JP2019105953A - Data analysis system and data analysis method - Google Patents

Abstract

To support selection of optimum processing of data analysis and combination of data.SOLUTION: A data analysis system stores data acquired for an analysis target and actual event data related to an event actually occurred for the analysis target; generates one or more first methods (i.e., programs and parameters for realizing pre-processing) that is first processing related to acquired data analysis, and one or more second methods (i.e., programs and parameters for realizing analysis processing) that is second processing related to acquired data analysis; generates one or more combinations of a series of analysis target data (i.e., a series of data) to execute a combination of the acquired data; generates analysis result data including a combination of each execution result; and collates the analysis result data with the actual event data to generate analysis performance data indicating each analysis performance of the combination.SELECTED DRAWING: Figure 10

Description

  The present invention relates to a data analysis system and a data analysis method.

  In Patent Document 1, "when performing data mining using a commercially available data mining tool, the purpose is to simplify the operation of the part that performs modeling and residual analysis", "for the commercially available data mining tool From the input data, judge the applicable prediction model for correlation finding, and apply an arbitrary parameter set from a plurality of parameter sets prepared in advance to each prediction model to calculate a prediction result, and the prediction result The residual is calculated from the above to determine the optimal correlation finding model and perform data mining. "

Japanese Patent Application Publication No. 2006-235848

  Recently, the introduction of IoT (Internet of Things) has been promoted in various fields such as management of manufacturing sites and inventory management of warehouses and stores. For example, at a manufacturing site, data is collected from sensors installed in devices and people, and the collected data is analyzed to grasp the operating status of the devices and lead wires of people and goods, thereby achieving overall operations and tasks. Attempts have been made to improve efficiency.

  There are various methods for analyzing the data collected here, and it is necessary to select an appropriate analysis method in accordance with the object of analysis and the purpose of analysis. Even when the same analysis method is used, analysis performance varies depending on the data preprocessing method and the parameter value to be set. Therefore, when newly constructing a data analysis system, it is necessary for the expert or the like to repeat trials to select a pretreatment method or analysis method, and to set parameters appropriately. In addition, in patent document 1, since a residual is used for evaluation of a model, when the objective variable which can be actually measured does not exist, it can not apply.

  The present invention has been made in view of such background, and it is an object of the present invention to provide a data analysis system and data analysis method that can support selection of an optimal processing method of data analysis.

  One of the present inventions for solving the above-mentioned subject is a data analysis system, which is acquired data which is data acquired about an analysis target, and an actual event which is data about an event which actually occurred about the analysis target. A storage unit for storing data, one or more first methods as a first process related to analysis of the acquired data, and one or more second processes as a second process related to analysis of the acquired data A first processing unit that generates one or more combinations of the methods of the above, and a second processing unit that executes the combination on the acquired data and generates analysis result data that is data including each execution result of the combination And a third processing unit that generates analysis performance data that is data indicating the analysis performance of each of the combinations by collating the analysis result data with the actual event data. .

  In addition, the subject which this application discloses, and its solution method are clarified by the column of the form for inventing, and drawing.

  According to the present invention, selection of an optimal processing method of data analysis can be supported.

It is a figure showing the rough composition of a data analysis system. It is a figure showing an example of the information processor which can be used as a component of a data analysis system. It is a figure which shows the main functions with which a data pre-processing apparatus is equipped, and the main data which a data pre-processing apparatus manages. It is a figure which shows the main functions with which a data management apparatus is equipped, and the main data which a data management apparatus manages. It is an example of analysis object data. It is an example of real event data. It is an example of analysis result data. It is an example of analysis performance calculation result data. It is a figure which shows an example of recommendation combination registration data. It is a figure which shows the main functions with which an analyzer is equipped, and the main data which an analyzer manages. It is an example of pre-processing method data. It is an example of analysis method data. It is an example of threshold pattern data. It is an example of selection combination registration data. It is an example of series data attribute data. It is an example of series group registration data. It is an example of weighting data. It is an example of initialization data. It is a sequence diagram explaining an initial setting process. It is an example of an initial setting screen. It is an example of an initial setting screen. It is an example of a series data attribute input screen. It is an example of a series group specification screen. It is a sequence diagram explaining analysis processing. It is a sequence diagram explaining analysis performance calculation & review processing. It is an example of a weighting setting screen. It is a flow chart explaining analysis performance calculation processing. It is an example of a combination review screen. It is an example of a selection method designation screen. It is an example of weight record data. It is an example of an initial setting screen. It is a flow chart explaining analysis performance calculation processing. It is an example of a combination review screen. It is a sequence diagram explaining an initial setting process.

  Hereinafter, embodiments will be described with reference to the drawings. In the following description, the same or similar configurations may be denoted by the same reference numerals and duplicate descriptions may be omitted.

First Embodiment
FIG. 1 shows a schematic configuration of a data analysis system 1 described as the first embodiment. As shown in the figure, the data analysis system 1 includes various sensors 3, a data preprocessing device 101, a data management device 102, an analysis device 103, and an input / output device 105.

  As shown in the figure, the data preprocessing apparatus 101, the data management apparatus 102, the analysis apparatus 103, the input / output apparatus 105, the analysis target 2, and the various sensors 3 are communicably connected via the communication network 5. . The communication network 5 is configured by, for example, a wired or wireless communication infrastructure such as a LAN (Local Area Network), a WAN (Wide Area Network), the Internet, an intranet, a dedicated line, a cellular phone network, an optical fiber, and the like.

  The analysis target 2 is, for example, a machine tool or a manufacturing apparatus (for example, a press machine, an NC machine tool, a robot arm or the like) provided in a product manufacturing plant or the like. The analysis target 2 may not necessarily be connected to the communication network 5.

  The various sensors 3 are installed, for example, around the analysis target 2 or the analysis target 2. Although the type of the various sensors 3 is not necessarily limited, for example, a current sensor, a vibration sensor, a temperature sensor, a humidity sensor, a position detection sensor, a speed sensor, a noise sensor, and the like. The data management device 102 acquires, for example, data (including signals) output from the various sensors 3 via the communication network 5. Hereinafter, the data (that is, sensor data) acquired by the data management apparatus 102 is referred to as acquired data.

  The data preprocessing apparatus 101 performs preprocessing (that is, the first process) necessary for the analysis apparatus 103 to perform a process related to data analysis (that is, a second process, hereinafter referred to as an analysis process) on acquired data. )I do.

  The data management apparatus 102 stores data (hereinafter, referred to as analysis target data) generated by the pre-processing. The data management device 102 also stores various data used in various processes performed in the data analysis system 1.

  The analysis device 103 performs, for example, analysis processing on the analysis target data to detect a sign of occurrence of an abnormality of the analysis target 2 (hereinafter, referred to as abnormality sign detection). Specifically, the analysis apparatus 103 is one or more methods (ie, a first method, hereinafter referred to as a pretreatment method) that is a pretreatment, and one or more methods (that is, a first method) that is an analysis process. Method 2. A combination of one or more combinations (hereinafter referred to as analysis method) (hereinafter referred to as method combinations) and a series group described later is generated. The analysis device 103 executes the generated combinations described above for the acquired data, and generates analysis result data which is data including the execution results of the combinations. The methods in the pre-processing and analysis processing indicate a program (that is, an algorithm) and parameters used in the program.

  In addition, the analysis device 103 is analysis result data which is data including the result of analysis processing, and data regarding an event (for example, failure, maintenance, etc .; hereinafter referred to as an actual event) actually occurring for the analysis target 2. By collating with the actual event data, analysis performance data is generated which is data indicating the analysis performance of each of the above combinations. Further, the analysis apparatus 103 specifies and outputs a combination having high analysis performance based on the analysis performance data, and performs evaluation of the combination, improvement proposal of analysis processing, and the like.

  The input / output device 105 provides a user interface for the user to access the data preprocessing device 101, the data management device 102, and the analysis device 103. The user can input data and perform various settings on the data preprocessing device 101, the data management device 102, and the analysis device 103 via the input / output device 105. Also, the user can refer to or acquire information (that is, data) output from the data preprocessing device 101, the data management device 102, and the analysis device 103 via the input / output device 105.

  The data preprocessing device 101, the data management device 102, the analysis device 103, and the input / output device 105 are all information processing devices. These may be independent hardware, or two or more of them may be configured by common hardware. Each of these devices may be configured in whole or in part using a virtual information processing resource such as a cloud server in a cloud system, for example.

  FIG. 2 shows hardware of an information processing apparatus (hereinafter referred to as an information processing apparatus 10) that can be used as each of the analysis apparatus 103, the data preprocessing apparatus 101, the data management apparatus 102, and the input / output apparatus 105. It is an example of composition. As shown in the figure, the information processing device 10 includes a processor 11, a main storage device 12, an auxiliary storage device 13, an input device 14, an output device 15, and a communication device 16. These are communicably connected to each other via communication means such as a bus (not shown).

