JP2018190366A - Retrieval device, retrieval method and retrieval program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a retrieval device, a retrieval method and a retrieval program by which retrieval effectiveness of documents can be improved.SOLUTION: A retrieval device 1 comprises: a classification part 11 classifying document data consisting of a plurality of items stored in a database on the basis of coincidence degree of the first item where a combination of one or more strings are described; and an extraction part 12 extracting a keyword group included in the second items from the document data in the same group classified by the classification part as related words of retrieving. When the retrieval device 1 is used as a distributing system on the network, it enables creation of related words about a plurality of kinds of machines, data storage thereof, and usage by a plurality of users.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a search device, a search method, and a search program for searching for a document using a keyword.

Conventionally, there is a case where information is acquired by keyword search from a past failure report that has been accumulated in order to solve a failure currently occurring in a machine tool or the like at a manufacturing site.
As a keyword search technique, for example, Patent Literature 1 proposes a technique for displaying guidance for narrowed keyword candidates in descending order of the cumulative number of keywords existing in a document group in order to guide the search narrowing in a desired direction. (For example, refer to Patent Document 1).

JP-A-5-314182

  By the way, when inputting the current phenomenon as a keyword, there are cases in which past documents cannot be appropriately searched because the same phenomenon may be expressed in different words depending on the user. For example, when a failure such as a sound from the machine occurs, it is conceivable that this phenomenon is expressed by different keywords such as “abnormal noise”, “vibrates”, “noise”, etc. by the user.

  In the prior art, these keywords are handled independently without considering the relevance of the keywords. Therefore, in order to widely search related documents, the number of searches increases, or “noise” or “vibrates” or It was necessary to perform a complex search using a logical operation expression such as “noisy”. In addition, if the user cannot come up with these related keywords, the user cannot obtain a desired search result.

  It is an object of the present invention to provide a search device, a search method, and a search program that can improve document search efficiency.

  (1) A search device according to the present invention (for example, “search device 1” described later) has one or more items in a database (for example, “document database 21” described later) in which document data including a plurality of items is stored. A classification unit (for example, “classification unit 11” described later) for classifying the document data based on the degree of match of the first item in which a combination of symbol strings is described, and within the same group classified by the classification unit An extraction unit (for example, “extraction unit 12” described later) that extracts a keyword group included in the second item as a related word for search.

  (2) In the search device according to (1), the classification unit may classify document data in which the first item is completely identical into the same group.

  (3) In the search device according to (2), the classification unit may classify document data in which at least a part of the first item matches into the same group.

  (4) The search device according to any one of (1) to (3) includes a reception unit (for example, “reception unit 13” described later) that receives a search request based on a first keyword, and the first keyword. A search unit that searches the document data and searches for the document data from a second keyword included in the same keyword group as the first keyword (for example, “search unit 14” described later). Also good.

  (5) In the search device according to (4), the accepting unit may display a keyword group including the first keyword and accept an input of the second keyword.

  (6) In the search device according to (4), the search unit sequentially extracts the second keyword from a keyword group including the first keyword, and extracts the document data by each of the second keywords. You may search.

  (7) The search device according to any one of (4) to (6) manages at least two or more different databases as one set of database groups, and each database included in the one set of database groups. A group management unit (for example, “group management unit 15” described later) that manages the keyword group extracted as a related word for search by the extraction unit as a set of keyword groups, In response to a search request based on the first keyword, the document data stored in the database included in the one set of database groups is searched, and in the one set of keyword groups, the same as the first keyword The document data may be searched by a second keyword included in the keyword group.

  (8) In the search device according to any one of (4) to (7), the reception unit receives one or more sentences as the search request, and then extracts the extraction unit from the one or more sentences. May extract the first keyword by a procedure common to the procedure for extracting the keyword group.

  (9) The search device according to any one of (4) to (7) is based on a sensor input unit (for example, “sensor input unit 16” described later) for inputting sensor information of the device and a predetermined conversion condition. A status acquisition unit (for example, “status acquisition unit 17” described later) that converts the sensor information into a status keyword that indicates the status of the device, and the reception unit converts the status keyword to the first keyword. It may be accepted as a keyword.

  (10) In the search device according to any one of (1) to (9), the document data is a device failure report, and the symbol string is a specification number and the number of parts to be repaired. May be.

  (11) In the search method according to the present invention, the document data is based on the matching degree of the first item in which a combination of one or more symbol strings is described in a database in which document data including a plurality of items is stored. The computer executes a classification step for classifying and the extraction step for extracting the keyword group included in the second item from the document data in the same group classified in the classification step.

