JP2003122754A - Retrieval supporting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a retrieval supporting device for supporting the implementation of retrieval proper to a user's purpose by supporting jobs such as selection of a search engine and an input of retrieval conditions to a user. SOLUTION: The retrieval supporting device is characterized by providing an information request analyzing means 1 for generating information request data including a request type expressing kinds of information requests regarding a purpose of the retrieval by analyzing input by the user, a kind of information to be retrieved and information to be a retrieval key and a retrieval instruction creating means 2 for selecting a proper information search engine based on the information request data and for automatically generating retrieval conditions to be provided to the selected information retrieval engine.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a search support device, and in particular, when a search request is input in natural language, what kind of search engine is selected and what kind of search is performed in order to obtain a necessary search result. The present invention relates to a search support device capable of instructing a user whether to input a condition.

[0002]

2. Description of the Related Art As the Internet has generally spread, there are more and more opportunities to search by using the Internet or the like to obtain necessary information. However, under the present circumstances, it is difficult to easily obtain this information without the search know-how for obtaining the information required by the user.

As a technique for supporting the search, there are means for preventing omission of search and means for expanding the search target. Specifically, (1) a technology to prevent omission by searching by expanding keyword phrases into synonyms (for example, including "Sakura" up to "Sakura" or "Galaxy" up to "Milky Way") (2) Technology that analyzes search conditions, requests search to multiple search systems, and collects the results to search within a range that cannot be covered by one search system.
(3) A technology in which a translation system is inserted between search systems to search information in other languages as well (for example, by translating a Japanese input in English, not only Japanese documents but also documents from the United States, the United Kingdom, etc. Search until) etc.

[0004]

The subject of these techniques is to assist the search system in search, not to support users who use the search system.

Therefore, according to the present invention, in order to support the user's search, a search support apparatus accumulating search know-how is provided, whereby a search engine selection process, a search condition input process, and a search result integration process are performed. It is intended to be semi-automated so that users can easily obtain the information they need.

[0006]

FIG. 1 shows a principle diagram of a search support device according to the present invention. In FIG. 1, 1 is an information request analysis means, and 2 is a search command creation means.

The above object of the present invention can be achieved by the following (1) to (5).

(1) Information request analysis that interprets user input to generate information request data including a request type indicating the type of information request related to the purpose of search, the type of information to be searched, and information serving as a search key Means 1 and search command creation means 2 for selecting an appropriate information search engine based on the information request data and automatically generating search conditions to be given to the selected information search engine. Search support device.

(2) The search support device according to (1), wherein the search command creating means 2 selects an appropriate information search engine according to the type of information to be searched.

(3) The search command creating means 2 adds a character string condition for narrowing down the search result to the information specified as the search key in the information request data, as a search expression given to the full-text search engine. The search support device according to (1) above, which generates a search condition expression.

(4) The search support device according to (1), wherein the information request analysis means 1 analyzes the user's input with reference to the state of the user at the time of input.

(5) The search support device according to (4), wherein the information request analysis means 1 analyzes the input of the user based on profile data in which the characteristics of the user are recorded.

As a result, the following operational effects can be obtained.

(1) Information request analyzing means for generating a request type and information request data, and search command creating means for selecting an appropriate information engine based on the information request data and generating a search condition to be given to the information search engine. With this feature, users who do not know the search know-how can receive assistance for the search procedure, and such users can easily obtain necessary information while using multiple search systems. You can ask.

(2) Since an information search engine corresponding to the type of information to be searched can be provided, the user can select an appropriate one from a large number of information search engines.

(3) As a search expression given to the full-text search engine, a character string condition for further narrowing down the search results is added to the information specified as the search key,
You can get more appropriate answers.

(4) Since the user's input is interpreted by referring to the user's situation at the time of input, an appropriate answer can be obtained from the user's standpoint.

(5) Since the user's input is interpreted based on the profile data recording the user's characteristics,
The profile data can be easily updated according to the user's condition, and an accurate answer from the user's standpoint can be easily obtained.