  The processor 11 is configured using, for example, a central processing unit (CPU) or a micro processing unit (MPU). The processor 11 reads and executes the program stored in the main storage device 12 to realize various functions of the analysis device 103, the data preprocessing device 101, the data management device 102, and the input / output device 105. .

  The main storage device 12 is a device for storing programs and data, and is, for example, a read only memory (ROM), a random access memory (RAM), a non-volatile semiconductor memory (NVRAM (non volatile RAM)) or the like.

  The auxiliary storage device 13 may be, for example, a hard disk drive, a solid state drive (SSD), an optical storage device (i.e., a compact disc (CD) or a digital versatile disc (DVD)), a storage system, an IC card, or an SD memory card. , An apparatus for reading / writing a recording medium such as an optical recording medium, and a storage area of a cloud server. The programs and data stored in the auxiliary storage device 13 are read into the main storage device 12 as needed.

  The input device 14 is, for example, a keyboard, a mouse, a touch panel, a card reader, a voice input device, and the like. The output device 15 is a user interface that provides the user with various information such as processing progress and processing results, and, for example, a screen display device (that is, a liquid crystal monitor, LCD (Liquid Crystal Display), graphic card, etc.), audio output Devices (ie, speakers, etc.), printing devices, etc. For example, the information processing apparatus 10 may be configured to input and output information with another apparatus via the communication apparatus 16.

  The communication device 16 is a wired or wireless communication interface for realizing communication with another device via communication means such as a LAN or the Internet, and, for example, a NIC (Network Interface Card), a wireless communication module, It is a USB (Universal Serial Interface) module, a serial communication module, or the like.

  The various functions of the analysis device 103, the data preprocessing device 101, the data management device 102, and the input / output device 105 are, for example, programs (auxiliary storage devices read by the processor 11 in the main storage device 12). 13 is implemented by executing the program read from the main storage unit 12 to the main storage unit 12).

  FIG. 3 shows the main functions of the data preprocessing apparatus 101 and the main data stored in the data preprocessing apparatus 101. As shown in the figure, the data preprocessing device 101 includes a data preprocessing unit 206 (that is, a part of the second processing unit). The data preprocessing apparatus 101 also stores the acquired data 207 and the analysis target data 208.

  The data preprocessing unit 206 can perform one or more types of preprocessing methods. The data preprocessing unit 206 preprocesses the acquired data received from the various sensors 3 via the communication network 5 by the preprocessing method instructed from the analysis apparatus 103 via the communication network 5 to obtain the analysis target data 208. The generated analysis target data 208 is sent to the data management apparatus 102.

  FIG. 4 shows main functions of the data management apparatus 102 and main data stored in the data management apparatus 102. As shown in the figure, the data management apparatus 102 includes the functions of a recommended combination search unit 306 and a data extraction unit 307. Further, the data management device 102 stores analysis target data 308, actual event data 309, analysis result data 310, analysis performance calculation result data 311, and recommended combination registration data 312. Each data is recorded, for example, in a table format.

  Among the above functions, the recommended combination search unit 306 is a combination recommended by the analysis device 103 (that is, a method combination recommended for the analysis purpose and the analysis target, and later-described series data used by the method combination (that is, analysis In the following, when the request for provision of the recommended combination (referred to as a recommended combination) is received (ie, the inquiry), the recommended combination registration data 312 corresponding to the analysis purpose and the analysis target designated by the user is The retrieved and searched recommended combination is transmitted to the analysis device 103.

  The data extraction unit 307 receives the data provision request from the analysis device 103, extracts the data corresponding to the provision request from the actual event data 309 and the analysis result data 310 according to the received provision request, and extracts the extracted data. Is sent to the analysis device 103.

  An example of the analysis target data 308 is shown in FIG. As shown in the figure, the analysis target data 308 is composed of one or more records having respective items of an acquisition date 501, a series name 502, a preprocessing method 503, a measurement item 504, and a measurement value 505. One record of the analysis target data 308 corresponds to one analysis target data.

  Among the above items, the acquisition date and time of acquisition data of the generation source of the analysis target data is set in the acquisition date and time 501. As a series name 502, a series data name to be described later is set. In the pre-processing method 503, information specifying the pre-processing method used to generate the analysis target data is set. In the measurement item 504, information indicating the content of the acquired data of the generation source of the analysis target data is set. In the measurement value 505, the value of the analysis target data (that is, the value after the pre-processing of the acquired data) is set.

  An example of the actual event data 309 is shown in FIG. The actual event data 309 is set by, for example, the user via the input / output device 105. As shown in the figure, the actual event data 309 is composed of one or more records having items of occurrence date 601, content 602, location 603, restart date 604, and failure occurrence predicted time 605. Each record of the real event data 309 corresponds to one real event.

  As the occurrence date 601, the occurrence date of the actual event is set. In the content 602, information indicating the specific content (for example, failure, maintenance, etc.) of the actual event is set. In the place 603, information indicating the position where the actual event has occurred (for example, the place or the position (ie, the place) in the analysis target 2) is set. In the restart date and time 604, information indicating the date and time when the analysis target 2 that has been stopped due to the actual event has resumed operation is set. In the failure occurrence prediction time 605, the next failure occurs if the appropriate response (ie, maintenance) is not performed, which is predicted for the device related to the actual event (ie, analysis target 2 or peripheral devices etc.) The time until occurrence (hereinafter referred to as failure occurrence prediction time) is set.

  An example of the analysis result data 310 is shown in FIG. In the analysis result data 310, the result of the analysis process performed by the analyzer 103 is managed. As shown in the figure, the analysis result data 310 is composed of one or more records including each item of data acquisition date 701, method combination ID 702, sequence group ID 703, index value 704, and presence / absence of warning 705.

  In the data acquisition date and time 701, the aforementioned acquisition date and time of the analysis target data used when executing the combination (that is, the combination of the method combination and the series data described later) is set. In the method combination ID 702, an identifier (hereinafter referred to as a method combination ID) assigned to each method combination is set. In the sequence group ID 703, an identifier (hereinafter, referred to as a sequence group ID) assigned to a group of analysis target data (hereinafter, referred to as a sequence group) used when executing the combination is set. One record of the analysis result data 310 corresponds to one of the combinations. As the index value 704, a value (hereinafter referred to as an index value) representing the degree of abnormality of the analysis target 2 calculated by execution of the combination is set. In the warning presence / absence 705, information (hereinafter referred to as warning presence / absence information) indicating the presence or absence of an abnormality of the analysis target 2 or the presence or absence of an abnormality sign is set.

  An example of the analysis performance calculation result data 311 is shown in FIG. In the analysis performance calculation result data 311, the results of the analysis device 103 calculating (i.e., evaluating) the analysis performance for each of the above combinations are managed. As shown in the figure, the analysis performance calculation result data 311 includes one or more records having respective items of analysis performance calculation date 801, method combination ID 802, sequence group ID 803, and analysis performance 804.

  In the analysis performance calculation date 801, the date and time when the analysis performance is calculated (hereinafter referred to as performance calculation date and time) is set.

  A method combination ID is set in the method combination ID 802. A series group ID is set in the series group ID 803.

  In the analysis performance 804, the analysis performance calculated (ie, evaluated) for the above combination is set. As shown in the figure, the analysis performance 804 has items of an F value 805, a recall 806, and a precision 807. In the recall rate 806, a recall rate (recall) which is a value indicating the exhaustivity of the analysis performance is set. In the precision 807, a precision that is a value indicating the accuracy of analysis performance is set. As the F value 805, an F value (F-measure) which is a harmonic mean of the recall rate and the precision rate is set.

  An example of the recommended combination registration data 312 is shown in FIG. The recommended combination registration data 312 manages the above-described recommended combination. As shown in the figure, the recommended combination registration data 312 includes items of an analysis purpose 901, an analysis target 902, an analysis target detail 903, a preprocessing method 904, an analysis method 905, a threshold 906, a measurement item 907, and an analysis performance 908. Composed of one or more records. One record of the recommended combination registration data 312 corresponds to one of the combination of analysis purpose and analysis target.

  The analysis purpose 901 is set with information indicating an analysis purpose (for example, abnormal sign detection, or improvement in work efficiency, etc.) An analysis target 902 is information for specifying the analysis target 2 (for example, a machine tool or line work In the analysis object details 903, detailed information on the analysis object 2 is set.