  (12) In the search method according to (11), the computer further accepts a search request with a first keyword, searches the document data with the first keyword, and the first keyword. And a search step for searching the document data from a second keyword included in the same keyword group.

  (13) In the search method according to (12), the computer further manages at least two different databases as a set of database groups, and for each database included in the set of database groups, Performing a group management step of managing the keyword group extracted as a related word for search in the extraction step as a set of keyword groups, and in the search step, in response to a search request by the first keyword, The document data stored in the database included in a set of databases is searched, and the document is searched by a second keyword included in the same keyword group as the first keyword in the set of keyword groups. Data may be searched.

  (14) In the search method according to (12) or (13), after the computer further receives one or a plurality of sentences as the search request in the reception step, from the one or a plurality of sentences, In the extraction step, the first keyword may be extracted by a procedure common to the procedure for extracting the keyword group.

  (15) In the search method according to (12) or (13), the computer further inputs a sensor input step of sensor information of the device, and the sensor information is determined based on a predetermined conversion condition. A state acquisition step of converting into a state keyword indicating, and in the reception step, the state keyword may be received as the first keyword.

  (16) A search program according to the present invention is for causing a computer to function as the search device according to any one of (1) to (10).

  According to the present invention, document retrieval efficiency can be improved.

It is a block diagram which shows the function structure of the search device which concerns on 1st Embodiment. It is a figure which illustrates the document data stored in the document database which concerns on 1st Embodiment. It is the schematic which shows the preparation procedure of the related term database which concerns on 1st Embodiment. It is a flowchart which illustrates the related term data process which concerns on 1st Embodiment. It is the schematic which illustrates the procedure of the keyword search which concerns on 1st Embodiment. It is a figure which shows the example of a screen display of the related word which concerns on 1st Embodiment. It is a block diagram which shows the function structure of the search device which concerns on 2nd Embodiment. It is a schematic diagram at the time of implementing the search device concerning 2nd Embodiment as a distributed system on a network. It is a 1st figure which shows the example of a display of the search screen which concerns on 2nd Embodiment. It is a 2nd figure which shows the example of a display of the search screen which concerns on 2nd Embodiment. It is a 3rd figure which shows the example of a display of the search screen which concerns on 2nd Embodiment. It is a block diagram which shows the function structure of the search device which concerns on 4th Embodiment. It is a figure which illustrates the conversion database which concerns on 4th Embodiment. It is a figure which shows the example of a display of the search result which concerns on 4th Embodiment. It is a schematic diagram at the time of implementing the search device concerning 4th Embodiment as a distributed system on a network.

[First Embodiment]
The first embodiment of the present invention will be described below.
The search device 1 of the first embodiment searches for document data stored in a database based on a keyword input from a user.
The document data to be searched is a device failure report accumulated as a past case, and includes, for example, each data of phenomenon, cause, countermeasure, and parts. The user obtains information on causes, countermeasures, and parts by searching for similar phenomena that have occurred in the past.

FIG. 1 is a block diagram illustrating a functional configuration of the search device 1 according to the first embodiment.
The search device 1 is an information processing device (computer) such as a server device or a PC, and includes a control unit 10 and a storage unit 20. Furthermore, the search device 1 includes various data input / output devices, communication devices, and the like, and searches a document database based on a search keyword input from a user and outputs a search result.

  The control unit 10 is a part that controls the entire search device 1 and implements various functions in the first embodiment by appropriately reading and executing various programs stored in the storage unit 20. The control unit 10 may be a CPU.

  The storage unit 20 is a storage area for various programs and various data for causing a hardware group to function as the search device 1, and may be a ROM, a RAM, a flash memory, a hard disk (HDD), or the like. Specifically, the storage unit 20 is synonymous with a document database (DB) 21 in which document data to be searched is stored, in addition to a search program for causing the control unit 10 to execute each function of the first embodiment. A related word file 22 or the like in which the classified keyword group is described in association with each other is stored. Note that these data may be provided outside the search device 1 and read and written by communication with the search device 1.

  The control unit 10 includes a classification unit 11, an extraction unit 12, a reception unit 13, and a search unit 14, and the search of document data using related words of the input keyword is realized by each of these function units. To do.

