JPH07334516A - Information retrieval device - Google Patents

Abstract

PURPOSE:To improve the efficiency of a whole system by sequentially issuing a retrieval request from a terminal machine which transmitted the response of a content whose similarity is high in the past based on a past retrieval history. CONSTITUTION:When the retrieval request is inputted to an input part 1, a similarity calculation part 3 calculates similarity between retrieval data on the history which a history storage part 2 stores and retrieval data of the retrieval request inputted from the input part 1. The similarity calculation part 3 sorts history data from data whose similarity is higher after similarity between the retrieval request and plural pieces of history data is obtained. The similarity calculation part 3 extracts the name of the terminal machine transmitting information from history data whose obtained similarity is higher. A network part 4 transmits the retrieval request to the terminal machine through a network. When the network part 4 receives a retrieval response, the obtained retrieval request, a retrieval result and the name of the transmission terminal machine are stored in the history storage part 2 as history data, and a display part 8 displays the retrieval result.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information retrieval device,
In particular, the present invention relates to an information search device such as an electronic notebook, a portable information terminal, a word processor, etc. connected by a network.

[0002]

2. Description of the Related Art Conventionally, various efficient methods have been proposed for retrieving information stored in a distributed manner in devices connected to a network. For example, Japanese Patent Laid-Open No. 5-
Japanese Patent No. 108432 discloses an access control device for a distributed database that enables efficient access based on a superposition broadcast protocol even in a duplicate distributed database processing device.

As shown in FIG. 8, this access control device analyzes a database access request 40 from the database processing request device 13 and creates cursor data 41, and an inquiry processing device 10 which processes the same data into a plurality of database processes. A distributed database protocol processing device 11 that duplicates and divides into devices and dispersively arranges them, a cursor state storage device 32 that stores a cursor state, cursor data 41, and state information 52 regarding the cursor data from the storage device 32. It is determined whether the data can be transmitted to the device 12, and if the data can be transmitted, the protocol data is created and the processing device 11
And a cursor state processing device 31 for requesting transmission of protocol data.

Next, the operation of this conventional device will be described.

The processor 10 analyzes the database access request 40 from the requesting device 13 and creates cursor data 41. The cursor data 41 is sent to the processing device 31 as it is, and the processing device 31 causes the storage device 3 to operate.
The state information 52 relating to the cursor data is extracted from the item 2 and it is determined whether or not the state information 52 is ready to be transmitted to the server device 12. If you are ready to send,
The processing device 31 creates protocol data and requests the processing device 11 to transmit the protocol data. If it is determined that the information should not be transmitted, the information is transmitted to the storage device 32, and the information is awaited until transmission becomes possible.

[0006]

However, in the access control device described in Japanese Patent Laid-Open No. 5-108432, when a search request is issued to all the terminals connected to the network, all the access control devices are Since the search is performed, the efficiency of the entire system becomes very poor.
When the user who inputs the search request simultaneously specifies the search target, it is necessary to input the search target device each time the search request is issued, and the operation is complicated.

For example, in the case of a system in which data is stored in a personal terminal or personal computer and all these devices are connected by a network, a search request issued by one user is indispensable. Searching on all mobile devices and personal computers is very inefficient.

When asking a person about something at work, he or she does not ask the same question to all of them at the same time, and only some of the people who are likely to know the answer to the question are asked. In this case, we are not searching all the accessible data. With this method, all the data is not accessed, so search omissions may occur, but since only some people are asked, the overall processing is extremely small. Therefore, there is a demand for an information retrieval device that can improve the efficiency of the entire system by applying this method.

The present invention has been made in order to solve the above problems, and issues a search request in order from a terminal that has sent a response with a high degree of similarity in the past, based on the past search history. It is therefore an object of the present invention to provide an information search device that can improve the efficiency of the entire system.

[0010]

According to the present invention, the above-mentioned object is to provide an input means for inputting a search request consisting of search data, a data storage means for storing searched data, and a storage means for storing the data in the data storage means. Search means for searching the search data from the searched data, display means for displaying the result of the search by the search means, network means for communicating data and commands via a network, the search request, the search Respondents who respond to the request, history storage means for storing history information of search results, and when the search request is input, a search request for the history information stored in the history storage means and the input search A similarity calculation means for calculating a similarity to a request, obtaining a similarity, and requesting a search request in order from an information search device that has sent a reply with a content having a high similarity in the past. It is achieved by the information retrieval apparatus according to claim 1 comprising.

According to the present invention, the above-mentioned object is to display the order of the information retrieval apparatus to which the similarity calculation means has sent the reply with the content having a high degree of similarity on the display means, and This is achieved by the information retrieving apparatus according to claim 2, further comprising an order confirmation means for requesting confirmation.