  In the pre-processing method 904, information specifying the pre-processing method is set. In the analysis method 905, information specifying the analysis method is set. As the threshold value 906, a value to be compared with the index value when determining the presence or absence of an abnormality or an abnormality precursor of the analysis target 2 is set. In the measurement item 907, a measurement item to be used for analysis processing is set.

  In the analysis performance 908, calculation (that is, evaluation) results of analysis performance in the case of performing data analysis with the combination of reference numerals 901 to 907 are set. As shown in the figure, the analysis performance 908 includes the items of F value 909, recall rate 910, and precision rate 911. The meaning of each of these items is the same as that described above, and thus the description thereof is omitted.

  FIG. 10 shows the main functions of the analyzer 103 and the main data stored in the analyzer 103. As shown in the figure, the analysis device 103 includes a combination determination unit 406 (that is, a first processing unit), a weight calculation unit 407 (that is, a part of a third processing unit), and an analysis processing unit 408 (that is, a second processing unit). A part of the processing unit) and each function of the performance calculation unit 409 (that is, the third processing unit) are provided. The analyzer 103 also stores pre-processing method data 411, analysis method data 412, threshold pattern data 413, selected combination registration data 414, series data attribute data 415, weighting data 417, initial setting data 418, and analysis target data 419. Do.

  Among the above functions, the combination determination unit 406 determines the above-described combination (that is, a combination of methods and a combination of methods and series data). The pretreatment method data 411 selectable in the above determination and the parameters used in the pretreatment method are shown in the pretreatment method data 411, and the analysis methods selectable in the above determination and parameters used in the analysis method are analysis method data 412 respectively. It is managed. Further, the series data that can be selected in the above determination is managed in series group registration data 416.

  The weight calculation unit 407 calculates (or manages) a weight used in calculation of analysis performance described later.

  The analysis processing unit 408 performs analysis processing on each of the combinations determined by the combination determination unit 406. The analysis object data 419 is acquired by the analysis device 103 from the data management device 102 via the communication network 5. The result of the analysis process is transmitted to the data management apparatus 102 and managed as analysis result data 310.

  The performance calculating unit 409 calculates the analysis performance for each combination based on the analysis result data 310. The analysis result data 310 is acquired by the analysis device 103 from the data management device 102 via the communication network 5.

  An example of the pre-processing method data 411 is shown in FIG. The pre-processing method data 411 stores information on selectable pre-processing methods. As shown in the figure, the preprocessing method data 411 is configured of one or more records having the preprocessing method 1001 and one or more parameters 1002.

  Among the above items, in the pre-processing method 1001, information specifying the pre-processing method is set. In the parameter 1002 (parameter 1 (1002 a) and parameter 2 (1002 b) in the example of FIG. 11), the value of the parameter used in the pre-processing is set. For example, one value may be set as the parameter value for the parameter 1002, or the minimum value, the maximum value, and the value of the step size are set, and values between the minimum value and the maximum value may be set. On the other hand, pre-processing may be performed with a specified step size. The parameter 1002 may also be set in other manners.

  An example of the analysis method data 412 is shown in FIG. The analysis method data 412 manages information on selectable analysis methods. As shown in the figure, the analysis method data 412 is composed of one or more records including an analysis method 1101 and one or more parameters 1102.

  Among the above items, in the analysis method 1101, information for specifying an analysis method is set. In the parameter 1102 (parameter 1 (1102 a) and parameter 2 (1102 b) in the example of FIG. 12), the value of the parameter given at the time of analysis by the analysis method is set. The parameter 1102 may set, for example, one value as the value of the parameter, or sets the minimum value, the maximum value, and the value of the step size, and for the values between the minimum value and the maximum value The pre-processing may be performed with a designated step size. The parameters 1102 may also be set in other manners.

  An example of the threshold pattern data 413 is shown in FIG. In the threshold pattern data 413, definition information of a pattern of a determination method (hereinafter referred to as a threshold pattern) used when the analyzer 103 determines the presence or absence of an abnormality or the presence or absence of an abnormality precursor for the analysis target 2 is managed. . As shown in the figure, the threshold pattern data 413 is composed of one or more records having items of threshold pattern ID 1201, threshold 1202, unit time 1203, upper limit 1204, and duration 1205.

  In the threshold pattern ID 1201, an identifier of the threshold pattern (hereinafter, referred to as a threshold pattern ID) is set. Information indicating whether to select each is set in each item of threshold 1202, unit time 1203, count upper limit 1204, and duration 1205 (in this example, “o” is set when selecting If not selected, "-" is set).

  In the same figure, for example, in the threshold pattern of which the threshold pattern ID 1201 is "T1", only the threshold 1202 is selected. When this threshold pattern is designated, the analyzer 103 determines that there is an abnormality or an abnormality sign when the calculated index value 704 exceeds the threshold.

  Further, for example, in the threshold pattern of which the threshold pattern ID 1201 is “T2”, the threshold 1202, the unit time 1203, and the upper limit 1204 are selected. When this threshold pattern is designated, the analyzer 103 compares the number of times the calculated index value 704 exceeds the threshold per unit time with the value of the upper limit 1204, thereby determining the presence or absence of an abnormality or the presence or absence of an abnormality sign. judge.

  An example of the selected combination registration data 414 is shown in FIG. As shown in the figure, the selection combination registration data 414 has one or more items including selection combination ID 1401, analysis purpose 1402, preprocessing purpose 1403, analysis method 1404, analysis method 1404, threshold 1405, measurement item 1406, and sequence group ID 1407. Composed of records.

  Among the above items, in the selection combination ID 1401, an identifier (hereinafter, referred to as a selection combination ID) assigned for each combination of analysis purpose, method combination, and sequence group is set.

  In the analysis purpose 1402, information indicating the analysis purpose is set. In the pre-processing method 1403, information specifying the pre-processing method is set. In the analysis method 1404, information specifying an analysis method is set. In the threshold 1405, information on a threshold to be referred to when the analyzer 103 determines the presence or absence of an abnormality or the presence or absence of an abnormality sign is set. In the measurement item 1406, information indicating the measurement item used in the analysis process is set. The series group ID described above is set in the series group ID 1407.

  An example of the series data attribute data 415 is shown in FIG. Acquisition data (i.e., data to be analyzed) includes data of a plurality of different groups of series. Hereinafter, the above group is referred to as a sequence group, and a data group belonging to the sequence group is referred to as sequence data.

  As shown in the figure, the series data attribute data 415 is composed of one or more records having respective items of series data ID 1501, series data name 1502, measurement position 1503, measurement item 1504, and attribute information 1505. One record of the series data attribute data 415 corresponds to one series data.

  Among the above items, in the series data ID 1501, an identifier (hereinafter referred to as a series data ID) assigned to each series data is set. In the series data name 1502, a name given to series data (hereinafter referred to as a series data name) is set. In the measurement position 1503, information (hereinafter referred to as measurement position information) indicating the position at which the data of the series data is measured (for example, the installation position of the various sensors 3) is set. Measurement items (for example, vibration value, temperature, etc.) are set in the measurement item 1504. In the attribute information 1505, information (hereinafter referred to as attribute information) relating to the attribute of the corresponding series data, such as a unit of data (for example, mm / s or ° C.), is set.

  An example of sequence group registration data 416 is shown in FIG. In the series group registration data 416, a data group (that is, a group of series data, hereinafter referred to as a series group) used when the analysis device 103 performs analysis processing is registered.

  As shown in the figure, the series group registration data 416 is composed of one or more records having respective items of a series group ID 1601, a series data name 1602, a measurement position 1603 and a measurement item 1604. One record of series group registration data 416 corresponds to one series data.

  Among the above items, a sequence group identifier (hereinafter referred to as a sequence group ID) is set in the sequence group ID 1601. The series data name 1602 is set to the series data name described above. The measurement position information described above is set in the measurement position 1603. The measurement items described above are set in the measurement item 1604. This configuration enables management of series data belonging to each series group. Arbitrary series data may belong to a plurality of series groups.

  An example of the weighting data 417 is shown in FIG. As shown in the figure, the weighting data 417 is composed of one or more records having items of time 1701 and weight 1702. Details of the weighting data 417 will be described later.

  An example of the initial setting data 418 is shown in FIG. The initial setting data 418 manages various information set by the user at the time of initial setting. As shown in the figure, the initial setting data 418 is composed of one or more records each having an initial setting ID 1801, an analysis purpose 1802, an analysis target 1803, and details 1804. In these items, values input by the user at the time of initial setting are set.

  Subsequently, processing performed in the data analysis system 1 will be described. At the time of introduction of the data analysis system 1 etc., the user etc. first perform initial setting, and specify a combination to be tried out of combinations of method combinations and sequence groups. The data analysis system 1 executes processing corresponding to each of the designated combinations. Then, the data analysis system 1 evaluates analysis performance for each combination as needed (e.g., periodically, when an event such as an abnormality occurs, or when a system user requests, etc.), and reviews each combination.