[0019]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described with reference to FIGS. 2 is an embodiment of the present invention, FIG.
Shows a detailed view of the answer template table, and FIG. 4 shows a detailed view of the search template table. In the figure, 1 is an information request analyzing unit, 2 is a search command creating unit, 11 is an answer template table, 12 is a search template table, and 13 is a user profile data holding unit.

The information request analysis means 1 analyzes a user's input, for example, an input requested to be searched in natural language, and a request type corresponding to the user's intention, a type of information to be searched, and information serving as a search key, that is, search The information request data including the condition is generated.

Based on the information request data, the search command creating means 2 selects an appropriate information search engine from one or more information search engines according to the request type and the type of information to be searched. Further, the search condition expression to be given to the selected information search engine is generated by using the search key in the information request data.

The answer template table 11 holds the type of information to be searched according to the request type and the search condition. As shown in FIG. 3, when the request type is "how to go", the answer Type, that is, the type of information to be searched is "route from the nearest station", "route to the nearest station", "nearest station", and in order to search the route from the nearest station, the search target is the destination “Facility name (for example, the British embassy if the destination is the Igis embassy) and α (nearest station name)” is required. In this case, the nearest station name is unknown,
This will be obtained by the method described later.

The search template table 12 holds a search system or search engine name on the Internet, which is actually used based on the request type and the type of information to be searched, and its search condition.
As shown in FIG. 4, for example, when the request type is “direction” and the type of information to be searched is “nearest station”, “full-text search engine” is used as the search system and the search condition is “facility name ( For example, when the destination is the British Embassy, it is understood that it is "English Embassy) +" walk or bus ".

The search engine around the station specified in the search template table 12 can search the map around the station and the route from other stations to reach that station. Descriptions can be searched, and newspaper article search engines can search information based on newspaper articles.

The user profile data holding unit 13
It holds user profile data, that is, information peculiar to the user. For example, the address of the user is registered.

The operation of the present invention will be described with reference to FIG. 5 when the user makes a search request in a natural language "I want to know how to get to the British Embassy".

First, the user profile data holding unit 13
As the address of the user, for example, the address of Tokyo is entered.

S0. The search request "I want to know how to get to the British Embassy" is input to the information request analysis means 1 shown in FIG.

S1. The information request analysis unit 1 performs a semantic analysis on this by natural language analysis, and recognizes that the request type is “how to” and the search condition is “Embassy of England”.

S2. The information request analysis means 1 searches the answer template table 11 based on this request type and search conditions.

S3. As a result, the request type is "Direction"
In case of, there are 3 types of information to be searched: "route from the nearest station", "route to the nearest station", "nearest station".
However, since the user's address entered in the user profile data holding unit 13 is Tokyo, it is determined that the "route to the nearest station" is unnecessary, and the "nearest station" and "nearest station" are determined. The search result for "route from" is output as information request data.

S3-1. As the information request data (No. 1), the request type is “direction”, the type of information to be searched, that is, the type of answer is “nearest station”, and the search condition is “Embassy of England (facility name) + α” (α Is a search template table). Note that FIG.
The nearest station _- 1 shown in S3-1 is that it is the first station.

S3-2. As the information request data (No. 2), the request type is “how to get”, the type of information to be searched is “route from the nearest station”, and the search conditions are “nearest station” and “Embassy of England (facility name) ) ”Is output. Note that $ in S3-2 of FIG. 5 indicates that the search condition, that is, the nearest station is not obtained.

Then, these S3-1 and S3-2 are output to the search command creating means 2.

S4. Since the search command creating means 2 is controlled to search the route from the nearest station by default with reference to the user profile data 13, in this case the default is to search the output of S3-2. .

S5. The search command creating means 2 obtains the search condition of “nearest station” from the search template table 12. Thus, it is recognized that the "full-text search engine" on the Internet is used and the search is performed by the search condition (facility name + "walk or bus").