The classification unit 11 includes a first item (for example, a combination of one or more symbol strings in the document database 21 in which document data including a plurality of items (for example, phenomena, causes, countermeasures, and parts) is stored. ), The document data is classified based on the degree of matching, and the classification result is provided to the extraction unit 12. The classification result may be stored in the document database 21.
Here, the symbol string is the specification number and the number of parts to be repaired or candidates.
The classification unit 11 may classify the document data whose first item completely matches into the same group, or classify the document data with at least a part (for example, pre-defined main parts) into the same group. May be.

  The extraction unit 12 extracts a keyword group included in the second item (phenomenon) from the document data within the same group classified by the classification unit 11 as a related word, and creates a related word file 22.

The accepting unit 13 accepts a search request for document data using the first keyword.
Further, the receiving unit 13 receives a selection input of the second keyword from the keyword group by extracting and displaying the keyword group of the related word including the first keyword from the related word file 22.

The search unit 14 searches the document data in the document database 21 using the first keyword, and in the related word file 22, the second word included in the keyword group of related words classified into the same group as the first keyword. Search document data by keyword.
At this time, the search unit 14 may wait for the reception unit 13 to receive the second keyword and perform a search using the second keyword.
Alternatively, the search unit 14 may sequentially extract the second keyword from the keyword group including the first keyword, and automatically search the document data by each of the extracted second keywords.

FIG. 2 is a diagram illustrating document data stored in the document database 21 according to the first embodiment.
The document data as the failure report includes a phenomenon, a search keyword A (n) extracted from the phenomenon, a cause B (n), a countermeasure C (n), and a management number for identifying the document data. , Part D (n) and total S (n).
The search device 1 searches the document data by matching the input keyword with the search keyword A (n) extracted from the phenomenon.

In the part item (first item) D (n), the number of pieces of information is added to each part, and for example, “X0015 (1)” is indicated by a number in parentheses. In this case, whether or not classification into the same group is possible is determined based on the degree of matching of the parts including the number.
In addition, the description format of components is not restricted to this, Moreover, an item may be divided for every components.

Since the item (first item) D (n) of the document data indicated by the management numbers 12, 145 and 332 is common to “X0015 (1) + V002 (1)”, it is classified into the same group. Then, the keyword A (n) described in the phenomenon (second item) (for example, “sounds abnormal”, “vibrates”, “noises”) is extracted as a related word.
In addition, since the item (first item) D (n) of the document data indicated by the management numbers 23, 775, and 980 is common to “AZX (1) + Z055 (1)”, the same group And keywords A (n) (for example, “does not move”, “does not work”, “stops”) described in the phenomenon (second item) are extracted as related words.

In this example, when the first items are completely matched, they are classified into the same group. However, the classification method based on the degree of match is not limited to this. The document data may be classified by partial matching including, for example, a common part (specification or specification and number), or may be classified by an inclusion relationship in which one is completely included in the other. The definition of the degree of match classified as the same group can be set as appropriate.
Total S (n) is flag data indicating that processing has been completed in related word data processing (FIG. 4) described later.

FIG. 3 is a schematic diagram illustrating a procedure for creating the related word file 22 according to the first embodiment.
The search device 1 performs a keyword creation process on the document database 21, and one or more keywords for search are set for each document data.

Next, keywords are grouped by related word data processing, and a related word file 22 is created.
For example, RF (1) = {A (1), A (5) for keywords A (1), A (2), A (3), A (4), A (5) for each document data. }, RF (2) = {A (2)}, RF (3) = {A (3), A (4)}, and the related word file 22 including the array RF in which a plurality of keywords are associated is obtained. Created.

FIG. 4 is a flowchart illustrating related word data processing according to the first embodiment.
In this example, perfect matching is adopted as the matching degree of parts classified into the same group.

  In step S1, the control unit 10 (classification unit 11) initializes variables r, m, and n (r = 0, m = 1, n = 1). In addition, the control unit 10 initializes the aggregation flag S (i) (1 ≦ i ≦ nMax) of each document data (management numbers 1 to nMax) to 0.

  In step S <b> 2, the control unit 10 (classification unit 11) determines whether or not the mth document data has been aggregated (S (m) = 1). If this determination is YES, the process proceeds to step S8, and if the determination is NO, the process proceeds to step S3.

  In step S3, the control unit 10 (extraction unit 12) extracts the keyword A (m) of the m-th document data and stores it in the related word file 22 (RF (m, 0) = A (m)). Then, the control unit 10 (extraction unit 12) updates the m-th aggregation flag to have been aggregated (S (m) = 1).

  In step S4, the control unit 10 (classification unit 11) compares the mth and nth document data, and the parts are common (D (m) = D (n) and m ≠ n and S (n) ≠ 1. ). If this determination is YES, the process proceeds to step S5, and if the determination is NO, the process proceeds to step S6.

