JP5533692B2 - Knowledge information extraction apparatus, method and program in manufacturing process - Google Patents

Description

  The present invention relates to a knowledge information extraction apparatus, method, and program in a manufacturing process for extracting knowledge information using a daily operation log.

  At manufacturing sites such as steelworks, an operation diary is created based on events that occur during operation. Using this operation diary, it is expected that operators who are not particularly skilled workers search for similar situations in the past to improve operational efficiency for improving the operation process.

JP 2007-272400 A JP 2010-108478 A JP 2007-226532 A

  However, at present, the past operation diaries are merely accumulated, and it is not easy to search the operation diaries and access the past operation diaries useful for examining the improvement of the operation process. By managing the operation log in a lump and providing an interface that allows easy access to useful operation diaries, operators can search past similar situations and improve operational efficiency to improve the operation process. .

  There are many facilities at manufacturing sites such as steelworks, and the state of the facilities and the items handled by the facilities are diverse, and they are related to each other. For example, there is a relationship that “heating device” is a subordinate concept of “apparatus” and “plasma heating device” is further subordinate concept thereof, a relationship that “a state (for example, temperature)” is an attribute of “apparatus”, etc. It is. Therefore, many concepts are used in the operation diary, and it is necessary to appropriately handle the concept system when the operation diary is a search target.

  As an operation support based on past knowledge, in Patent Document 1, when an operation result is lowered, an item that is considered to be a factor of decrease is searched from the phenomenon database, and the factor of decrease is narrowed down based on past results Is disclosed. However, in Patent Document 1, it is necessary to create in advance a phenomenon database in which information about the operational performance lowering factor definition is stored, and knowledge information is extracted using an operational diary created on a daily basis. Not what you want.

  Moreover, in patent document 2, utilizing the report regarding the past trouble is disclosed. That is, the trouble report including the title and the cause of the trouble is subjected to a morphological element analysis of the title, and a classification word for classifying the trouble name that has occurred is extracted. Then, morphological element analysis is performed on the cause of the trouble recorded in the trouble report in which the extracted classification word is included in the title, and a cause candidate word that is a candidate of the cause of the trouble is extracted. However, Patent Document 2 is a technique based on the premise that a trouble has occurred, and does not extract knowledge information using an operation diary that is created regardless of the presence or absence of the trouble. Moreover, the classification words and the cause candidate words are simply extracted based on the frequency of occurrence of these terms in the trouble report, and the related concepts are not taken into consideration.

  Further, in Patent Document 3, a defect database composed of facility data and predicate data, a cause database composed of target data and predicate data, and a maintenance database composed of element data and predicate data are associated with ID numbers in a tree shape. To construct an ontology data structure to be generated. However, in Patent Document 3, when there is a specific problem, the cause and the maintenance measures are judged against the tree, and when searching for a concept from a large number of past operation diaries, There is no disclosure of considering related concepts.

  The present invention has been made in view of the above points, and aims to manage operation diaries collectively and make it easy to access useful operation diaries, and more specifically, a large number of past operation diaries. The purpose is to make it possible to consider related concepts when searching for a certain concept.