S6. Based on this, the following search is provided to obtain the required search results.

S6-1. First, using the "full-text search engine" as the search system, recognizing that the search condition is (Embassy of England + "walk or bus"),
This is output to the user to support the search.

S6-2. The user sees this output and
Using the "full-text search engine", search by the search condition (Embassy of Eggis + "walk or bus"), check the "nearest station" from the results, and get the nearest station (Hanzomon station).

S6-3. Next, the search command creating means 2 uses the station surroundings search engine as a search system to obtain the route from the nearest station from the search template table 12, and searches by (facility name + nearest station) as the search condition. Is recognized and output to the user to assist the search. The user looks at this output and uses the "station area search engine" to search by the search conditions (Embassy of England + Hanzomon Station).

S6-4. As a result, the "station periphery search engine" outputs the route to the British embassy around Hanzomon Station to the user, so that the user can know the "route from the nearest station".

The second operation of the present invention will be described with reference to FIG. The operation explanation of FIG. 5 is for the case where the user has an address in Tokyo where the facility exists, but FIG. 6 shows the case where the user's address is far away from the place where the facility exists, such as a resident of Kyushu. explain. For such a user, it is necessary to show where the target facility is, first, its location, and then show the route from the user's address to the nearest station of the facility.

Therefore, if the user has an address at a remote location defined in advance, it is necessary to fill in the user profile data holding unit 13 in advance that the location of the facility and the route to the nearest station are also obtained.

S10. When the user inputs a search request "I want to know how to get to the British Embassy",
The information request analysis means 1 performs the same operation as in FIG.

S11. The search command creating means 2 receives this output result.

S12. The search command creating means 2 refers to the user profile data holding unit 13 and states that as the user information “how to go” and if the place is a remote place, “route to the nearest station” and “place” are also requested. , Recognize this. Then, based on this, the search condition and the search system according to the type of information to be searched are recognized from the search template table 12.

S13. Then, according to this, the following search results are obtained: S13-1. It outputs to the user a search by (facility name + "address or zip code") with the "full text search engine" to support the search, and thus the user uses the "full text search engine" to search conditions. Search by (British Embassy + "Address or Zip Code").

S13-2. This allows you to look up the "address" and get the address of the British Embassy.

S13-3. To the user, the "full-text search engine" is used to output a search by (facility name + "walk or bus") to support the search, and thus the user uses the "full-text search engine" to search conditions ( Search at the British Embassy + "walk or bus").

S13-4. According to this result, "nearest station"
You can find out that this is Hanzomon Station.

If the user's location is far from the facility, it is often impossible for the user to know only the nearest station. Therefore, the route to the nearest station is calculated as follows.

S14. The "station periphery search engine" is output to the user to search for the nearest station of the user and the nearest station of the facility obtained in S13-4 to assist the search of the route to the nearest station.

S15. This allows the user to know the route from his nearest station to the nearest station of the British Embassy.

S16. Next, the "station surroundings search engine" is used for the user to output a search based on the search condition (facility name + nearest station) to support the search.

S17. This allows the user to use the "station surroundings search engine" to perform a search using the search conditions (Embassy of England + Hanzomon Station) and obtain "directions from the nearest station."

As described above, there are cases where it is desired to change the search procedure depending on the search result, but this can also be controlled by managing the user profile data.

Although the above description is an example of a route search relationship, the present invention is not limited to this, and can be applied to other fields as long as the search know-how is stored in the two template tables. is there.

The third operation of the present invention will be described with reference to FIG. 7 when the user makes a search request in the natural language "I want to know where the company Fujikawa is".

S20. Information request analysis means 1 shown in FIG.
The search request of "I want to know the place where the company Fujikawa is" is entered.

S21. The information request analysis means 1 performs a semantic analysis on this by natural language analysis, and recognizes that the request type is “location” and the search condition is “Fujikawa (company name)”.

