Data processor, inter-system coordination system, data processing method, and program

Abstract

To suppress the influence of the deterioration of the accuracy of coordination data by determining the deterioration of the accuracy of the coordination data with an own system using only the coordination data received from a coordination destination system.SOLUTION: A data processor according to an embodiment comprises: a data accumulation part that receives and accumulates time series data corresponding to the coordination data from a coordination destination system whose value changes with a predetermined periodicity; and an abnormality detection part that determines whether or not the accuracy of the received time series data is deteriorated based on whether or not a time series variation of the received time series data is similar to a time series variation of the time series data corresponding to the data normality.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] An embodiment of the present invention relates to a data processing device, an inter-system cooperation system, a data processing method, and a program. [Background technology]

[0002] In a system of systems, in which multiple independent systems interact with each other, the accuracy of the data exchanged between the systems may not be guaranteed due to the independence of each system.

[0003] In an inter-system collaboration system, the management of the local system and the linked system is independent, so the local system generally cannot determine to what extent the linked data passed from the linked system can be trusted. For this reason, in an inter-system linkage system, there is a risk of unintended operations being performed due to the use of inaccurate linkage data.

[0004] For this reason, while conventional technology has achieved improved accuracy of data passed to partner systems, the partner system that received the data was unable to grasp that accuracy. In addition, conventional technologies have been proposed to detect errors in linked data, but in order to do so, it is necessary to obtain additional information in addition to the linked data that was originally exchanged. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2010-273363 A [Patent Document 2] JP 2014-187914 A [Patent Document 3] Patent No. 7166395 [Patent Document 4] Patent No. 7034038 [Patent Document 5] Patent No. 5813317 Summary of the Invention [Problem to be solved by the invention]

[0006] In a system-to-system collaboration system, when data from a linked system that is not managed by the local system is used for the operation of the local system, in order to prevent the local system from referring to data with low accuracy and performing unintended operations, it is desirable for the linked system that passes on the data to strive to improve accuracy, as with the technology described in Patent Document 1, while the local system that receives the data also detects a deterioration in the accuracy of the linked data and processes it by correcting or discarding values. However, when implementing such a function in an existing system later, it is not always possible to obtain new data from the linked system to detect deterioration in the accuracy of the linked data.In addition, considering the cost and risk of modifying the system, it is desirable to avoid large-scale changes such as adding linked data as much as possible. For this reason, it has been desired to reduce the impact of deterioration in accuracy of the linked data while using only the linked data that has already been exchanged.

[0007] The present invention has been made in consideration of the above, and aims to provide a data processing device, an inter-system linkage system, a data processing method, and a program that are capable of determining a deterioration in accuracy of linked data in its own system using only the linked data received from the linked system, and suppressing the effects of the deterioration in accuracy of the linked data. [Means for solving the problem]

[0008] A data processing device according to an embodiment includes a data storage unit that receives and stores time series data corresponding to linked data from a linked system, the linked data being data whose values ​​change with a predetermined periodicity, and an anomaly detection unit that determines whether or not the accuracy of the received time series data has deteriorated based on whether or not time series changes in the received time series data are similar to time series changes in time series data corresponding to normal data. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 is a schematic configuration diagram of a linkage system group according to an embodiment. [Diagram 2] FIG. 2 is a schematic configuration block diagram of the system of the first embodiment. [Diagram 3] FIG. 3 is a diagram illustrating the general operation of the data accuracy deterioration detection unit. [Figure 4] FIG. 4 is a schematic block diagram of the system of the second embodiment. [Diagram 5] FIG. 5 is a diagram illustrating the classification of the data change classification function. [Figure 6] FIG. 6 is a process flowchart of the second embodiment. [Figure 7] FIG. 7 is a schematic block diagram of the system of the third embodiment. [Figure 8] FIG. 8 is an explanatory diagram of the accuracy deteriorating section identifying process. [Figure 9] FIG. 9 is a diagram for explaining the accuracy degradation data processing in the first example. [Figure 10] FIG. 10 is a diagram for explaining the accuracy degradation data processing in the second example. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Next, an embodiment will be described with reference to the drawings. FIG. 1 is a schematic configuration diagram of a linkage system group according to an embodiment. For ease of understanding, FIG. 1 illustrates an example in which a linked system group is made up of two systems, a local system and a linked system, but the number of linked systems is not limited to this.