The knowledge information extraction device in the manufacturing process of the present invention stores operation diary data having a result data column in which the value of the result data is described and a special note column in which the operator can describe individual specific items. A knowledge information extraction apparatus in a manufacturing process for extracting knowledge information using a database, and when a concept defined in a previously constructed ontology is input, the concept is stored in the database based on the input concept A search means for performing a search on the special mention column of the operation diary that has been entered and displaying a list of operation diaries including the input concept in the special notice column on a display device, wherein the search means includes the inputted concept displayable der on the display device a list of logbook be included logbook including special mention column defined concepts to a predetermined relationship with respect to is, the search unit, the actual When the data name is input, only the operation diary including the input concept in the special note column, or the operation diary including the input concept or the concept defined as having a predetermined relationship in the special remarks as a target only, generating a graph of the performance data corresponding to the actual data name that is the input, and wherein the you to view.
Further, another feature of the knowledge information extracting apparatus in the manufacturing process of the present invention is that the search means displays all or a part of the concepts defined as related to the inputted concept. The point is to display on the device.
Further, another feature of the knowledge information extracting apparatus in the manufacturing process of the present invention is that the search means has a concept defined as being related to the input concept or the input concept. A related keyword extracting means for extracting related keywords that are related to all or a part from the special note column and displaying them on the display device is provided.
Another feature of the knowledge information extracting apparatus in the manufacturing process of the present invention is that a narrowing condition can be input to the actual data corresponding to the input actual data name, and the narrowing is performed based on the narrowing condition. However, the search is performed on the operation diary including the actual data in the actual data column.
The knowledge information extraction method in the manufacturing process of the present invention accumulates operation diary data having a result data column in which the value of the result data is described, and a special note column in which the operator can describe individual specific items. A knowledge information extraction method in a manufacturing process for extracting knowledge information by a knowledge information extraction device using a database, wherein the knowledge information extraction device receives the concept defined in the ontology constructed in advance, based on the input concepts, a search step of performing a search on the subject contains legal field of accumulated logbook in the database, and displays a list of logbook including concepts the input special mention column display device performed, the search step, logbook including concepts defined a predetermined relationship with respect to the input concepts Special field including in the logbook List Ri displayable der on the display device, the search step, when actual data name is entered, only the logbook including concepts the input to otherwise column, or the entered concepts or contrast the concepts defined and predetermined relationship Te as a target only logbook including Notices, generating a graph of the performance data corresponding to the actual data name that is the input, and wherein the you to view.
The program of the present invention provides knowledge using a database in which data of operation diaries having an actual data column in which values of actual data are described and a special column in which an operator can describe individual specific items is stored. A program for extracting information, and when a concept defined in an ontology constructed in advance is input, based on the input concept, the special column of the operation log stored in the database is targeted. A search is executed, and a search process for displaying a list of operation diaries including the input concept in a special note column on a display device is executed by a computer, and the search process has a predetermined relationship with the input concept. Ri displayable der on the display device also a list of logbook including logbook including concepts defined in the Special column there, the search process, when the actual data name is input, pre Only the operation diary including the input concept in the special column or the operation diary including only the operation diary including the input concept or the concept defined as having a predetermined relationship with the input concept. generating a graph of the performance data corresponding to the actual data name, wherein you to view.

  According to the present invention, when searching for a certain concept from a large number of past operation diaries, related concepts can also be taken into account, so that the concept system when the operation diary is a search target can be handled appropriately. Can do. In this way, it is possible to manage the operation diaries collectively and provide an interface that allows easy access to useful operation diaries, allowing operators to search for past similar situations and improve operational efficiency for improving operational processes. .

It is a figure which shows the function structure of the knowledge information extraction apparatus in the manufacturing process which concerns on embodiment. It is a figure which shows the example of an operation log. It is a figure which shows a keyword search screen. It is a figure for demonstrating a Dice coefficient. It is a figure which shows an ontology construction screen. It is a figure which shows an ontology search screen. It is a figure which shows a data search screen. It is a figure which shows a data search screen. It is a figure which shows a data search screen. It is a figure which shows a data search screen. It is a figure which shows a data search screen.

Preferred embodiments of the present invention will be described below with reference to the accompanying drawings.
FIG. 1 shows a functional configuration of the knowledge information extraction device 100 in the manufacturing process according to the present embodiment. In this embodiment, the steelmaking process is targeted. Reference numeral 101 denotes a database in which electronic data of operation diaries (operation / equipment / quality diary) is accumulated. FIG. 2 shows an example of the operation diary. The operation diary is created on a computer by document creation software or the like and stored in the database 101. In the manufacturing site that operates 24 hours, such as steelworks, etc., it is divided into number 8, number 2 and number 8 every 8 hours. Indicates.