  FIG. 19 is a sequence diagram for explaining a process (hereinafter referred to as an initial setting process S1900) related to the initial setting of the analyzer 103, which is performed when the data analysis system 1 is introduced or the like. In the following description, the letter “S” attached to the beginning of the code means a processing step.

  First, the combination determination unit 406 of the analysis device 103 receives an input of analysis purpose and analysis target via the input / output device 105 (S1901), and provides the data management device 102 with data of recommended combination corresponding to analysis purpose and analysis target. Request (S1902).

  When the recommended combination search unit 306 of the data management apparatus 102 receives the above request, the recommended combination search unit 306 searches the recommended combination registration data 312 for data corresponding to the analysis purpose and the analysis target designated by the above request (S1903). When the recommended combination is searched as a result of the search, the recommended combination search unit 306 transmits the recommended combination to the analysis apparatus 103 as a result of no recommended combination if the recommended combination is not searched (S1904).

  The combination determination unit 406 of the analyzer 103 confirms the presence or absence of the recommended combination (S1905). If the recommended combination is received (S1905: YES), the combination determining unit 406 displays the received recommended combination on the input / output device 105 (S1906). Also, when presenting recommended combinations to the user, information on the analysis performance of each recommended combination may be presented. Also, if there is no recommended combination (S1905: NO), for example, combination determining section 406 provisionally selects all possible method combinations and series data combinations, and enters information indicating that all combinations have been selected. It is displayed on the output device 105 (S1907).

  Subsequently, the combination determination unit 406 determines a method combination based on the user's operation input (ie, selection operation from the recommended combination or all combinations) received via the input / output device 105 (S1908), Sequence groups to be used for analysis processing are determined (S1909).

  Subsequently, the combination determination unit 406 transmits (i.e., instructs) the data preprocessing apparatus 101 on the preprocessing method based on the determined combination (S1910).

  Subsequently, the combination determination unit 406 stores the information set in the above processing as the initial setting data 418 (S1911).

  FIGS. 20 and 21 are examples of screens (hereinafter referred to as an initial setting screen 2000) presented to the user via the input / output device 105 in the initial setting process S1900. FIG. 20 is an example in the case where the recommended combination does not exist, and FIG. 21 is an example in the case where the recommended combination exists.

  First, FIG. 20 will be described. As shown in the figure, the initial setting screen 2000 has an analysis purpose input field 2001, an analysis object input field 2002, a measurement item registration field 2003, an analysis method designation field 2004, and a combination selection field 2005.

  In the analysis purpose input field 2001, an analysis purpose (for example, equipment abnormality sign detection, work process abnormality monitoring, or quality estimation) is input. The analysis purpose input field 2001 may be set by, for example, selection by a pull-down menu or direct input.

  In the analysis target input field 2002, information on the analysis target 2 (for example, device type, device name, phenomenon of the target to be analyzed, etc.) is input. These may be set by selection by pull-down or direct input. In addition, when an object type is set, an apparatus corresponding to the object may be displayed in a pull-down. Also, the corresponding device name may be displayed in the pull-down by the selection of the device type, and the corresponding target phenomenon may be displayed in the pull-down by the selection of the device name. Each column of the analysis target input column 2002 may be left blank.

  When the user inputs contents in the analysis purpose input field 2001 and the analysis object input field 2002 and operates the determination button 2006, a measurement item registration field 2003a and an analysis method specification field 2004 are displayed.

  The user selects a measurement value that can be input to the data analysis system 1 in the measurement item registration field 2003a. When the user operates the execute button 2007 in all combinations in the analysis method designation column 2004, all combinations are selected and combination determination is made (S1908 in FIG. 19). When the user operates the combination selection button 2008 in the analysis method designation column 2004, the combination selection column 2005 is displayed, and a pretreatment method, an analysis method, a threshold used to determine the presence or absence of an abnormality or the presence or absence of an abnormality sign, and measurement It becomes possible to specify the item.

  The check box 2009 of the combination selection box 2005 is a combination selection box, and is used to determine whether or not to select an analysis by the displayed combination. The selection combination ID described above is set in the selection combination ID 2010.

  The preprocessing 2011, the analysis method 2012, the threshold 2013, and the measurement item 2014 are fields in which the user selects each item. When the user sets each item and operates the determination button 2015, the combination is determined (S1908 in FIG. 19). In the measurement item 2014, items other than the items registered in the measurement item registration field 2003a may not necessarily be displayed.

  The analysis target input field 2002 may be displayed after the input of the analysis purpose is completed. The measurement item registration field 2003, the analysis method designation field 2004, and the combination selection field 2005 may be displayed after the input of the analysis object input field 2002 is completed. Also, not all elements need to be displayed. Also, it is not necessary to fill in all the input fields, and blank spaces may exist.

  Subsequently, an initial setting screen 2000 when there is a recommended combination will be described with reference to FIG.

  If the recommended combination exists, the user inputs in the analysis purpose input field 2001 and the analysis object input field 2002 and operates the determination button 2006, the recommended measurement items are displayed in the measurement item registration field 2003b.

  The recommended analysis method selection field 2101 displays a recommended combination. If there is a combination for which analysis is desired to be additionally performed, the user designates preprocessing, an analysis method, a threshold value, and a measurement item in the additional combination selection field 2102.

  If the user adds a check in the recommended analysis method selection field 2101 and the additional combination selection field 2102 and operates the determination button 2105, the combination is determined (S1908 in FIG. 19). The execute button 2104 may be operated in all combinations to select all combinations.

  FIG. 22 is an example of a screen (hereinafter referred to as a series data attribute input screen 2200) displayed by the input / output unit 105 when the input / output unit 105 reads an attribute of series data.

  As shown in the figure, in the series data attribute input screen 2200, input fields of series data ID 2201, series data name 2202, measurement position 2203, measurement item 2204, attribute information 2205, and file reading unit 2206 are provided. It is done.

  The series data ID is set in the series data ID 2201 among the above items. The series data name 2202 is set to the series data name described above. Measurement position information is set in the measurement position 2203. The measurement items described above are set in the measurement item 2204. The attribute information described above is set in the attribute information 2205. Attribute data may be input from the input screen for each series data, or a file reading unit 2206 may be used to read a file in which attribute data is described.

  FIG. 23 is a screen (hereinafter referred to as a series group designation screen 2300) displayed by the input / output device 105 when the user designates a series group (that is, one or more series data) used for each method combination. is there. As shown in the figure, the series group designation screen 2300 has a method combination designation column 2301 and an analysis target series selection column 2302 (in the example of FIG. 23, 2302a and 2302b).

  In the method combination designation column 2301, the user selects a series group from among the method combinations selected in the combination selection column 2005 of FIG. 20, the recommended analysis method selection column 2101 of FIG. 21, and the additional combination selection column 2102 of FIG. Select the method combination to be set.

  The analysis object series selection column 2302 specifies one or more series data to be used for the method combination selected in the combination designation column 2301 by putting a check in the selection column 2303.

  The contents read in the series data attribute input screen 2200 of FIG. 22 are displayed in the series name 2304, the measurement position 2305, and the measurement item 2306. When setting these items (that is, reference numerals 2304 to 2306), a search function, an alignment function, or filtering by specifying conditions may be used in combination. Further, the selection column 2303 may be able to check collectively.

  In this example, it is specified that analysis using series data designated in group 1 and analysis processing using series data designated in group 2 are performed for both the combination of method combination ID “A1” and “B1”. It is done. As the method combination ID in the combination designation column 2301, for example, all combinations may be designated by a simple operation. Further, a combination or series group may be added as needed by using a combination designation addition button or a group addition button.

  The combination determination unit 406 of the analysis apparatus 103 sets the information set in the analysis purpose input field 2001 and the analysis object input field 2002 of the initial setting screen 2000 as the initial setting data 418, and also selects the combination selection field 2005 and the recommended analysis method selection field 2101. The information input in the additional combination selection column 2102 is set as selection combination registration data 414, and the information set via the series data attribute input screen 2200 is set as series data attribute data 415, and is set in the series group specification screen 2300. The stored information is stored as sequence group registration data 416, respectively.

  FIG. 24 is a sequence diagram illustrating a process (hereinafter, referred to as analysis process S2400) performed in the data analysis system 1 when the analysis apparatus 103 performs the analysis process. The analysis processing S2400 will be described below with reference to FIG.

  As shown in the figure, acquired data is transmitted from the various sensors 3 to the data preprocessing device 101 (S2401 to S2402).

