the first embodiment.

FIG. 5 is a schematic diagram illustrating a position information extraction method in a position information extraction unit according to the first embodiment.

FIG. 6 is a schematic diagram illustrating an overview of a distributed network environment according to the first embodiment.

FIG. 7 is a view showing a display screen on which a search result of an agent according to the second embodiment is displayed and selected by a user.

FIG. 8 is a schematic diagram showing a storage format of position information in a position information management unit according to the second embodiment.

FIG. 9 is a schematic diagram illustrating a storage format of position information having a sublabel indicating an internal structure in a position information management unit according to the fourth embodiment;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Information search apparatus 2 Unit information input means 3 Position information extraction means 4 Search means 5 Search result display means 6 Position information management means 7 Unit selection input means 8 Importance adjustment means 9 Search position determination means 10 Display screens 11-16 Personification Example of Agent Display 17-17 Example of Title of Unit Information on Program 21-23 Example of Unit Information on Program 31 Address Extraction Unit 32 Location Information 33 Template Matching Unit 34 Template 35 Label 36 Name Server Inquiry Unit 37 Name Server 41 -44 Calculator of each search position 51-53 Example of title of unit information related to travel

Continuation of the front page (72) Inventor Mitsuhiro Sato 1006 Kazuma Kadoma, Kadoma City, Osaka Prefecture Matsushita Electric Industrial Co., Ltd. (JP, A) Japanese Patent Laid-Open No. 4-314171 (JP, A) Mitsunori Wada et al., "Information Filtering on the Internet (1)-Overview", Information Processing Society of Japan 51st (late 1995) National Assembly Lecture Papers (4), 1995 (Heisei 07-09-20), pp. 85-86 Mitsunori Wada et al., "Information Filtering on the Internet (2)-How to Organize Information", Proc. Of the 51st (the latter half of 1995) Information Processing Society of Japan (4), 1995. Hei 07-09-20), pp. 139-181. 87-88 (58) Field surveyed (Int. Cl. ⁷ , DB name) G06F 17/30 JICST file (JOIS)

Claims (6)

(57) [Claims] 1. An agent knows a user's search request by inputting unit information whose content is similar to the unit information that the user wants to search as an example, and the related unit information is often included in the input unit information. Other units to
A label indicating the name of the computer where the information is stored and the
The location information consisting of the access destination to the
From the extracted unit information, and
Access the specified computers to get unit information,
The acquired unit information is combined with the input unit information.
An information search method for displaying to a user as a search result based on similarity . 2. An importance is assigned to each of the extracted position information , and the importance is included in the position information.
The agent selects the unit information that matches the user's search request from the search results displayed by the agent to the user, and places importance on the position information from which the selected unit information is derived. 2. The information retrieval method according to claim 1, wherein the importance held as knowledge of the agent is adjusted. 3. A unit information input means for inputting unit information similar in content to a unit information to be searched by a user as an example, and related unit information in a distributed network environment often included in the input unit information. Stored
Label showing the name of the computer and access to the computer
From the input unit information.
Output position information extracting means, and the extracted position information
Get unit information by accessing each specified computer from
Searching means for searching, and
As a search result based on the similarity to
An information search device comprising: a search result display means for displaying a search result on a search result . 4. Each of the extracted location information has a weight.
A position information management means for assigning a degree of importance, including the degree of importance in the position information, and holding and managing the knowledge as agent knowledge; and unit information matching the user's request from the displayed search results. Unit selection input means for the user to select
And severity adjustment means for adjusting the degree of importance which is held in the position information management means based on the selection result, the important
Search to determine the computer to get unit information based on degree
Positioning means, wherein the searching means determines the search position.
The information search apparatus according to claim 3, wherein the unit information is obtained only from the computer determined by the determining means . 5. The unit selection input means further includes a related other unit which is often included in the unit information selected by the user.
A label indicating the name of the computer in which
Extracting position information consisting of an access destination to the computer, it is held in the position information management means as knowledge of the agent 請
The information search device according to claim 4 . 6. The method according to claim 1, wherein said location information management means further comprises:
Information at each access destination for each
Sub-label expressing service form or meaning of information
And the importance of each sub-label are included in the position information.
Selected by the user through the unit selection input means.
The information search device according to claim 4 , wherein the importance added to the pair of the label and the sublabel is adjusted based on unit information or information on a sublabel.