  In the operation diary, a diary date 201, a last data update date 202, and a last print date 203 are described. In addition, production No. A production number, process, production symbol, code, production width, size, etc. are described for each (204), and the actual data column 205 in which values such as start time, production time, operation speed, etc., which are actual data, are described. Is provided. The value in the result data column 205 may be automatically input from the facility, or may be input by an operator. In addition, the production plan and the actual result 206 are described. Further, special notes columns 207 for operation special notes, facility special notes, and quality special notes are provided. In these special notes columns 207, an operator can describe items such as individual specific findings (special notes).

  Returning to FIG. 1, reference numeral 102 denotes an input device such as a keyboard and a mouse. Reference numeral 103 denotes a display device that displays various screens.

  Reference numeral 104 denotes a keyword search unit. When an arbitrary keyword is input via the input device 102, a search is performed on the special entry column 207 of the operation log based on the input keyword, and the search result is displayed. Display on the device 103.

  FIG. 3 shows a keyword search screen displayed on the display device 103. In the keyword search screen, a keyword input field 301 and a period input field 302 are provided. When a keyword is entered in the keyword input field 301 and the search button 303 is pressed, keyword search is executed. That is, the keyword search unit 104 searches the special entry column 207 of the operation diary stored in the database 101 for the keyword (referred to as input keyword) input in the keyword input column 301, and the operation diary including the input keyword in the special entry column 207. List 304 is displayed. When a period is input in the period input field 302, an operation diary with a date (diary date 201) included in the period is set as a search target.

  In the operation diary list 304, for example, special items described in the date, group, writer, and special note column 207 are extracted and displayed so that the outline of each operation diary can be understood together with the number of hits. In this case, the input keyword in the special note may be highlighted such as changing the color. The operation diary list 304 is linked to the operation diary data, and when a date is selected, details of the selected operation diary are displayed.

  In addition, the co-occurrence word extraction unit 104a extracts and displays the co-occurrence word 305 that co-occurs with the input keyword from the special entry column 207 of the operation diary stored in the database 101. Specifically, the morphological element analysis of the description in the special note column 207 of the operation log is performed, and the part of speech is extracted. Then, using one description in the special note column 207 as a unit, terms appearing simultaneously with the input keyword are treated as co-occurrence. For the determination of the co-occurrence degree, for example, a Dice coefficient as shown in FIG. 4 is used. When displaying the co-occurrence words 305, they are listed in descending order of co-occurrence.

  When any one of the co-occurrence words 305 is designated, the keyword search unit 104 displays the designated co-occurrence word in the keyword input field 301 and performs a search again as an input keyword. When (+) displayed next to each co-occurrence word 305 is designated, the keyword search unit 104 performs an AND search between the input keyword and the designated co-occurrence word. Although not shown here, a logical search may be executed by a combination of OR search, AND search, OR search, and NOT search.

  Using the keyword search as described above, for example, when the keyword “clogged” is input in the keyword input field 301, co-occurrence words co-occurring with “clogged” are displayed. This makes it possible to know terms that frequently appear when “clogging” appears in the special note column 207, and even if you are not an expert in the process, you can estimate what is likely to be clogged. Become.

  Returning to FIG. 1, reference numeral 105 denotes an ontology search unit. When a concept defined in a previously constructed ontology is input via the input device 102, an operation diary is displayed based on the input concept. A search is executed for the special note column 207 and the search result is displayed on the display device 103.

  Here, an ontology is a systemized dictionary that provides a common vocabulary (concept), so that the concepts that make up the target world (in this embodiment, the steelmaking process) are extracted so that the target world can be expressed well. It describes the relationship between concepts. Examples of relationship notation methods in the ontology include an is-a relationship, a part-of relationship, an attribute-of relationship, a relationship, an instance, and a role. The is-a relationship represents a superordinate concept-subordinate concept. For example, there is a relationship that “heating device” is a subordinate concept of “apparatus” and “plasma heating device” is a subordinate concept thereof. The part-of relationship represents an overall concept and a partial concept, and represents that an object is composed of a plurality of parts (parts). For example, the relationship “burner” is a part constituting a “heating device”. The attribute-of relationship represents an attribute of a concept. For example, there is a relationship that “a certain state (for example, temperature)” is an attribute of “device”. The relationship represents a relationship between two concepts, for example, “bicycle” is the property of “human”. An instance represents a concrete object of a concept. A role represents a role associated with a certain concept. For example, “hose” has a relationship of “passing through”.