According to the present invention, for the above-mentioned object, the history storage means stores the history information in a file and copies a part or all of the file to the history storage means of another information retrieval device. This is achieved by the information retrieval device according to claim 3, which comprises history copying means.

[0013]

According to the information retrieving apparatus of the first aspect, the retrieval request consisting of retrieval data is inputted by the input means of the information retrieving apparatus on the transmission side, and the history information stored in the history storage means is referred by the similarity calculating means. The similarity between the content of the search history and the input search request is obtained, and the search request is issued in order from the information search device that has sent the reply with the content with the high similarity in the past. Is transmitted to a plurality of information retrieval devices on the receiving side via. This search request is received by the network means of each information retrieval device on the reception side, the retrieval means retrieves the searched data stored in the data storage means, and the retrieval result or the like is transmitted by the network means to the information on the transmission side via the network. It is sent to the search device. This search result is received by the network means of the information search device on the transmission side,
The display unit displays the search result and the history storage unit stores the history information such as the search request, the respondent, and the search result. Therefore, it is possible to improve the search efficiency by estimating which information search device has the answer to a new search request from the history information, and to reduce the time and effort of user input, unnecessary calculation time, and network load. You can

According to the information retrieving apparatus of the second aspect, the order confirming means displays the names of the information retrieving apparatuses that have answered with a high degree of similarity on the display means in descending order of the degree of similarity to request confirmation of the retrieval order. Therefore, the user can arbitrarily set the search order.

According to the information retrieving apparatus of claim 3, the history copying means stores the history information in a file and copies a part or all of the file to the history storing means of another information retrieving apparatus. Therefore, even if the device is purchased for the first time, by copying the search history from another terminal, it is possible to obtain the priority order of who to ask.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the information retrieval apparatus according to claim 1 will be described below with reference to the drawings. This embodiment provides an information search device capable of improving the efficiency of the entire system by issuing a search request in order from a terminal that has sent a response with a high degree of similarity in the past, based on the past search history. Is an issue.

In the information retrieval apparatus of this embodiment, a plurality of terminals are connected to each other via a network. As shown in FIG. 1, the terminal includes an input unit 1 as an input unit for inputting a search request having search data, and a history storage as a history storage unit for storing a history of search requests, respondents, search results, etc. Connected to the unit 2, the input unit 1 and the history storage unit 2,
The similarity between the search request of the history stored in the history storage unit 2 and the search request input from the input unit 1 is calculated, and the search request is requested in order from the terminal that has sent the response with the high similarity in the past. Connected to the similarity calculation unit 3 as the similarity calculation unit, the data storage unit 5 as the data storage unit for storing the searched data, the input unit 1 and the data storage unit 5, and stored in the data storage unit 5. A search unit 6 as a search unit for searching the search data from the searched data, a search result buffer 7 as a search result buffer unit that is connected to the search unit 6 and temporarily holds a search result, and is connected to the search result buffer 7. , A display unit 8 as a display unit for displaying an input status, a search status, a search result, etc., a history storage unit 2, a similarity calculation unit 3, a search unit 6, and a search result buffer 7, and a search request and a search are made. And a network unit 4 of the fruit such as a network means for communicating over a network. Note that the search answer may not be stored or only the summary of the answer may be stored because of the storage capacity.

Next, the operation of this embodiment will be described with reference to FIGS.
Description will be given along the flowchart of FIG.

In the transmitting terminal, the input unit 1 waits for a search request to be input (step S1). When the search request is input, the similarity calculation unit 3 calculates the similarity between the search data of the history stored in the history storage unit 2 and the search data of the search request input from the input unit 1 (step S
2). For example, the similarity calculation method uses the distance calculation method disclosed in the document classification device of Japanese Patent Application No. 5-181264 filed by the present applicant. However, the present invention is not limited to this distance calculation method. An example of the similarity calculation method will be described later.

After the similarity between the search request and the plurality of history data is obtained, the similarity calculator 3 sorts the history data in descending order of similarity (step S3). Similarity calculator 3
Extracts the name of the terminal that transmitted the information from the obtained history data having a high degree of similarity. The network unit 4 sends a search request to the terminal via the network (step S4). For example, when the terminal C in the past has sent a response to a search request similar to the current search request, the terminal C is selected as a result of the similarity calculation between the history data and the search request. When the network unit 4 receives the search response (step S5), the obtained search request, search result,
The transmission terminal name is stored in the history storage unit 2 as history data (step S6). The display unit 8 displays the search result (step S7).

On the other hand, when the receiving side terminal receives the search request from the transmitting side terminal (step S8), it is held in the data storage section 5 by the searching section 6 in accordance with the searching request from the transmitting side terminal. The searched data is searched (step S9). The retrieved data is temporarily stored in the retrieval result buffer 7 and then transmitted by the network unit 4 to the transmitting terminal via the network (step S1).
0).