  The data preprocessing unit 206 of the data preprocessing device 101 performs preprocessing on the acquired data received from the various sensors 3 by one or more preprocessing methods instructed from the analysis device 103 in S1910 of FIG. S2403) The generated analysis object data is transmitted to the data management apparatus 102 (S2404).

  The data management apparatus 102 stores the received analysis target data as analysis target data 308 (S2405), and transmits the analysis target data to the analysis apparatus 103 (S2406).

  The analysis processing unit 408 of the analysis apparatus 103 stores the analysis target data received from the data management apparatus 102 as the analysis target data 419, and for the analysis target data 419, a combination determined in the initialization or the combination review described later. The analysis processing is performed based on the above to calculate an index value representing the degree of abnormality of the analysis target 2 (S2407).

  Further, the analysis processing unit 408 determines the presence or absence of an abnormality or an occurrence of an abnormality of the analysis target 2 based on the threshold pattern (S2408), and as a result of the determination, if a warning (ie, alert) is necessary, the warning It outputs to 105 (S2409).

  Subsequently, the analysis processing unit 408 transmits the analysis result to the data management apparatus 102 (S2410), and the data management apparatus 102 stores the received analysis result as the analysis result data 310 (S2411).

  FIG. 25 is a sequence diagram showing an example of processing (hereinafter referred to as analysis performance calculation & review processing S2500) performed in the data analysis system 1 when the analysis apparatus 103 performs processing regarding calculation of analysis performance and revision of the combination. It is. The analysis performance calculation & review process S2500 will be described below with reference to FIG.

  When the user inputs actual event data (S2501), the input / output device 105 transmits the actual event data to the data management device 102 (S2502) and transmits an input notification of the actual event data to the analysis device 103 (S2503) .

  The data management apparatus 102 stores the received real event data as real event data 309 (S2504).

  When the performance calculation unit 409 of the analysis device 103 receives the input notification of the actual event data, the performance calculation unit 409 transmits a provision request of the actual event data and the analysis result data to the data management device 102 (S2505). At this time, the performance calculation unit 409 of the analysis apparatus 103 may specify a time range and request real event data and analysis result data within the specified time range.

  When receiving the provision request, the data extraction unit 307 of the data management device 102 extracts the requested data from each of the actual event data 309 and the analysis result data 310 (S2506), and transmits the extracted data to the analysis device 103. Send (S2507).

  The performance calculation unit 409 of the analysis device 103 calculates analysis performance based on the received actual event data and analysis result data (S2508). The details of this process (hereinafter, referred to as analysis performance calculation process S2508) will be described later.

  Subsequently, the combination determination unit 406 of the analysis device 103 reviews the combination (S2509). The details of this process (hereinafter, referred to as combination review process S2509) will be described later.

  Subsequently, the combination determination unit 406 transmits the information on the calculated analysis performance (hereinafter referred to as analysis performance calculation result data) to the data management apparatus 102 (S2510), and the data pre-processing apparatus 101 before reviewing it An instruction on the processing method is issued (S2512).

  The performance calculation method may differ depending on the purpose of analysis. In that case, the performance calculation unit 409 executes the performance calculation method for each analysis purpose, and calculates the analysis performance using the performance calculation method suitable for the analysis purpose.

  When the data management apparatus 102 receives the analysis performance calculation result data, it stores the received analysis performance calculation result data as analysis performance calculation result data 311 (S2511). In addition, the data management apparatus 102 stores the combination having the higher performance in the recommended combination registration data 312 (S2511).

  Note that, for example, the analysis device 103 notifies the data management device 102 of a combination of high performance together with information such as analysis purpose and analysis target 2 as a recommended combination, and the data management device 102 sets this as recommended combination registration data 312 It may be registered. Further, FIG. 25 shows an example in which performance calculation and method combination review are triggered by input of actual event data (S2501), but this processing is executed periodically or triggered by a user request. You may

  Subsequently, the analysis performance calculation process S2508 shown in FIG. 25 will be described in detail. In general, it is desirable in the detection of abnormality sign that there is a time from the detection point of the abnormality sign to the time when the failure actually occurs. Therefore, in the following, a case will be described as an example in which the analysis performance is calculated by weighting according to the time from the detection of an abnormality sign to the occurrence of a failure.

  FIG. 26 is an example of a screen (hereinafter referred to as a weighting setting screen 2600) displayed by the input / output device 105 when setting the weighting. As shown in the figure, the weighting setting screen 2600 has a basic information input field 2602, a weighting setting field 2603, and a weighting graph display field 2604.

  When the user inputs the maintenance frequency / interval and the time required to stop the device in the basic information input column 2602 and operates the OK button 2605, the weight calculation unit 407 of the analysis device 103 selects a weight based on the input information. Calculate and generate a weighting table 2606. The weight calculation unit 407 may display a weighting graph 2607 as shown in the figure to visualize the contents of weighting. Also, the user may operate the weighting table 2606 or the weighting graph 2607 to easily change the weight. Also, when the user corrects one of the weighting table 2606 or the weighting graph 2607, the same correction may be automatically reflected in the other. When the user operates the determination button 2608, the weight calculation unit 407 stores the content input on the weighting setting screen 2600 as weighting data 417.

  Although four times are specified in this example, the mode of specifying time is not necessarily limited. Also, based on the time specified (ie, set) in the weighting table 2606, the weight for the time between each specified time may be automatically interpolated with a temporary function or the like. The setting of the weighting may be performed, for example, at the time of the initial setting described above.

  FIG. 27 is a flowchart for explaining an example of the analysis performance calculation process S2508 in FIG. 25 when the analysis purpose is abnormality sign detection. The analysis performance calculation processing S2508 will be described below with reference to FIG.

  First, the performance calculation unit 409 of the analysis apparatus 103 selects one combination from among one or more combinations (that is, a combination of methods and series groups) for which analysis has been performed (S2701). Next, the performance calculation unit 409 selects the record with the earliest occurrence date 601 among the unselected records of the actual event data 309 (S 2702), and the warning presence or absence 705 among the unselected records of the analysis result data 310. “Available” is set, and the data acquisition date 701 is the earliest record (S2703). Note that, in the analysis performance calculation process S2508, the record with alert indicates a record in which an alert is set as a result of analysis by a combination during analysis performance calculation (that is, a combination of method combination and series group).

  Subsequently, the performance calculation unit 409 determines the data acquisition date 701 (hereinafter also referred to as alert occurrence date) of the record of the analysis result data 310 being selected and the occurrence date 601 of the record of the actual event data 309 being selected (hereinafter, The front-rear relationship with the failure occurrence date and time is also compared (S2704).

  As a result of comparison, if “alert occurrence date> failure occurrence date” (S2704: No), the performance calculation unit 409 adds 1 to the missing number (S2715), and the record of the selected real event data 309 is selected. It is determined whether the record of actual event data 309 whose occurrence date and time 601 is later is (presence or absence of failure) (S2716), and if there is a record (S2716: YES), the record is selected (S2717), and the process of S2704 Return to On the other hand, if there is no record (S2716: NO), the performance calculation unit 409 sets the record of the unselected analysis result data 310 in which “presence” is set to the warning presence 705 in the false alarm count (analysis result data 310 during selection (S2718), and then the process proceeds to S2712.

  In S2704, if “alert occurrence date ≦≦ fault occurrence date” (S2704: YES), the performance calculation unit 409 performs within a certain period (for example, time-out time in reference numeral 2606 in FIG. 26) after the alert occurs. It is checked whether or not there is a failure (that is, the presence or absence of such a real event data 309 record) (S2705).

  If a failure has not occurred within a predetermined time (S2705: NO), the performance calculating unit 409 adds 1 to the number of false alarms (S2719), and thereafter proceeds to the processing of S2709.

  On the other hand, if a failure has occurred within a predetermined time (S2705: YES), the performance calculation unit 409 determines from the data acquisition date 701 (that is, the alert occurrence date) of the analysis result data 310 under selection the failure occurrence date 601 The weight according to the time until is determined (S2706), the determined weight is added to the weight sum, and 1 is added to the detection number (S2707).

  Subsequently, the performance calculating unit 409 skips all the records from the selected analysis result data record to the failure occurrence date 601 (that is, from the selected analysis result data record to the failure occurrence date 601) All the records in the list are selected (S2708).

  Subsequently, the performance calculating unit 409 determines the presence / absence of a record of actual event data (that is, the presence / absence of a fault) whose occurrence date / time 601 is later (S2709). If there is no next record (that is, there is no later failure) (S2709: NO), the performance calculation unit 409 sets the number of false alarms to the number of unselected alerts (that is, “Yes” is set to the warning 705). The number of records of unselected analysis result data) is added (S2720), and thereafter, the process proceeds to the processing of S2712.