  In step S5, the control unit 10 (extraction unit 12) extracts the keyword A (n) of the nth document data as the rth related word for the mth document data and stores it in the related word file 22 ( RF (m, r) = A (n)). Then, the control unit 10 (extraction unit 12) counts up the variable r and updates the nth aggregation flag to have been aggregated (S (n) = 1).

In step S6, the control unit 10 (classification unit 11) counts up the variable n (n = n + 1).
In step S7, the control unit 10 (classification unit 11) determines whether or not the variable n has exceeded the maximum value nMax. If this determination is YES, the process proceeds to step S8, and if the determination is NO, the process proceeds to step S4.

  In step S8, the control unit 10 (classification unit 11) determines whether nMax and the variable m are equal. If this determination is YES, the process ends. If the determination is NO, the process moves to step S9.

  In step S9, the control unit 10 (classification unit 11) counts up the variable m (m = m + 1) and initializes the variable n (n = 1). Thereafter, the process proceeds to step S2.

Thus, the related word data processing groups the keywords of the document data as related words and totals them in the related word file 22.
Further, the extraction unit 12 may perform a determination process, for example, using a high-level predetermined number of keywords that are aggregated in the same group as related words, and obtain only reliable information from the aggregation result. Good. Even when a plurality of keywords are extracted from the document data, a determination process based on the frequency of association may be performed independently of each other.

FIG. 5 is a schematic diagram illustrating a keyword search procedure according to the first embodiment.
The search device 1 searches the document data stored in the document database 21 in response to receiving a keyword input.
At this time, the search device 1 reads the related word file 22 and displays the related words collected in the same group as the input keyword as guidance, thereby prompting the user to perform an additional search using the related words.

FIG. 6 is a diagram illustrating a screen display example of related terms according to the first embodiment.
The search device 1 displays related words (eg, “noise”, “abnormal noise”) for the input keyword (eg, “vibration”).
At this time, the related words to be displayed may be arranged in descending order of the degree of relevance according to the frequency totaled in the same group, for example.
The search device 1 accepts a selection input from the displayed related words, and performs a search process using the selected related words as a new keyword.

  According to the first embodiment, the search device 1 classifies document data based on the degree of match of the first item (part) among a plurality of items constituting the document data, and performs search from the document data of the same group. A keyword group included in item 2 (phenomenon) is extracted as a related word, and a related word file 22 is created. Thereby, the search device 1 can provide a user with related terms for searching a wide range of documents described using different terms, thereby improving search efficiency.

The search device 1 can provide a wide range of case documents related to the keywords by performing a search using related words stored in the related word file 22 in addition to the keywords first received from the user.
For example, when the specifications and the number of repair parts used in the device are the same, the failure phenomenon is likely to be the same. Since the search device 1 treats keywords expressing the same phenomenon as synonymous related words and searches for related words as search keywords, document search efficiency can be improved.

  The search device 1 can improve accuracy by adopting perfect match as the degree of match of the first item classified into the same group, and can extract a wide range of related words by adopting partial match it can.

The search device 1 displays related words of the input first keyword, and accepts selection input of the second keyword from these. Therefore, since the search device 1 can provide guidance to the keyword related to the user, it is possible to reduce the burden on the user who devise and input the search keyword and improve the search efficiency.
In addition, the search device 1 can perform a search automatically using related words sequentially, thereby simplifying the operation by the user and searching a wide range of documents.

[Second Embodiment]
As described above, the document data accumulated in the document database 21 according to the first embodiment is, for example, document data including phenomena, causes, countermeasures, parts, and the like accumulated as past cases regarding one or a plurality of devices having the same specifications. there were. As described above, the search device 1 according to the first embodiment can search the document data based on related words related to the document data related to, for example, one or a plurality of devices having the same specification.
By using the search device 1 according to the first embodiment, the user searches for similar phenomena that occurred in the past in one or a plurality of devices having the same specifications, thereby obtaining information on causes, countermeasures, and parts. I was able to get it.
On the other hand, at a manufacturing site, for example, a device related to a mechanism (hereinafter also referred to as “mechanical device”), which is a different manufacturer, and a common device (hereinafter also referred to as “control device”) that are the same manufacturer related to control. In many cases, a single machine tool is used in combination. In such a case, it has not been easy to comprehensively manage and search the failure report of the mechanical device and the failure report of the control device as a set of related failure reports.