  The ontology is built in advance on the knowledge information extraction apparatus 100. FIG. 5 shows an example of an ontology construction screen for constructing an ontology. Here, “immersion nozzle”, “sliding nozzle”, “long nozzle”, “electrode nozzle”, “spray nozzle” and the like are defined as subordinate concepts of “nozzle” (is-a relationship). Further, “ejection port” is defined as a part constituting “nozzle” (part-of relationship). Further, “flow rate attribute value” is defined as the attribute of “nozzle” (attribute-of relationship). In addition, “ejection function” is defined as the role of “nozzle” (roll). In each concept, an alias (notation fluctuation) can be described. For example, with regard to “nozzle”, it is defined that the English notation “nozzle” is treated as the same thing.

  FIG. 6 shows an ontology search screen displayed on the display device 103. In the ontology search screen, a concept input field 601 and a period input field 602 are provided. In addition, a check box 603 is provided for selecting whether to perform an extended search including sibling concepts (concepts having the same superordinate concept) with respect to the concept input in the concept input field 601 or to perform an extended search including a subordinate concept. It has been. When a concept is input in the concept input field 601 and the search button 604 is pressed, an ontology search is executed. That is, the ontology search unit 105 displays the concept 605 defined in the ontology as related to the concept (referred to as an input concept) input in the concept input field 601. In addition, the ontology search unit 105 searches the operation diary special entry column 207 stored in the database 101 for the input concept, and displays the operation diary list 606 including the input concept in the special entry column 207. When the check box 603 is checked, the sibling concept and the subordinate concept are also searched from the special diary column 207 of the operation diary stored in the database 101, and the operation including the operation diary including the sibling concept and the subordinate concept in the special note column 207 is also performed. A diary list 606 is displayed. Here, an example has been described in which an extended search including subordinate concepts and sibling concepts can be performed, but check boxes for concepts having other relationships may be provided. When a period is input in the period input field 602, an operation diary with a date (diary date 201) included in the period is set as a search target.

  In the illustrated example, a concept path, a superordinate concept, a sibling concept, and a subordinate concept are displayed as the concept 605 related to the input concept. In the concept path, the superordinate concept up to the input concept is displayed. Here, the superordinate concept up to the input concept “crane” is displayed, such as machinery / equipment / equipment / transport / transportation device / crane. doing. In the superordinate concept, a superordinate concept that is higher than the input concept is displayed. Here, “conveyance / transportation device” that is a superordinate concept of “crane” that is the input concept is displayed. In the sibling concept, the sibling concept of the input concept is displayed, and here, the sibling concepts of “crane” as the input concept, such as “cart”, “conveyor”, and the like are displayed. In the subordinate concept, subordinate concepts below the input concept are displayed. Here, “conveying crane”, “maintenance crane”, etc., which are subordinate concepts of “crane” as the input concept are displayed. It should be noted that although a single higher level concept or lower level concept is displayed as the higher level concept or lower level concept, a plurality of higher level concepts or lower level concepts may be displayed.

  Here, as the concept 605 related to the input concept, only the concept in the is-a relationship representing the superordinate concept-subordinate concept is displayed, but other relationships in the ontology, that is, the part-of relationship and the attribute-of relationship are displayed. If relationships between relationships, instances, and roles are defined, these concepts are also displayed.

  When a concept 605 related to any input concept is specified, the ontology search unit 105 displays the specified concept in the concept input field 601 and performs a search again as an input concept. When (+) displayed next to a concept displayed as a sibling concept or a subordinate concept among the concepts 605 related to the input concept is designated, the ontology search unit 105 specifies the input concept and the designated concept. Perform an OR search with sibling concepts and subordinate concepts. Although not shown here, an AND search, or a logical search by a combination of AND search, OR search, and NOT search may be executed.