  On the other hand, when there is the next failure (S2709: YES), the performance calculation unit 409 determines whether or not there is an unselected alert (that is, the presence or absence of a record of unselected analysis result data for which “presence” is set in the warning 705). ) (S2711), and if there is an unselected alert (S2711: YES), the process returns to S2703. On the other hand, if there is no unselected alert (S2711: NO), the process proceeds to S2712.

  In S2712, the performance calculating unit 409 adds the number of unselected actual event data records (that is, the number of unselected failures) to the missed count.

Subsequently, the performance calculation unit 409 calculates, for example, various performance values from the following formula (S2713).
Reproduction rate = total weight / number of failure occurrences Equation 1
Accuracy rate = number of detections / (number of detections + number of false alarms) ··· Equation 2
F value = (2 x precision ratio x recall ratio) / (accuracy rate + recall ratio) Formula 3

  Subsequently, the performance calculation unit 409 determines the presence or absence of an unprocessed combination among one or more combinations (that is, a combination of methods and series groups) subjected to the analysis (S2714). If there is an unselected combination (S2714: YES), the processing returns to S2701. If there is no unselected combination (S2714: NO), the processing ends.

  Through the above process, the performance calculation unit 409 calculates the number of missed calls, the number of false alarms, the recall rate, the precision rate, and the F value for each combination (that is, a combination of method combination and series data). Also, the performance calculation unit 409 may calculate, for each combination, the average and variance of weights, or the number of false alarms per predetermined unit time. Also, in this case, the performance calculating unit 409 may set the weight for missing a fault as 0 and obtain the average and the variance. The performance calculating unit 409 transmits these pieces of performance information to the data management apparatus 102 together with information such as the analysis purpose and the analysis target 2, and the data management apparatus 102 accumulates the information in the analysis performance calculation result data 311. The failure information may reflect not only a failure that actually occurred, but also a failure that would have occurred if repair or parts replacement were not performed based on maintenance information. By including the maintenance information in this manner, it is possible to accurately calculate the analysis performance including the case where the occurrence of a warning is performed and maintenance is performed to avoid the occurrence of a failure in advance.

  Subsequently, the combination review process S2509 shown in FIG. 25 will be described in detail.

  FIG. 28 shows an example of a screen (hereinafter referred to as a combination review screen 2800) displayed by the input / output device 105 in the combination review process S2509 of FIG. As shown in FIG. 2802, selection column 2803, combination identifier 2804, sequence group 2805 used for analysis, pre-processing method 2806, analysis method 2807, and threshold value 2808, analysis performance (reproducibility 2809, accuracy 2810, F-number 2811, weighted average 2812 A line format setting field is provided which has items of weight distribution 2813, false alarm number 2814 per unit time, and the like. One row in the figure corresponds to one combination.

  Each combination may be displayed in an ascending or descending order, for example, based on the value of the numerical value such as the recall rate or the precision rate. Also, filtering may be performed according to the value of each item. By outputting the result of the alignment and the filtering, the user can easily grasp, for example, the upper combination of analysis performance. When the user sets a check in the selection column 2803 and operates the determination button 2815, the combination determination unit 406 stores the checked combination, and the analysis processing unit 408 performs processing based on the combination.

  In this example, the user selects the content of the combination, but the combination determination unit 406 may automatically select the combination according to the method specified by the user.

  FIG. 29 is an example of a screen (hereinafter referred to as a selection method designation screen 2900) displayed by the input / output device 105 when the combination determination unit 406 automatically configures a combination. As shown in the figure, the selection method designation screen 2900 includes a selection column 2902 and a combination selection method setting column 2903. The user can select one or more combination selection methods by inputting a check in the selection field 2902. When reviewing the combination, the performance calculation unit 409 automatically sets the content of the combination according to the selection method specified by the user on the selection method specification screen 2900. If there is no combination that satisfies the content (that is, the condition) set on the selection method designation screen 2900 by the user, the combination determination unit 406 sets the upper predetermined on the condition of F value priority or weighted recall ratio priority and the like. A combination of numbers may be selected automatically.

  Further, the combination determination unit 406 may automatically select the combination based on, for example, the distribution of weights for each actual event determined in the analysis performance calculation process S2508.

  In FIG. 30, data to be referred to when the combination determination unit 406 automatically selects the contents of the method combination based on the determined distribution of weights (S2706) for each actual event (hereinafter referred to as weight recording data 3000). An example of As shown in the figure, the weight recording data 3000 includes an actual event 3002, a method combination ID 3003, a sequence group ID 3004, and a weight 3005. The actual event 3002 is No. Each item of 3006, content 3007, and place 3008 is included.

  No. The identifier of the actual event is set in 3006, the information indicating the content of the actual event is set in the content 3007, and the information indicating the occurrence position of the actual event is set in the place 3008. Further, the method combination ID is set in the method combination ID 3003, the sequence group ID is set in the sequence group ID 3004, and the weight determined by the combination determination unit 406 in S2706 of FIG.

  For example, the combination determination unit 406 may specify a condition for an average value, a variance value, a 99% value or the like of weights, calculate a weight value for each combination, and select a combination that satisfies the condition. According to this, for example, when the weight is set to indicate that the precursor is detected earlier as the value is closer to "1" and indicates that the detection of precursor is slower as the value is closer to "0", It is possible to select a combination using an arbitrary condition such as a 99% value of 0.7 or more, a weight of 0.5 or more of 90%, or the like.

  For example, while selecting a combination of upper predetermined% (for example, 20%) in analysis performance and learning as a recommended combination, try and narrow down on other combinations (for example, upper 20% × 2 in analysis performance) You may Further, the combination may be narrowed down gradually by excluding several% lower in performance each time in the evaluation of the analysis performance. The review of the combination does not necessarily have to be performed every time the evaluation of the analysis performance is performed, and may be terminated for a certain period or a certain number of times after the analysis processing is started, or at a stage where the number of combinations is narrowed. Further, a combination having a high value such as the recall ratio or F value may be learned as a recommended combination. Also, although the above describes the abnormality sign detection using data collected using various sensors 3 as an example, the analysis target data is, for example, the load of a processor included in the device, operating time, or image data, etc. It may be collected by a method. The preprocessing method, the analysis method, and the threshold pattern may be added and deleted by the user even after the data analysis system 1 is operated.

  Further, the pre-processing method does not necessarily have to select the same method for each sequence group, and different pre-processing methods may be selected for each measurement item. Further, the data analysis system 1 may be configured to be able to analyze one or more analysis targets 2.

  In addition, when the new data analysis system 1 is constructed, if the analysis target data already collected for the analysis target 2 exists, the analysis processing S2400 or the analysis performance calculation & review processing S2500 using the collected analysis target data To select a method with good performance and perform data analysis thereafter. This makes it possible to shorten the lead time when constructing the data analysis system 1.

  Alternatively, a plurality of data analysis systems 1 may be communicably connected, and the plurality of data analysis systems 1 may share information of a recommended combination. For example, the data management apparatus 102 is provided with a function of sharing information of recommended combinations, and the data management apparatus 102 of each data analysis system 1 shares information of recommended combinations by transmitting and receiving information of recommended combination registration data 312. Let's do it. The same recommended combination registration data 312 may be shared by each data analysis system 1. As a result, for example, the performance information learned for each analysis purpose or target device in the existing data analysis system 1 can be used for the other data analysis system 1 newly constructed, and the analysis performance in the other data analysis system 1 is The top combinations can be easily identified. In addition, the lead time can be shortened when constructing a new data analysis system 1 by giving priority to analysis with a combination of top analysis performances, and resource consumption when constructing a new data analysis system 1 The amount can also be reduced.

  When data is transmitted and received among a plurality of data analysis systems 1, all or one of the data preprocessing device 101, the data management device 102, the analysis device 103, and the input / output device 105 of each data analysis system 1 is provided. The configuration of the part may be shared with one or more other data analysis systems 1. As a result, for example, surplus data held by each data analysis system 1 can be shared, and resources of each data analysis system 1 can be effectively used.

  In addition to the pretreatment method, the analysis method, and the threshold pattern, the analyzer 103 sets a plurality of combinations for each analysis purpose (that is, a combination of the pretreatment method, the analysis method, the threshold pattern, and each parameter) in advance. It may be held (i.e. stored). The above combination may be based on empirically obtained findings. For example, when no data is stored in the recommended combination registration data 312 and a recommended combination can not be obtained from the data management apparatus 102, the combination determination unit 406 presents the combination as a preset combination. According to this, it becomes possible to use knowledge other than the data learned by the data analysis system 1, and even when the recommended combination is not registered, a combination which can be expected to have high performance without executing all combinations is possible. It is possible to perform selectively.