The search device 1 according to the second embodiment manages document data stored in two or more different document databases 21 as a set of related document data, and uses the document data stored in each document database 21 for search. The keyword group extracted as a related word is comprehensively managed as a set of keyword groups. Thereby, the search device 1 enables two or more pieces of document data to be simultaneously searched as a set of document data groups.
For example, as described above, when a mechanical device with a different manufacturer is combined with the same common control device and used as a single machine tool, the failure report for the mechanical device and the failure report for the control device are sent. It becomes easy to manage and search comprehensively as a set of failure reports.
Hereinafter, the search device 1 according to the second embodiment of the present invention will be described.
In addition, about the component same as the search device 1 which concerns on 1st Embodiment, it shows using the same code | symbol and the description is abbreviate | omitted.

FIG. 7 is a block diagram illustrating a functional configuration of the search device 1 according to the second embodiment.
As in the first embodiment, the search device 1 is an information processing device (computer) such as a server device or a PC, and includes a control unit 10 and a storage unit 20. Further, the search device 1 includes various data input / output devices, communication devices, and the like. Based on the input of a search keyword from a user, the search device 1 stores the document data stored in each document database 21 included in a set of databases. Search and output the search results.

  The storage unit 20 is synonymous with two or more document databases (DB) 21 each storing document data to be searched in addition to a search program for causing the control unit 10 to execute each function of the second embodiment. Two or more related word files 22 or the like in which the classified keyword group is described in association with each other are stored. Note that these data may be provided outside the search device 1 and read and written by communication with the search device 1.

  The control unit 10 includes a classification unit 11, an extraction unit 12, a reception unit 13, a search unit 14, and a group management unit 15, and related words of keywords input from the user by these functional units. Is used to search the document data stored in each document database 21 included in a set of databases.

  The classification unit 11 and the extraction unit 12 perform the same processing as in the first embodiment, for example, on the document database 21 to be processed by an instruction from the user. By doing so, a related word file 22 is created for each document database 21.

For example, when the phenomenon data of the document data (failure report) relating to the mechanical device includes “concentrated lubricating oil is declining fast” and “cannot release the door lock”, the classification unit 11 and the extracting unit 12 may The word file 22 describes {concentrated lubricating oil = lubricating oil}, {cannot release the door lock = cannot open the door}, and the like.
Further, when the phenomenon data of the document data (failure report) related to the control device includes “fan alarm of control device” and “the screen display has become darker since yesterday”, the classification unit 11 and the extraction unit 12 {Fan = fan motor = fan motor}, {Screen is dark = Screen is not visible = Power is not turned on, etc. are described in the related word file 22 of the control device.

For example, the group management unit 15 manages at least two or more different document databases 21 as a set of database groups based on an instruction from the user. At the same time, the group management unit 15 manages a keyword group extracted as a related word for search by the classification unit 11 and the extraction unit 12 for each document database 21 included in the set of database groups as a set of keyword groups. To do.
In order to manage at least two or more different document databases 21 as a set of database groups, the group management unit 15 creates a management table for the document database 21 managed as a set of database groups, and virtually You may make it handle as one file. For example, the management table may be configured to link (link) identification information of a set of database groups and identification information of two or more document databases 21 included in the set of database groups. . In addition, the document database 21 managed as a set of databases may be stored as one physical file.

The accepting unit 13 accepts a search request for document data using the first keyword. More specifically, the reception unit 13 receives, for example, identification information of a set of database groups corresponding to a set of machine tools, a first keyword, and the like, which are input by the user, as input to the search device 1. Accept through. The receiving unit 13 may receive these input data from a client terminal via a network, for example.
Furthermore, the reception unit 13 extracts and displays the keyword group including the first keyword from all the related word files 22 included in the set of database groups, and thereby displays the second keyword from the keyword group. Accept keyword selection input. By doing so, the user does not need to be aware of which document database 21 the first keyword is the keyword related to the document data.

The search unit 14 searches the document data stored in all the document databases 21 included in the set of database groups by the first keyword, and also includes the second keyword included in the keyword group including the first keyword. Thus, the document data stored in the document database 21 corresponding to all or the keyword group is searched.
As in the first embodiment, the search unit 14 may wait for the reception unit 13 to receive the second keyword and perform a search using the second keyword.
Alternatively, the search unit 14 may sequentially extract the second keyword from the keyword group including the first keyword, and automatically search by each of the extracted second keywords.