  In the operation diary list 606, for example, special items described in the date, group, writer, and special note column 207 are extracted and displayed so that an outline of each operation diary can be understood together with the number of hits. In this case, the input concept (including the sibling concept and the subordinate concept if the check box 603 is checked) in the special note may be highlighted such as by changing the color. The operation diary list 606 is linked to operation diary data, and when a date is selected, details of the selected operation diary are displayed.

  By using ontology search as described above, when searching for a certain concept, related concepts are also extracted from a large number of past operation diaries, so the concept system for searching operation diaries can be Can be handled appropriately. For example, “crane” is a kind of “conveying / conveying device” and other types of “conveying / conveying device” include “cart”, “conveyor”, etc. It is possible to easily identify the relationship.

  Returning to FIG. 1, reference numeral 106 denotes a data search unit. When a desired result data name is input via the input device 102, an operation data record data column of the operation log is input based on the input result data name. A search is executed for 205, and a graph of actual data corresponding to the input actual data name is generated and displayed on the display device 103.

  7 to 10 show data search screens displayed on the display device 103. FIG. 7 to 9 are data search screens for displaying a histogram representing the range distribution of the result data. In the data search screen, an actual data name input field 701, a value range input field 702, a keyword input field 703, and a period input field 704 are provided. When a result data name is selected and input from the pull-down menu in the result data name input field 701 and a search button 705 is pressed, data search is executed. That is, the data search unit 106 retrieves the performance data corresponding to the performance data name (referred to as the input performance data name) input in the performance data name input field 701 from the performance data field 205 of the operation log accumulated in the database 101. A search is performed, and a histogram 706 representing the range distribution is generated and displayed. In the illustrated example, “operation speed” is entered in the result data name input field 701, and in the histogram 706, the horizontal axis is the value of actual data (here, the operation speed), and the vertical axis is the number of cases. Further, the data search unit 106 displays a list 707 of operation diaries in which the result data corresponding to the input result data name is included in the result data column 205. When a period is input in the period input field 704, an operation diary with a date (diary date 201) included in the period is set as a search target.

  As shown in FIG. 8, when a value range is input to the value range input field 702, the histogram 706 is generated and displayed by narrowing down to the actual data of the input value range (referred to as input value range). The value range on the horizontal axis of the histogram 706 shown in FIG. 8 is narrower than that of the histogram 706 shown in FIG. The operation diary list 707 displays operation diaries corresponding to the input result data name and including the result data in the input value range in the result data column 205.

  Further, a range of a desired range can be designated by dragging the mouse on the currently displayed histogram 706 without using the range input field 702. When a desired value range is designated in the histogram 706 (referred to as an input value range) and the search button 705 is pressed, the data search unit 106 narrows down to the actual data in the input value range and regenerates and displays the histogram 706. To do. The operation diary list 707 displays operation diaries corresponding to the input result data name and including the result data in the input value range in the result data column 205.

  As shown in FIG. 9, when a keyword is input in the keyword input field 703, only the operation diary including the input keyword in the special note field 207 is used as the target of the actual data corresponding to the input actual data name. A histogram 706 is generated and displayed, and an operation diary list 707 is displayed.

  Here, as shown in FIG. 8, when the narrowing is performed by the range, the related keyword extraction unit 106 a is related keywords 708 that are considered to be related to the operation log including the narrowed down result data in the result data column 205. Is extracted and displayed. Similarly, when a keyword is input as shown in FIG. 9, the keyword is included in the special note column 207, and the operation log including the record data corresponding to the input record data name is included in the record data column 205. A related keyword 708 that is assumed to be present is extracted and displayed.

  For the extraction of the related keyword 708, the idea of the Dice coefficient shown in FIG. 4 is used. That is, it is assumed that the number of operation diaries satisfying the narrowing down by range or keyword is “A”, and the number of operation diaries not satisfying is “¬A”. For each related keyword candidate in the related keyword candidate group, the number of operation diaries including the related keyword candidate is “B”, and the number of operation diaries not including the related keyword candidate is “¬B”. Then, the degree of association is obtained as degree of association = 2I / {(I + II) + (I + III)}, and the related keyword candidates having the highest degree of association are listed as the related keywords 708 in order. The related keyword candidate group may be a part of speech extracted by performing morphological element analysis described in the special entry column 207 of the operation diary, a concept defined by an ontology, or the like.