In this flow, the search requests are issued in order to the terminal names in descending order of similarity between the search request and the history data. There are various methods for timing of a search request for one device, reception of the search result, and transmission of a search request for the next terminal. For example, when a search request is issued to a terminal having a high degree of similarity, the search response is waited for, and when the desired response is not obtained, the search request is issued to the next device.
A search request is transmitted to a plurality of terminals in descending order of similarity, and a search reply is awaited. If the desired search response is not obtained, the user is asked whether or not to search for another terminal, and when searching for another terminal, the search is continued according to the similarity order again. A search request is issued to a terminal that does not exist. It is also possible to evaluate both the similarity and the search cost using an evaluation formula and start the search in an appropriate order.

Next, as shown in FIG. 3, the search history stored in the history storage unit 2 stores four items of a history number, a search keyword, a respondent, and the reply content of the keyword. For these items, one history corresponds to one answer to a search request transmitted in the past. The history number can be displayed to the user and can be selected by the user when the calling party of the search request is determined, for example. By displaying this table to the user, the user can know what kind of information was obtained from which terminal in the past, and can issue a search request transmission request only by specifying the number.

The search keyword is a keyword used for a search request input in the past, but here, not only the normal search but also the contents at the time of registering the schedule may be used. For example, when the schedule of the manufacturing confirmation meeting is distributed to the related persons, “manufacturing confirmation meeting” is stored in the item of this search keyword, and the item of the respondent is
The person or terminal name of the destination of the schedule is entered. In this case, nothing is included in the response content. Even when the report of the manufacturing confirmation meeting is sent to the related persons, it may be stored in the same manner.

In the item of the respondent, the name of the person who has sent the reply to the search keyword is stored. Here, the name of one person is illustrated as one item, but there may be a plurality of people. The answer content is the answer content sent by the respondent to the search keyword. The answer content stored here may be all the sent information, a pointer indicating the place where the information is stored, or a keyword extracted from the sent information. In order to extract the keywords, it is possible to extract the independent words from the sent sentences by matching them with the pre-stored independent word dictionary and select some from the most frequently used words. . Do not memorize the items of the reply contents as history,
It is also possible to store only the search keyword.

Next, a similarity table for obtaining the similarity between the search request and the search history will be described with reference to FIG.

The similarity between two nouns is preset in the similarity table, and "0" has the lowest similarity,
“9” has the highest degree of similarity. For example, the similarity between an image and a chip is 6, and the similarity between an image and overtime is 0. This table may be created manually, but in reality, a vocabulary of hundreds to tens of thousands of words is required, so it is desirable to obtain it by some learning means. For example, assuming that the user has issued a search request of “overtime” and “notification”, the similarity between the search keyword and response content of history number 1 and the search keyword of this search request is similarity = ((overtime, Image) + (overtime, processing) + (overtime,
Chip) + (Overtime, New Product) + (Overtime, Release) + (Overtime, High Speed) + (Notification, Image) + (Notification, Processing) + (Notification, Chip) + (Notification, New Product) + (Notification, Release) +
(Notification, high speed)) / 12. Here, (overtime, image) and the like represent the degree of similarity between the overtime and the image in the similarity table. (overtime,
The reason why the sum of the similarities of (images) etc. is divided by 12 is to obtain the average. The value of 12 is determined by the number of independent words in the search request and the number of independent words in the search history.

When the degree of similarity with only the part of the search keyword is not stored and the answer contents are not stored, the degree of similarity = ((overtime, image) + (overtime, processing) + (overtime,
Chip) + (overtime, new product) + (overtime, release) + (overtime, high speed)) / 6. Each history in this case and "overtime", "notification"
The degree of similarity with the search request is as shown in FIG. 5, and the similarity order is 3, 1, 2, and 4 in that order. It should be noted that the procedure for obtaining the degree of similarity in consideration of the response content is also the same. As a result, the search request is first transmitted to Mr. Kasagi's terminal with history number 3. In order to determine to which terminal in the order of similarity the search request is to be sent, a threshold is set for the similarity, and if the first similarity ratio is equal to or higher than the threshold, the search request is sent. There is also a way to

The present invention is a system that assumes that a plurality of terminals will search using a network. For example, a search request is input from the terminal A, and the search request is transmitted to the terminals B, C, and D connected to the network. Each of the terminals B, C and D performs a search, and the obtained answer is transmitted to the terminal A. At this time, regardless of the contents of the search, if the search requests are simultaneously sent to the terminals B, C, and D, the search is always performed by the three terminals,
ineffective. Therefore, in the present invention, it is possible to improve the efficiency of the entire system by sequentially making a search request to the terminals that have sent the answers with high similarity in the past.

According to the information retrieving apparatus of the present embodiment, the retrieval efficiency is improved by estimating which information retrieving apparatus has the answer to a new retrieval request from the history information, and the labor of user input is increased. It is possible to reduce unnecessary calculation time and network load.