[0011] In the following explanation, "precision" refers to the degree of agreement with the true value of the measured quantity, including the accuracy and precision of the measurement result, based on the definition in the Japanese Industrial Standards. Therefore, highly accurate data refers to highly reliable data. In the following description, "deterioration in accuracy" refers to a decrease in the degree of agreement with the true value of the measurand, and data with deteriorating accuracy refers to data whose reliability has decreased due to a decrease in the degree of agreement with the true value of the measurand.

[0012] The linkage system group 10 includes a local system 11, a first linkage destination system 12-1, and a second linkage destination system 12-2. In the above configuration, the local system 11, the first destination linkage system 12-1, and the second destination linkage system 12-2 exchange linkage data, but operate independently of each other.

[0013] The local system 11 controls the local system 11 using the link data CD10 received from the first link destination system 12-1 and the link data CD20 received from the second link destination system 12-2.

[0014] Furthermore, the local system 11 transmits the generated linked data CD01 to the first destination system 12-1, and transmits the generated linked data CD02 to the second destination system 12-2.

[0015] Similarly, the first destination system 12-1 controls the first destination system 12-1 using the link data CD01 received from its own system 11 and the link data CD21 received from the second destination system 12-2. Furthermore, first destination system 12-1 transmits the generated link data CD12 to second destination system 12-2.

[0016] Moreover, the second destination system 12-2 controls the second destination system 12-2 using the link data CD02 received from the system 11 itself and the link data CD12 received from the first destination system 12-1. Furthermore, second destination system 12-2 transmits the generated link data CD21 to first destination system 12-1.

[0017] In the above configuration, the local system 11, the first destination system 12-1, and the second destination system 12-2 have the same basic configuration, so the configuration of the local system 11 will be described below. [1] First embodiment FIG. 2 is a schematic configuration block diagram of the system of the first embodiment. The local system 11 includes a local system operation unit 21, a data storage unit 22, a data accuracy deterioration detection unit 23, and an output unit 24.

[0018] The local system operation unit 21 acquires link data CD10, CD20 from the first link destination system 12-1 or the second link destination system 12-2, which are link destination systems, and performs control operations of the local system 11.

[0019] Furthermore, the local system operation unit 21 generates a future value prediction model etc. 21A based on the time series data stored in the data storage unit 22, and also performs feedforward control. The data storage unit 22 stores the acquired linked data CD10, CD20 as time series data CDT10, CDT20, respectively.

[0020] In this case, the values ​​of the linked data CD10 and the linked data CD20 have a predetermined periodicity, and the time series data CDT10 and the time series data CDT20 each show a similar periodic pattern. For example, when the time series data CDT10 and the time series data CDT20 are the amounts of power used in each system, the values ​​show similar increase and decrease patterns depending on the season and the time of day.

[0021] The data accuracy deterioration detection unit 23 monitors the time series data CDT10 and the time series data CDT20, detects deterioration of the linked data CD10 or the linked data CD20, and notifies the local system operation unit 21, or notifies the administrator of the local system 11 via the output unit 24. Here, the functional configuration of the data accuracy deterioration detection unit 23 will be described. The data accuracy deterioration detection unit 23 has a model creation function 23A, a normal model 23B, and an abnormality detection function 23C. At a predetermined timing, the model creation function 23A creates and updates the next normal model 23B based on multiple time series data CDT10, CDT20 stored in the data storage unit 22, based on which it has been detected that no deterioration in accuracy has occurred based on the previously created normal model 23B. The anomaly detection function 23C detects a deterioration in accuracy of the time series data based on the input time series data CDT10-Y and the normal model 23B as an anomaly, and notifies the local system operation unit 21 and also notifies the administrator of the local system 11 via the output unit 24.

[0022] Here, the operation of the data accuracy deterioration detection unit 23 will be described. FIG. 3 is a diagram illustrating the general operation of the data accuracy deterioration detection unit. In the example of FIG. 3, a case will be described in which the normal model 23B is generated and deterioration of data accuracy is detected based on the time-series data CDT10 corresponding to the first link destination system 12-1. The data accuracy deterioration detection unit 23 generates a normal model 23B based on a plurality of (three in the example of FIG. 3) aligned time-series data CDT10-1 to CDT10-3 using a model creation function 23A. For example, the weight of newer time series data (eg, time series data CDT10-3 in the above example) is increased, and an average (weighted average) is generated as the normal model 23B. Next, a method for detecting deterioration in accuracy in the first embodiment will be described. The time series data CDT10-Y stored in the data storage unit 22 is divided into blocks at equal intervals. The intervals are preferably set to match the fluctuation period of the time series data to be detected, and for example, the intervals may be one day for time series data on the amount of electricity used in a certain house. Of the divided blocks, the block for which data up to the end of the block has been acquired and which is the newest in the time series becomes the time series data of the monitored block.