  As described above, according to the data analysis system 1 of the present embodiment, the optimum combination of method combinations (i.e., pre-processing method, analysis method, parameters) and sequence groups (i.e., one or more sequence data) It is possible to select a combination. In addition, it becomes possible to select an optimal combination of method combination and series group for analysis purpose and analysis target. In addition, it is possible to select a high-performance combination according to the purpose of analysis, regardless of the skill of the user such as a system builder.

Second Embodiment
In the first embodiment, the case where the analysis purpose is abnormal sign detection is described as an example, but in the second embodiment, the case where the analysis purpose is work process control will be described. The basic configuration of the data analysis system 1 of the second embodiment and the flow of processing performed by the data analysis system 1 are basically the same as those of the first embodiment. Hereinafter, differences from the first embodiment will be mainly described.

  FIG. 31 is an example of an initial setting screen 2000 displayed by the input / output device 105 at the time of initial setting processing S1900 of the second embodiment. As in the first embodiment, the analysis purpose input field 2001, the analysis object input field 2002, the measurement item registration field 2003c, the analysis method designation field 2004, the combination selection field 2005, and the recommended analysis also for the initial setting screen 2000 of the second embodiment. A method selection field 2101 and an additional combination selection field 2102 are included. About these, since it is the same as that of a 1st embodiment, explanation is omitted. The recommended measurement item in the measurement item registration column 2003c may be a measurement item different from that of the first embodiment in accordance with the analysis purpose 2001, the analysis target 2002, and the like.

  As shown in the figure, in this example, the analysis purpose is "work process control", the analysis target is "line", the work content is "assembly work", and the analysis details are "work abnormality" and "efficiency". . As in the present example, the number of analysis details is not necessarily one, and a plurality of analysis details may be specified. The “operation error” is, for example, “in wrong order of operation steps”, “out of operation”, “additional operation is included”, and “operation is correct but failed”. Further, “efficiency” is, for example, the number of failures such as “the number of times the material is taken out”, “the table on which the material is dropped”, and the like.

  FIG. 32 is a flowchart for explaining analysis performance calculation processing S2508 in the second embodiment. The following description will be made with the same figure.

  The processes of S3201 to S3204 in the same figure are the same as the processes of S2701 to S2704 of FIG.

  In S3215, the performance calculation unit 409 adds 1 to the missing number and 1 to the missing number for each type of work abnormality.

  The processes of S3216 to S3217 are the same as the processes of S2716 to S2717 of FIG.

  In S3218, the performance calculation unit 409 adds the number of unselected alerts including the selected alert to each of the number of false alarms and the number of false alarms for each work abnormality type, and then the process proceeds to S3212.

  In step S3205, the performance calculation unit 409 determines whether the alert occurrence date and time and the work abnormality occurrence date and time match. If the alert occurrence time does not match the work abnormality occurrence time (S3205: NO), the performance calculation unit 409 adds 1 to the false alarm count and 1 to the false alarm count for each work abnormality type (S3219), and thereafter The process proceeds to step S3209. If the alert occurrence date and time and the work abnormality occurrence date coincide (S3205: YES), the performance calculation unit 409 adds 1 to the detection number and 1 to the detection number for each work abnormality type (S3206), and thereafter The process proceeds to step S3209.

  The processes of S3209 to S3211 are the same as the processes of S2709 to S2711 of FIG.

  In step S3220, the performance calculation unit 409 adds the number of false alarms and the number of unselected alerts to the number of false alarms for each type of work or more, and then proceeds to the process of step S3212.

  In S 3212, the performance calculation unit 409 sets the number of missed actual event data records (that is, the number of unselected failures) as the missed count as the missed count for each operation abnormality type. The number of records of (ie, the number of unselected failures) is added to each.

In the subsequent S3213, the performance calculating unit 409 calculates the recall, the precision, the F value, and the like. Here, the performance calculation unit 409 calculates the relevance ratio and the F value according to Equations 2 and 3 of the first embodiment, and calculates the recall ratio using the following equation.
Reproduction rate = number of detections / number of failures

  In addition, the performance calculation unit 409 also uses the number of missed calls, the number of false alarms, and the number of occurrences for each work type for the recall rate, precision rate, and F value for each work abnormality type, respectively. Calculated by 3. At this time, the number of false alarms per unit time or another index value may be calculated.

  Subsequently, the performance calculation unit 409 determines the presence or absence of an unprocessed combination based on one or more combinations (that is, a combination of methods and series groups) subjected to the analysis (S3214). If there is an unselected combination (S3214: YES), the processing returns to S3201, and if there is no unselected combination (S3214: NO), the processing ends.

  The comparison between S3204 and S3205 may not necessarily be exact, and may be made with some margin in consideration of the error of the date and time recording of the actual event. In addition, when one or more alerts occur in a short time, one or more alerts may be collectively counted as one. The actual event data may also include occurrence date and time such as work abnormality and falling of the work table, and types of cases (i.e., wrong order, missing process, falling on the work table, etc.).

  FIG. 33 is an example of the combination review screen 2800 in the second embodiment. The items denoted by reference numerals 2802 to 2811 and 2814 are the same as those in the first embodiment. Unlike the first embodiment, in the combination review screen 2800 of the second embodiment, items of weight average 2812 and weight distribution 2813 do not exist, and items of work abnormality type 2820 are separately provided. According to the combination review screen 2800, for example, it is possible to select a specific work abnormality by filtering by the work abnormality type 2820 and select a combination most suitable for the abnormality.

  According to the data analysis system 1 of the second embodiment, it is possible to select one or more combinations each having good analysis performance with respect to all types of abnormality and one or more combinations having good analysis performance for each type of abnormality. is there. According to this, even if there are various types of work abnormality and there is no combination for covering and detecting all abnormalities, it is possible to select a combination with good analysis performance for each type of abnormality. Moreover, the analysis performance of each combination can be learned for each type of work abnormality, and the detection performance of the work abnormality can be improved. In addition, when using the learned analysis performance when constructing a new data analysis system 1, it is possible to efficiently select a good combination of analysis performance for the whole and every work abnormality, and lead time can be shortened. And reduction of computational resources. Furthermore, when the type of abnormality detected for each combination is significantly different, it is also possible to estimate the type of abnormality by examining the combination determined to have a warning.

Third Embodiment
In the third embodiment, processing performance is improved by performing load distribution using a plurality of analyzers 103 in the first embodiment or the second embodiment. In the present embodiment, one of the plurality of analyzers 103 is used as a representative analyzer 103a, and the other analyzers 103 are distinguished as analyzers 103b. The representative analyzer 103a is provided with a process assignment function. The representative analysis device 103a manages the information of the other analysis device 103b, and distributes the load by assigning the process to the other analysis device 103b.

  FIG. 34 is a sequence diagram for explaining the process (hereinafter referred to as the initial setting process S3400) corresponding to the initial setting process S1900 of the first embodiment in the second embodiment.

  First, the processing of reference numerals S3401 to S3409 in the figure is the same as the processing of S1901 to S1909 in FIG. 19 except that the processing of the analysis device 103 is performed by the representative analysis device 103a, so the description will be omitted.

  In S3410, the combination determination unit 406 of the representative analysis device 103a (or the processing assignment unit or the like may be provided in the analysis device 103a) assigns the combination processing to a plurality of other analysis devices 103b (S3410) and assignment The result is notified to each analyzer 103b (S3411). The combination determination unit 406 of the representative analysis device 103a instructs the data preprocessing device 101 on the preprocessing method based on the determined combination (S3412).

  The other analysis apparatuses 103b request the data management apparatus 102 for the analysis target data 419 corresponding to the assigned process (S3413), and the data management apparatus 102 sends the requested analysis target data 419 to the other requester. Is transmitted to the analysis device 103b (S3414).

  The processing of S3415 is the same as that of S1911 of FIG.

  In the analysis processing S2400, the other analyzer 103b performs, for example, processing assigned to itself (for example, processing of a combination of a specific method combination and a sequence group), and the result is sent to the representative analyzer 103a. Send. The representative analyzer 103a aggregates the processing results sent from each of the other analyzers 103b, and performs analysis performance calculation & review processing S2500 based on the aggregated information. The representative analysis device 103a reassigns the process to another analysis device 103b based on the result of the analysis performance calculation and review process S2500, for example, and notifies the other analysis device 103b of the assignment result. The other analyzer 103b executes the newly assigned process.

  As described above, by load-distributing the combination processing using the plurality of analyzers 103, the processing can be efficiently executed, and the turn around time can be shortened.

  As mentioned above, although this invention was concretely demonstrated based on embodiment, this invention is not limited to said embodiment, It can change variously in the range which does not deviate from the summary. For example, the above embodiments have been described in detail in order to explain the present invention in an easy-to-understand manner, and are not necessarily limited to those having all the described configurations. In addition, with respect to a part of the configuration of the above embodiment, other configurations can be added, deleted, and replaced.