An embodiment of the information retrieval apparatus according to claim 2 will be described below with reference to the flowchart of FIG. In addition,
The same steps as those in FIG. 2a are designated by the same reference numerals and the description thereof will be omitted. The present embodiment presents the candidate of the terminal name to be searched using the history data to the user, but it is an object to provide an information search device having an order confirmation function that allows the user to specify the final search order. .

The terminal names that have transmitted similar information in the past are obtained from the history data, and then the terminal name list in descending order of similarity is displayed on the display unit 8 and the list is searched in that order. Ask the user for confirmation (step S1)
1). At this point, the user can modify the search order if desired, and can also modify the search order to be output to all terminals. According to the present embodiment, candidates are presented using history data for the order of the terminals to be searched, but the final decision is left to the user.

In the display example displayed on the display unit 8, as shown in FIG. 7, the user obtains information regarding overtime work and notification and refers to the history information. As a result, Mr. Kasagi's terminal is given first priority. This indicates that the ranking is selected as a candidate for sending a search request. If the user can keep the display, "O
If "K" is designated by a pointing means such as a light pen, a search request is sent to Mr. Kasagi. If you specify the name of another person, the search request will be sent to that person as well. If you specify the "Continue" part, the next page of the name list is displayed.

When setting the priorities, it is preferable to list only the names of terminals that can currently communicate. When using a LAN, the communicable partner varies depending on the current location. Even if the name of the terminal that cannot communicate on the spot is displayed, communication cannot be actually performed, so terminals that cannot communicate are excluded.

An embodiment of the information retrieving apparatus according to claim 3 will be described below. It is an object of the present embodiment to provide an information search device having a search history file copy function for storing a search history in a file format and copying the file to the history storage means of another information search device.

The search history stored in the history storage unit 2 is saved in the form of a file. Part or all of the search history file stored in the history storage unit 2 is transferred to another terminal device via the network and copied or added to the history storage unit 2 of the other terminal device. Further, part or all of the search history file may be copied or added to the history storage unit 2 of another terminal by a floppy disk or a memory card without going through the network. By doing so, even if the device is purchased for the first time, it is possible to obtain the priority order of who to ask by copying the search history from another terminal.

It is also possible to prepare a plurality of history files and switch and use them. For example, when used in the office 1, a history file for the office 1, and when used in the office 2, a history file for the office 2, such as a terminal that is easy to communicate on the spot or a terminal with low communication cost. Can be selected.

[0038]

According to the information retrieving apparatus of the first aspect, the retrieval efficiency is improved by estimating which information retrieving apparatus has the answer to a new retrieval request from the history information, and the retrieval efficiency is improved. It is possible to reduce the effort, the wasteful calculation time, and the load on the network.

According to the information retrieval apparatus of the second aspect, the user can arbitrarily set the retrieval order.

According to the information retrieving apparatus of claim 3, even if the apparatus is purchased for the first time, by copying the retrieval history from another terminal, it is possible to obtain the priority order of who to ask. it can.

[Brief description of drawings]

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

FIG. 2a is a flowchart showing the operation of the transmitting side of the information search device of the present invention.

FIG. 2b is a flowchart showing the operation of the receiving side of the information search device of the present invention.

FIG. 3 is a diagram showing a search history of the information search device according to the present invention.

FIG. 4 is a diagram showing a similarity table of the information search device according to the present invention.

FIG. 5 is a diagram showing a history number and a degree of similarity.

FIG. 6 is a flowchart showing the operation of the present invention.

FIG. 7 is a diagram showing a display example when a search request is input.

FIG. 8 is a block diagram showing a configuration of a conventional access control device.

[Explanation of symbols]

 1 input section 2 history storage section 3 similarity calculation section 4 network section 5 data storage section 6 search section 7 search result buffer 8 display section

Claims (3)

[Claims] 1. Input means for inputting a search request comprising search data, data storage means for storing searched data, and search means for searching the search data from the searched data stored in the data storage means. Display means for displaying the results retrieved by the retrieval means, network means for communicating the retrieval request and the retrieved results via a network, and the respondent responding to the retrieval request and the retrieval request , And history storage means for storing three histories of the searched results, and when the search request is input, the history search request stored in the history storage means and the input search request An information search device comprising a similarity calculation means for calculating a similarity and requesting a search request in order from an information search device that has sent a reply with a high similarity in the past. 2. The similarity calculation means includes an order confirmation means for displaying on the display means the order of the information retrieval apparatus that has sent the reply with the content of high similarity in the past and for confirming the retrieval order. The information search device according to claim 1. 3. The history storage means includes history copy means for storing the history information in a file and copying a part or all of the file to the history storage means of another information retrieval device. Alternatively, the information search device according to claim 2.