[0023] In the example of FIG. 3, the time series data of the monitored block is time series data CDT10-4. The data accuracy deterioration detection unit 23 compares this time series data CDT10-4 with the time series changes of the previously generated normal model 23B, and determines that the data accuracy has deteriorated if the similarity is below a certain value (the similarity is lower than a predetermined standard similarity and can be considered to be dissimilar). The result of the determination by the data accuracy deterioration detection unit 23 is notified to the administrator of the local system 11 via the output unit 24, and is also notified to the local system operation unit 21. Therefore, since the obtained time series data CDT10-4 has low accuracy and is not suitable for use in controlling the local system 11, the local system operation unit 21 applies the time series data CDT10-4 with reduced accuracy to the future value prediction model, etc. 21A and does not use it for control. That is, since time-series data with low data accuracy due to deterioration of data accuracy is not used for control, control can always be performed using time-series data with high data accuracy (associated data).

[0024] As described above, according to the first embodiment, it is possible to prevent the local system operation unit 21 of the local system 11 from performing unintended operations based on inaccurate linked data (in the example of Figure 3, time series data CDT10-4) upon detection of a deterioration in data accuracy.

[0025] In this case, by detecting deterioration in data accuracy based only on the time series data received from the linked systems, it is possible to eliminate the constraints of linked systems 12-1, 12-2, such as the ability to provide specific data for detecting deterioration in data accuracy, and it is possible to detect deterioration in accuracy of linked data provided from a larger number of linked systems.

[0026] In addition, since there is no need to provide additional linked data to detect deterioration in data accuracy, there is no need to make major modifications to the linked systems 12-1, 12-2 and the system itself 11, and it is possible to retrofit an accuracy deterioration detection function to an existing system relatively easily.

[0027] [2] Second embodiment FIG. 4 is a schematic block diagram of the system of the second embodiment. In FIG. 4, the same parts as those in FIG. 2 are denoted by the same reference numerals. The local system 11X of the second embodiment includes a local system operation unit 21, a data accumulation unit 22, a data accuracy deterioration detection unit 23X, an output unit 24, and an input unit 25.

[0028] The local system operation unit 21 and the data storage unit 22 are the same as those in the first embodiment, and therefore the detailed explanation thereof will be used.

[0029] The data accuracy deterioration detection unit 23X monitors the time series data CDT10 and the time series data CDT20, detects deterioration of the linked data CD10 or the linked data CD20, and notifies the local system operation unit 21, or notifies the administrator of the local system 11 via the output unit 24.

[0030] Furthermore, the data with degraded accuracy detected is processed (corrected, etc.) based on instructions from the administrator of the local system 11 via the input unit 25, and the processed time-series data is provided to the local system operation unit.

[0031] Here, the functional configuration of the data accuracy deterioration detection unit 23X will be described. The data accuracy deterioration detection unit 23 has a model creation function 23A, a normal model 23B, an abnormality detection function 23C, and a data change classification function 23D.

[0032] At a predetermined timing, the model creation function 23A creates and updates the next normal model 23B based on multiple time series data CDT10, CDT20 stored in the data storage unit 22, based on which it has been detected that no deterioration in accuracy has occurred based on the previously created normal model 23B.

[0033] The anomaly detection function 23C detects a deterioration in accuracy of the time series data CDT10-Y based on the input time series data CDT10-Y and the normal model 23B as an anomaly, notifies the system operation unit 21, and also notifies the administrator of the system 11 via the output unit 24.

[0034] The data change classification function 23D classifies the type of data change (accuracy deterioration) for the time series data CDT10-Y in which accuracy deterioration has been detected, relative to normal data (e.g., normal model 23B), based on the detection result of the anomaly detection function 23C, and notifies the administrator of the system 11 via the output unit 24.

[0035] Here, the operation of the data accuracy deterioration detection unit 23 will be described with reference to FIGS. In the example of FIG. 4 as well, a case will be described in which the normal model 23B is generated and deterioration of data accuracy is detected based on the time-series data CDT10 corresponding to the first linkage destination system 12-1.