  In addition, each of the configurations, function units, processing units, processing means, etc. described above may be realized by hardware, for example, by designing part or all of them with an integrated circuit. Further, each configuration, function, etc. described above may be realized by software by a processor interpreting and executing a program that realizes each function. Information such as a program, a table, and a file for realizing each function can be placed in a memory, a hard disk, a recording device such as a solid state drive (SSD), or a recording medium such as an IC card, an SD card, or a DVD.

  Further, in each drawing, control lines and information lines indicate what is considered to be necessary for explanation, and not all the control lines and information lines in mounting are necessarily shown. For example, in practice it may be considered that almost all configurations are mutually connected.

  The arrangement of the various functional units, the various processing units, and the various databases of the information processing apparatus described above is merely an example. The arrangement form of the various functional units, the various processing units, and the various databases can be changed to an optimum arrangement form of each information processing apparatus in terms of hardware and software performance, processing efficiency, communication efficiency, and the like.

  Further, the configurations of the various databases described above can be flexibly changed from the viewpoints of efficient use of resources, improvement of processing efficiency, improvement of access efficiency, improvement of search efficiency and the like.

  Further, the data described above is not limited to the table structure, as long as each data can be appropriately associated and stored, and may be recorded in another structure such as a queue structure or a list structure. In addition to numerical values, numerical values may be used as a method of recording values. Further, each item recorded in each table may be different depending on the application, and is not limited to the items described as an example. Also, the data may be distributed and arranged in a plurality of information processing apparatuses.

  Also, in the above, when referring to the number of elements (for example, the number, numerical value, quantity, range, etc.), unless otherwise specified or the principle is clearly limited to a specific number, etc. The number is not limited to a specific number, and may be a specific number or less.

  Further, in the above description, the constituent elements (eg, each function, table, element step, etc.) are not necessarily essential unless particularly clearly stated or considered to be obviously essential in principle. Absent.

  Further, each processing described above may be executed as batch processing as well as real time processing.

  In addition, each of the embodiments described above may be applied independently, or all or a part of a plurality of embodiments may be combined and applied.

Reference Signs List 1 data analysis system 2 analysis target 3 various sensors 101 data pre-processing device 102 data management device 103 analysis device 206 data pre-processing unit 207 acquisition data 208 analysis target data 306 recommended combination search unit 307 Data extraction unit, 308 analysis target data, 309 real event data, 310 analysis result data, 311 analysis performance calculation result data, 312 recommended combination registration data, 406 combination determination unit, 407 weight calculation unit, 408 analysis processing unit, 409 performance calculation Part, 411 preprocessing method data, 412 analysis method data, 413 threshold pattern data, 414 selected combination registration data, 415 series data attribute data, 416 series group registration data, 417 weighting data, 418 initial setting data, 419 analysis target Data, S1900 initial setting process, S2400 analysis process, S2500 analysis performance calculation & review process

Claims (16)

Acquisition data which is data acquired for an analysis target, and
Actual event data, which is data relating to an event that actually occurred for the subject of analysis;
A storage unit for storing
One or more first methods, which are first processes related to analysis of the acquired data;
One or more second methods, which are second processes related to analysis of the acquired data;
A first processing unit that generates one or more combinations;
A second processing unit that executes the combination on the acquired data and generates analysis result data that is data including the respective execution results of the combination;
A third processing unit that generates analysis performance data that is data indicating analysis performance of each of the combinations by collating the analysis result data with the actual event data;
, Data analysis system. The data analysis system according to claim 1, wherein
The acquired data includes data of a plurality of groups different in series,
The first processing unit generates one or more combinations of the first method, the second method, and the group;
The second processing unit executes the combination on the acquired data, and generates analysis result data which is data including each execution result of the combination,
The third processing unit collates the analysis result data with the actual event data to generate analysis performance data which is data indicating analysis performance of each of the combinations.
Data analysis system. The data analysis system according to claim 2, wherein
The analysis result data includes an index value indicating the degree of abnormality of the object to be analyzed, which is determined for each of the combinations;
The storage unit stores a threshold for determining the index value.
The second processing unit generates warning presence / absence information which is information indicating presence / absence of abnormality of the analysis target or presence / absence of abnormality sign for each of the combinations by comparing the index value with the threshold value. Include in analysis result data,
Data analysis system. The data analysis system according to claim 3, wherein
The analysis result data includes information indicating an acquisition date and time of the acquired data used when generating the analysis result data,
The actual event data includes information indicating the date and time of occurrence of an event related to a failure that actually occurred for the analysis target,
The third processing unit compares the acquisition date and time of the analysis result data, in which the warning presence / absence information indicates that there is an abnormality or abnormality sign of the analysis target, with the occurrence date and time of the actual event data. Generate the analysis performance data by
Data analysis system. The data analysis system according to claim 4, wherein
The third processing unit compares the acquisition date and time with the occurrence date and performs weighting according to the time between the acquisition date and the occurrence date and time to obtain the analysis performance by calculating the analysis performance. Generate data,
Data analysis system. The data analysis system according to claim 2, wherein
The first method is realized by a program that implements preprocessing of the acquired data, and a parameter that the program refers to when executed.
The second method is realized by a program for realizing an analysis process on the acquired data, and a parameter which the program refers to at the time of execution.
Data analysis system. The data analysis system according to claim 2, wherein
The storage unit stores information specifying the analysis target, information indicating the purpose of analysis of the acquired data, information specifying the combination, and information in which each analysis performance of the combination is associated.
Data analysis system. The data analysis system according to claim 7, wherein
When the object to be analyzed and the purpose to analyze the acquired data are specified, the combination of good performance corresponding to the object to be analyzed and the purpose of analysis specified based on the analysis performance recorded in the storage unit It further comprises a processing unit that selects and outputs one or more
Data analysis system. The information processing apparatus
Acquisition data which is data acquired for an analysis target, and
Actual event data, which is data relating to an event that actually occurred for the subject of analysis;
Memory step,
One or more first methods, which are first processes related to analysis of the acquired data;
One or more second methods, which are second processes related to analysis of the acquired data;
Generating one or more combinations,
Executing the combination on the acquired data, and generating analysis result data which is data including the respective execution results of the combination;
Generating analysis performance data which is data indicating analysis performance of each of the combinations by collating the analysis result data with the actual event data;
Perform data analysis methods. The data analysis method according to claim 9, wherein
The acquired data includes data of a plurality of groups different in series,
The information processing apparatus
Generating one or more combinations of the first method, the second method, and the group;
Executing the combination on the acquired data, and generating analysis result data which is data including the respective execution results of the combination;
Generating analysis performance data which is data indicating analysis performance of each of the combinations by collating the analysis result data with the actual event data;
Perform data analysis method further. The data analysis method according to claim 10, wherein
The analysis result data includes an index value indicating the degree of abnormality of the object to be analyzed, which is determined for each of the combinations;
The information processing apparatus
Storing a threshold for determining the index value;
Generating warning presence / absence information which is information indicating presence / absence of abnormality of the analysis target or presence / absence of abnormality sign for each of the combinations by comparing the index value with the threshold, and including in the analysis result data;
Perform data analysis method further. The data analysis method according to claim 11, wherein
The analysis result data includes information indicating an acquisition date and time of the acquired data used when generating the analysis result data,
The actual event data includes information indicating the date and time of occurrence of an event related to a failure that actually occurred for the analysis target,
The information processing apparatus compares the acquisition date and time of the analysis result data, in which the warning presence / absence information indicates that there is an abnormality or abnormality sign of the analysis object, with the occurrence date and time of the actual event data. Generating the analysis performance data,
Perform data analysis method further. The data analysis method according to claim 12, wherein
The information processing apparatus compares the acquisition date and time with the occurrence date and time, and performs weighting according to the time between the acquisition date and the occurrence date and time to obtain the analysis performance data To generate
Perform data analysis method further. The data analysis method according to claim 10, wherein
The first method is realized by a program that implements preprocessing of the acquired data, and a parameter that the program refers to when executed.
The second method is realized by a program for realizing an analysis process on the acquired data, and a parameter which the program refers to at the time of execution.
Data analysis method. The data analysis method according to claim 10, wherein
Storing the information identifying the analysis target, the information indicating the analysis purpose of the acquired data, the information identifying the combination, and the information associated with each analysis performance of the combination;
Perform data analysis method further. The data analysis method according to claim 15.
When the information processing apparatus specifies the analysis target and the analysis purpose, one of the combinations of good performance corresponding to the specified analysis target and analysis purpose based on the stored analysis performance Step to select and output
Perform data analysis method further.

Images