FIG. 8 is a schematic diagram when the search device 1 according to the second embodiment is implemented as a distributed system on a network and applied to a mechanical device and a control device.
It should be noted that when the search device 1 is used via a network, known access user authentication, access control of the user (granting access authority permitted to the user), and the like are performed.
As shown in FIG. 8, the manufacturer or the maintenance manager of the mechanical device (i) (1 ≦ i ≦ n) and / or the control device can send document data (that is, A related word file corresponding to the document data of each device can be generated from a failure report (which includes data on phenomena, causes, countermeasures, parts, and the like accumulated as past cases). Here, the control device is common to a plurality of mechanical devices (i).

Then, the manufacturer or maintenance manager of the mechanical device (i) and / or the control device can identify the control device identification information and the mechanical device (i) configured as a set of machine tools (i) via the network. The document data (failure report) of the mechanical device (i) configured as a set of machine tools (i), the document data (failure report) of the control device, and the mechanical device (i ) Related word file 22 related to the document data and related word file 22 related to the document data of the control device can be managed in association with a set of related databases (i).
By doing so, the user, the mechanical device (i), the manufacturer and the maintenance manager of the control device, and the like can control the control device and the mechanical device (i ) Document data (failure report) can be efficiently searched.

9A to 9C are diagrams illustrating display examples of a search screen of the search device 1 according to the second embodiment.
With the input screen illustrated in FIG. 9A, the reception unit 13 receives input of identification information (device ID) of a set of machine tools to be searched. By doing so, the reception unit 13 specifies the document database 21 in which the document data (failure report) of the control device and the mechanical device constituting the machine tool is accumulated.
Next, on the input screen shown in FIG. 9B, the accepting unit 13 accepts an input of the inquiry content (document data search request using the first keyword).
By doing so, the reception unit 13 can extract and display the related words of the first keyword based on the related word file 22 related to the control device and the related word file 22 related to the mechanical device.
At this time, the reception unit 13 determines whether the extracted related word is a related word extracted from the related word file 22 related to the control device, a related word extracted from the related word file 22 related to the mechanical device, or common to both. It is preferable to display the related words. By doing so, the user can easily grasp the relationship between the first keyword, the control device, and the mechanical device.

Thereafter, the search unit 14 performs document data (failure report) relating to the control device and / or document data relating to the mechanical device (based on the related word selected by the user or the related word having a high priority set in advance ( Search for failure reports.
The search unit 14 extracts document data (failure report) related to the control device and / or document data (failure report) related to the mechanical device as a search result.
As a result, as shown in FIG. 9C, the search unit 14 can display search results, for example, cause and countermeasure candidates.

  According to the second embodiment, when the search device 1 combines a mechanical device with a different manufacturer and a common control device with the same manufacturer into one machine tool, the document data of the mechanical device and the control device is obtained. (Fault report) can be comprehensively managed as a set of databases.

Further, by implementing the search device 1 as a distributed system on the network, document data (failure reports) of mechanical devices and control devices related to a plurality of machine tools can be referred to by a plurality of users. Thereby, it is not necessary to construct a plurality of systems, and the system can be constructed at low cost.
Further, the search device 1 separately generates and stores the document data of the control device and the document data of the mechanical device, and thereafter, the document database 21 in which the document data of the control device is accumulated, and the document data of the mechanical device are stored. By managing the stored document database 21 as a set of databases, it is possible to support a plurality of machines. Thereby, for example, in a group of machines having the same control device but different mechanical devices, information on the control device can be shared, and the related word file creation time can be omitted.

  In the second embodiment, a case has been described in which a mechanical device with a different manufacturer and a common control device with the same manufacturer are combined to form one machine tool, but the present invention is not limited to this. The second embodiment can also be applied to a case where an arbitrary mechanical device and an arbitrary control device are combined to form a single machine tool.

  In the second embodiment, the document data (failure report) of the mechanical device and the control device has been described as an example. However, the document data is not limited to machine information. The search device 1 includes two or more different document databases 21 in which document data having a first item in which a combination of at least one or more symbol strings is described and a second item including a keyword are stored in common. Can be managed as a set of databases.

[Third Embodiment]
Hereinafter, a third embodiment of the present invention will be described.
In 3rd Embodiment, the structure of the reception part 13 in the search device 1 differs from 1st and 2nd embodiment.

In the first and second embodiments, the data received by the receiving unit 13 as a search request is the first keyword. In this case, the user himself / herself considers the keyword, and narrows down desired document data using a logical operation such as AND or OR. However, there are cases in which the user wants to input a conversation sentence or the like explaining the state of the device as it is.
Therefore, the search device 1 of the third embodiment accepts input of one or more sentences as a search request.