S22. The information request analysis unit 1 determines the answer template table 11 according to the request type and the search condition.
To search.

S23. This makes the request type "location"
In the case of, it is recognized as a search result that there are two types of information to be searched, “address” and “region”. Then, the following request data (1) and (2) are output.

S23-1. As the information request data (No. 1), the request type is “place”, the type of information to be searched, that is, the type of answer is “region”, and the search condition is “Fujikawa (company name) + α” (α is a search) (See template table) is output.

S23-2. As the information request data (No. 2), the request type is “place”, the answer type is “address”, and the search condition is “Fujikawa (company name) + α” (α
Is a search template table).

S24. The information request data (1),
(No. 2) is output to the search command creating means 2. At this time, since the user profile data holding unit 13 has transmitted to the search command creating means 2 that the default type is “location” when the request type is “location”, the information request data 1 Performs a search process for it.

S25. The search command creating means 2 obtains the search condition of “region” from the search template table 12.
As a result, it is recognized that the "newspaper article search engine" on the Internet is used and the search is performed using the search condition (company name).

S26. Based on this, you can obtain the required search results by supporting the following.

S26-1. First, the search command creating means 2
The "newspaper article search engine" is used as the search system, and it recognizes that the company name, that is, (Fujikawa), is used as the search condition, and outputs this to the user to support the search. Based on this output, the user uses the "newspaper article search engine" to search by the search condition (Fujikawa).

S26-2. The user can check the “region” based on this search result, and can obtain an appropriate search result.

In the above description, the present invention has been described with respect to stations and areas, but the present invention is not limited to these.

[0071]

According to the present invention, the following effects can be obtained.

(1) Information request analysis means for generating a request type and information request data, and a search instruction creating means for selecting an appropriate information engine based on the information request data and generating a search condition to be given to the information search engine. With this feature, users who do not know the search know-how can receive assistance for the search procedure, and such users can easily obtain necessary information while using multiple search systems. You can ask.

(2) Since an information search engine corresponding to the type of information to be searched can be provided, the user can select an appropriate one from a large number of information search engines.

(3) As a search expression given to the full-text search engine, a character string condition for further narrowing down the search results is added to the information specified as the search key,
You can get more appropriate answers.

(4) Since the user's input is interpreted by referring to the user's situation at the time of input, an appropriate answer can be obtained from the user's standpoint.

(5) Since the user's input is interpreted based on the profile data recording the user's characteristics,
The profile data can be easily updated according to the user's condition, and an accurate answer from the user's standpoint can be easily obtained.

[Brief description of drawings]

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

FIG. 2 is an embodiment of the present invention.

FIG. 3 is a configuration diagram of an answer template table of the present invention.

FIG. 4 is a configuration diagram of a search template table according to the present invention.

FIG. 5 is an operation explanation (1) of the present invention.

FIG. 6 is a second operation explanation of the present invention.

FIG. 7 is an operation explanation (3) of the present invention.

[Explanation of symbols]

1 Information request analysis means 2 Search command creation means 11 Answer template table 12 Search template table 13 User profile data storage

Claims (5)

[Claims] 1. An information request analysis means for interpreting a user's input and generating information request data including a request type indicating the type of information request regarding the purpose of search, the type of information to be searched, and information serving as a search key. And a search command creating means for selecting an appropriate information search engine based on the information request data and automatically creating search conditions to be given to the selected information search engine. apparatus. 2. The search support apparatus according to claim 1, wherein the search command creating means selects an appropriate information search engine according to the type of information to be searched. 3. The search condition creating means adds a character string condition for narrowing down a search result to information specified as a search key in the information request data as a search expression to be given to a full-text search engine. The search support device according to claim 1, wherein an expression is generated. 4. The search support apparatus according to claim 1, wherein the information request analysis unit analyzes the user's input with reference to the state of the user at the time of input. 5. The search support apparatus according to claim 4, wherein the information request analysis means analyzes the input of the user based on profile data in which the characteristics of the user are recorded.