  When any related keyword 708 is designated, the data search unit 106 displays the designated related keyword in the keyword input field 703 and executes the search again. In other words, a histogram 706 of performance data corresponding to the input performance data name is generated and displayed for only the operation diary including the keyword in the special note column 207, and a list of operation diaries 707 is displayed.

  In the operation diary list 707, for example, special items described in the date, group, writer, and special note column 207 are extracted and displayed so that the outline of each operation diary can be understood together with the number of hits. In this case, the keywords and related keywords in the special notes may be highlighted such as by changing the color. The operation diary list 707 is linked to the operation diary data, and when a date is selected, details of the selected operation diary are displayed.

  FIG. 10 is a data search screen that displays a line graph representing the time transition of the performance data. Although it is basically the same as that described with reference to FIGS. 7 to 9, the actual data corresponding to the actual data name input in the actual data name input field 701 is used as the actual data of the operation log accumulated in the database 101. A search is made from the column 205, and a line graph 709 representing the time transition is generated and displayed. In the illustrated example, “operation speed” is input in the actual data name input field 701, and in the line graph 709, the horizontal axis is time, and the vertical axis is the value of actual data (the operation speed here). In the initial display, the time transition of the average value for each month is displayed, but by narrowing down the period, the total range of the average value becomes finer like month → week → day... And details can be confirmed.

  Even when the line graph 709 is displayed, for example, a desired period can be specified by dragging the mouse on the currently displayed line graph 709. Thereby, the line graph 709 is generated and displayed by narrowing down to the actual data of the input period (referred to as the input period). The operation diary list 707 displays operation diaries corresponding to the input result data name and including the result data in the input period in the result data column 205.

  If the data search as described above is used, the actual data corresponding to the actual data name input in the actual data name input field 701 is visualized, that is, the histogram 706 indicating the range distribution of the actual data or the time transition. Since the line graph 709 is displayed, the operator can easily grasp the performance data.

  In addition, since it is possible to perform a combined search including keywords, it is easy to grasp the relationship between the performance data and the keywords. Moreover, since the related keyword 708 is displayed by inputting a keyword or narrowing down a range or a period, it is easy to grasp the relationship between the result data and the keyword. As a result, for example, it is possible to grasp a keyword appearing in a range or period having an outlier and estimate what caused the outlier due to the cause. Alternatively, for example, when a plurality of peaks appear in the histogram 706 or the line graph 709, it is possible to grasp the keywords appearing in the respective peaks and estimate the factors causing the peaks.

  When narrowing down sequentially in a series of operations, if the history of the histogram 706 (or line graph 709) remains in the screen, how the graph changes depending on the keyword, range, and time (ie, , How the value of the result data changes).

  In the present embodiment, an example in which a compound search including a keyword can be performed in a data search has been described, but a compound search including an ontology may be performed. FIG. 11 shows a data search screen displayed on the display device 103. In place of the keyword input field 703 in FIGS. 7 to 9, whether to perform an extended search including a concept input field 710 and sibling concepts for the concept input to the concept input field 710 or not, includes an extended search. A check box 711 for selecting whether or not is provided.

  When keywords and concepts are input, co-occurrence words and related concepts may be displayed as in the case of keyword search and ontology search described above. In FIG. 11, similar to FIGS. 7 to 9, a record data name input field 701, a value range input field 702, a keyword input field 703, a period input field 704, a search button 705, a histogram 706, and a list of operation logbooks. 707 is provided, and as described in paragraphs 0033 to 0039, it is possible to narrow down by the range of the performance data, create a histogram, extract related keywords, and list the searched operation diaries.

  Further, instead of the histogram 706 of FIG. 11, a line graph 709 representing the time transition as shown in FIG. 10 may be displayed to enable narrowing down by period as described in paragraphs 0040 to 0041. .