After accepting one or more sentences as a search request, the accepting unit 13 extracts a first keyword from the one or more sentences by a procedure common to the extraction unit 12 extracting a keyword group.
Morphological analysis is usually used as a procedure for extracting a keyword group. At this time, a keyword group is determined by referring to a database of unnecessary words that are excluded from extraction targets and synonyms (thesaurus) that are collected as synonyms.
By referring to this database in common in the extraction unit 12 and the reception unit 13, the search device 1 matches the keyword group of the related word file 22 with the first keyword as the search request, and as a result, the search accuracy is increased. Will improve.

[Fourth Embodiment]
The fourth embodiment of the present invention will be described below.
In addition, about the component same as 1st-3rd embodiment, it shows using the same code | symbol and the description is abbreviate | omitted.

In the first and second embodiments, the data received as a search request is the first keyword. In the third embodiment, the first keyword is extracted from one or more accepted sentences. In these cases, for example, when searching for knowledge information about a machine failure, the user needs to interpret the state of the failure and create a description or a keyword.
Therefore, the search device 1 according to the fourth embodiment automatically generates the first keyword according to the conversion condition based on the empirical rule using the sensor information obtained from the device as an input.

FIG. 10 is a block diagram illustrating a functional configuration of the search device 1 according to the fourth embodiment.
The control unit 10 of the search device 1 includes a sensor input unit 16 and a state acquisition unit 17 in addition to the classification unit 11, the extraction unit 12, the reception unit 13, the search unit 14, and the group management unit 15.
The storage unit 20 includes a conversion database (DB) 23 in addition to the document database 21 and the related word file 22.

The sensor input unit 16 receives sensor information including measurement data as input from various sensors installed in a device to be monitored (for example, the above-described mechanical device or control device).
The state acquisition unit 17 converts the received sensor information into a state keyword indicating the state of the device based on a predetermined conversion condition stored in the conversion database 23. This status keyword is input to the receiving unit 13 as a first keyword for document search.

Here, the conversion condition stored in the conversion database 23 is a condition for comparing a sensor value or a calculation result (for example, an integral value) of the sensor value with a threshold value and converting it into a state keyword. It is preset from an empirical rule based on it.
Further, the state acquisition unit 17 may calculate a feature amount from the sensor information using, for example, FFT or principal component analysis, and may determine a state keyword based on a conversion condition based on the feature amount.

FIG. 11 is a diagram illustrating a conversion database 23 according to the fourth embodiment.
In this example, conditions for converting the temperature and pressure acquired as sensor information into state keywords are set.
For example, when the number of times the temperature exceeds the threshold value within a predetermined measurement time is 5 times or more and the actual integrated value of the temperature is equal to or greater than a set value, the state acquisition unit 17 acquires the state keyword “overheat”. .
Similarly, when the number of times the temperature exceeds the threshold value within the predetermined measurement time is 5 times or more and the actual integrated value of the temperature is less than the set value, the state acquisition unit 17 sets the state keyword “load abnormality”. get.
Further, when the current pressure exceeds the threshold value and the actual integrated value of the pressure within a predetermined measurement time is equal to or greater than the set value, the state acquisition unit 17 acquires the state keyword “pressure abnormality”.

FIG. 12 is a diagram illustrating a display example of search results by the search device 1 according to the fourth embodiment.
The search unit 14 searches the document database 21 after acquiring a related word from the related word file 22 using the state keyword converted by the state acquisition unit 17 as a first keyword.

When the document data is searched based on the state keyword converted from the sensor information, for example, “overheat” and the related word, information such as a phenomenon, a cause, a countermeasure, and a part is displayed as a search result.
In this example, the control device that monitors the sensor data senses the state “overheat” and a corresponding case that is a search result is displayed on the screen.
Also, for example, when another document related to the part replacement method is searched for “replacement” described in the corresponding case, the link “replacement method” is displayed with an underline, and this link is selected. Thus, the screen may transition to display of another document.

FIG. 13 is a schematic diagram when the search device 1 according to the fourth embodiment is implemented as a distributed system on a network.
In this case, the sensor input unit 16, the state acquisition unit 17, and the conversion database 23 may be provided outside the search device 1.

For example, the control device 100 that controls the device to be monitored includes the sensor input unit 16 and the state acquisition unit 17. The control device 100 transmits the status keyword converted based on the conversion database 23 to the search device 1 (accepting unit 13) via the communication device 110 as the first keyword of the search request.
When receiving the search result via the communication device 110, the control device 100 displays the search result on the display device 120 as an answer to the inquiry and notifies the user.