  As mentioned above, although this invention was demonstrated with various embodiment, this invention is not limited only to these embodiment, A change etc. are possible within the scope of the present invention. For example, in FIG. 1, the apparatus is configured to include one apparatus, but may be configured to include a plurality of apparatuses. For example, a configuration may be adopted in which a plurality of terminals are connected to the server, the server performs various searches in response to requests from the terminals, and the terminals receive the search results.

  The knowledge information extraction apparatus to which the present invention is applied is specifically configured by a computer apparatus or computer system including a CPU, a RAM, a ROM, and the like. Therefore, the present invention includes a program itself installed in a computer apparatus or a computer system in order to realize each function processing of the present invention.

  100: Knowledge information extraction device, 101: Database, 102: Input device, 103: Display device, 104: Keyword search unit, 104a: Co-occurrence word extraction unit, 105: Ontology search unit, 106: Data search unit, 106a: Related Keyword extractor

Claims (6)

A manufacturing process that extracts knowledge information using a database in which data of operation diaries having an actual data column in which the values of actual data are described and a special item column in which an operator can describe individual specific items is stored. A knowledge information extraction device in
When the concept defined in the ontology built in advance is input, a search is performed on the special column of the operation log stored in the database based on the input concept, and the input concept is A search means for displaying a list of operation diaries included in the special note column on the display device,
Said retrieval means, Ri displayable der on the display device logbook also include a list of logbook including concepts defined a predetermined relationship with respect to the input concepts otherwise column,
When the result data name is input, the search means includes only the operation log including the input concept in the special note column, or the input concept or a concept defined as having a predetermined relationship with the input concept. targeting only logbook including Notices, generating a graph of the performance data corresponding to the actual data name that is the input knowledge information extracting device in the manufacturing process, wherein you to view.   2. The knowledge information extraction in the manufacturing process according to claim 1, wherein the search means displays all or a part of the concepts defined as having a relationship with the input concept on the display device. apparatus.   The search means extracts, from the special note column, related keywords that are related to the input concept or all or a part of the concept defined to be related to the input concept. The knowledge information extracting device in the manufacturing process according to claim 1, further comprising related keyword extracting means for displaying on the display device. It is possible to input a narrowing condition for the actual data corresponding to the input actual data name, and to perform a search for the operation log including the actual data narrowed down by the narrowing condition in the actual data column. The knowledge information extraction device in the manufacturing process according to claim 1, wherein the knowledge information extraction device is a manufacturing process. Knowledge information is extracted by a knowledge information extraction device using a database in which data of operation diaries having an actual data column in which the value of actual data is described and a special item column in which an operator can describe individual specific items is stored. A knowledge information extraction method in a manufacturing process for extracting
When the concept defined in the ontology constructed in advance is input, the knowledge information extraction device executes a search for the special entry column of the operation diary stored in the database based on the input concept. , Performing a search step of displaying a list of operation diaries including the input concept in a special note column on a display device ;
The search step is Ri displayable der on the display device logbook also include a list of logbook including concepts defined a predetermined relationship with respect to the input concepts otherwise column,
In the search step, when an actual data name is input, only the operation diary including the input concept in the special note column, or the input concept or a concept defined as having a predetermined relationship with the input concept. targeting only logbook including Notices, generating a graph of the performance data corresponding to the actual data name that is the input knowledge information extraction method in the manufacturing process which features you to view. For extracting knowledge information using a database in which data of operation diary having an actual data column in which the value of actual data is described and a special note column in which an operator can describe individual specific items is stored A program,
When the concept defined in the ontology built in advance is input, a search is performed on the special column of the operation log stored in the database based on the input concept, and the input concept is Let the computer execute a search process to display a list of operation diaries included in the special note column on the display device,
The search process, Ri displayable der a list of logbook the concepts defined a predetermined relationship with respect to the input concepts, including logbook including special mention column on said display device,
In the search process, when an actual data name is input, only the operation diary including the input concept in the special note column, or the input concept or a concept defined as having a predetermined relationship with the input concept. targeting only logbook including Notices, generating a graph of the performance data corresponding to the actual data name that is the input, program characterized you to view.

Images