According to the fourth embodiment, the user's load is reduced because the keyword is automatically generated from the sensor information as compared with the case where the user determines the state of the device and creates a keyword for search. .
Furthermore, by using a plurality of sensor information, the amount of information for grasping the state of the device is increased, and a keyword expressing the state more accurately is generated. As a result, the search device 1 can provide useful search results in a timely manner.

  As mentioned above, although embodiment of this invention was described, this invention is not restricted to embodiment mentioned above. Further, the effects described in the present embodiment are merely a list of the most preferable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the present embodiment.

  The search method by the search device 1 is realized by software. When realized by software, a program constituting the software is installed in a computer (search apparatus 1). Further, these programs may be recorded on a removable medium and distributed to the user, or may be distributed by being downloaded to the user's computer via a network.

DESCRIPTION OF SYMBOLS 1 Search apparatus 10 Control part 11 Classification part 12 Extraction part 13 Reception part 14 Search part 15 Group management part 16 Sensor input part 17 Status acquisition part 20 Storage part 21 Document database 22 Related word file 23 Conversion database

Claims (16)

A classification unit that classifies the document data based on the degree of match of the first item in which a combination of one or more symbol strings is described in a database in which document data including a plurality of items is accumulated;
A search device comprising: an extraction unit that extracts a keyword group included in a second item as a related word for search from document data in the same group classified by the classification unit.   The search device according to claim 1, wherein the classification unit classifies the document data in which the first item is completely matched into the same group.   The search device according to claim 2, wherein the classification unit classifies document data in which at least a part of the first item matches into the same group. A reception unit for receiving a search request based on the first keyword;
A search unit that searches the document data by the first keyword and searches the document data by a second keyword included in the same keyword group as the first keyword. 4. The search device according to any one of 3.   The search device according to claim 4, wherein the reception unit displays a keyword group including the first keyword and receives an input of the second keyword.   The search device according to claim 4, wherein the search unit sequentially extracts the second keyword from a keyword group including the first keyword, and searches the document data by each of the second keyword. At least two different databases are managed as one set of database groups, and one set of keyword groups extracted as related words for search by the extraction unit for each database included in the one set of database groups. A group management unit that manages as a keyword group set of
The search unit searches for document data stored in the database included in the set of databases in response to a search request based on the first keyword, and in the set of keyword groups, The search device according to claim 4, wherein the document data is searched by a second keyword included in the same keyword group as the one keyword.   The reception unit receives the one or more sentences as the search request, and then extracts the first keyword from the one or more sentences by a procedure common to the extraction unit extracting the keyword group. The search device according to any one of claims 4 to 7. A sensor input unit for inputting sensor information of the device;
A state acquisition unit that converts the sensor information into a state keyword indicating the state of the device based on a predetermined conversion condition;
The search device according to claim 4, wherein the receiving unit receives the state keyword as the first keyword. The document data is a device failure report,
The search apparatus according to any one of claims 1 to 9, wherein the symbol string is a specification number and the number of parts to be repaired. A classification step of classifying the document data based on the degree of match of the first item in which a combination of one or more symbol strings is described in a database in which document data including a plurality of items is accumulated;
A search method in which a computer executes an extraction step of extracting a keyword group included in a second item from document data in the same group classified in the classification step. The computer further comprises:
An accepting step for accepting a search request by the first keyword;
The search step for searching the document data by the first keyword and searching the document data by a second keyword included in the same keyword group as the first keyword is executed. Search method. The computer further comprises:
At least two or more different databases are managed as a set of database groups, and for each database included in the set of database groups, one set of keyword groups extracted as related words for search in the extraction step Execute a group management step to manage as a keyword group set of
In the search step, in response to a search request based on the first keyword, the document data stored in the database included in the set of database groups is searched, and in the set of keyword groups, the first data The search method according to claim 12, wherein the document data is searched by a second keyword included in the same keyword group as the one keyword. The computer further comprises:
In the receiving step, after receiving one or more sentences as the search request, the first keyword is extracted from the one or more sentences by a procedure common to the procedure for extracting the keyword group in the extracting step. The search method according to claim 12 or claim 13. The computer further comprises:
A sensor input step for inputting sensor information of the device;
Performing a state acquisition step of converting the sensor information into a state keyword indicating the state of the device based on a predetermined conversion condition;
The search method according to claim 12 or 13, wherein in the receiving step, the state keyword is received as the first keyword.   The search program for functioning a computer as a search device in any one of Claims 1-10.

Images