[0036] The data accuracy deterioration detection unit 23 generates a normal model 23B based on a plurality of (three in the example of FIG. 3) aligned time-series data CDT10-1 to CDT10-3 using a model creation function 23A. For example, the weight of newer time series data (eg, time series data CDT10-3 in the above example) is increased, and an average (weighted average) is generated as the normal model 23B.

[0037] Next, a classification method of the data change classification function in the second embodiment will be described. FIG. 5 is a diagram illustrating the classification of the data change classification function. FIG. 5A is an explanatory diagram of an example of normal data of time-series data. FIG. 5B is an explanatory diagram of an example of abnormal data type 1 of time-series data.

[0038] The abnormal data type 1 is abnormal data that occurs when the time resolution of the linked data in the linked system 12-1 or the linked system 12-1 deteriorates. The cause of abnormal data type 1 is a decrease in the number of pieces of linked data constituting the time-series data due to a decrease in the processing performance of linked system 12-1 or linked system 12-1.

[0039] FIG. 5C is an explanatory diagram of an example of abnormal data type 2 of time-series data. The abnormal data type 2 is abnormal data that occurs when the resolution of the data value of the linked data in the linked system 12-1 or the linked system 12-1 deteriorates. Causes of abnormal data type 2 include a deterioration in the sensor sensitivity in the linked system 12-1 or the linked system 12-1, which makes it easier for the difference between adjacent linked data that make up the time series data to become a constant value.

[0040] FIG. 5D is an explanatory diagram of an example of abnormal data type 3 of time-series data. The abnormal data type 3 is abnormal data that occurs when noise is mixed into the linked data in the linked system 12-1 or the linked data in the linked system 12-1. Possible causes of abnormal data type 3 include the occurrence of a noise source in the linked system 12-1 or in the linked system 12-2, and the occurrence of a peak in which the difference between the data of adjacent linked data that make up the time series data is significantly large.

[0041] Here, the operation of the data accuracy deterioration detection unit 23X will be described. The data precision deterioration detection unit 23X generates a normal model 23B based on a plurality of (three in the example of FIG. 3) aligned time series data CDT10-1 to CDT10-3 using the model creation function 23A, in the same manner as in the first embodiment.

[0042] Next, a method for detecting deterioration in accuracy in the second embodiment will be described. FIG. 6 is a process flowchart of the second embodiment. The time series data CDT10-Y of the monitored object stored in the data storage unit 22 is divided into blocks at equal intervals.

[0043] The abnormality detection function 23C of the data accuracy deterioration detection unit 23X compares the input time series data CDT10-4 with the time series changes of the previously generated normal model 23B, and determines that the data accuracy has deteriorated if the similarity is below a certain value (if the similarity is lower than a predetermined standard similarity and can be considered to be dissimilar).

[0044] The result of the determination by the data accuracy deterioration detection unit 23 is notified to the administrator of the local system 11 via the output unit 24, and is also notified to the local system operation unit 21 and the data change classification function 23D (step S11). The data change classification function 23D measures the number of pieces of linked data in the time-series data CDT10-Y (step S12).

[0045] Next, the data change classification function 23D judges whether the number of data differs from the number of data in the linked data constituting the normal model (step S13). In this case, the number of data is judged to match not only when the number of data is completely the same, but also when it is within a predetermined range.

[0046] In the judgment of step S13, if the number of data in the linked data constituting the normal model is different from the number of data in the linked data (step S13; Yes), the data change classification function 23D classifies the time series data of the monitored object as abnormal data type 1 of time series data (step S17), and transitions the processing to step S21.

[0047] In the judgment of step S13, if the number of data in the linked data constituting the normal model matches the number of data (step S13; No), the data change classification function 23D measures the difference between adjacent data for the linked data constituting the time series data and plots it as time series data (step S14).

[0048] As a result of the process in step S14, data change classification function 23D compares the plot of normal model 23B and determines whether the distribution of values ​​is biased (step S15). In the judgment of step S15, if the distribution of values ​​is biased compared to the plot of the normal model 23B (step S15; Yes), the data change classification function 23D classifies the time series data of the monitored object as abnormal data type 2 of time series data (step S18), and transitions the processing to step S21.

[0049] In the judgment of step S15, if the distribution of values ​​is not biased compared to the plot of the normal model 23B (step S15; No), the data change classification function 23D determines whether or not there is a peak that is significantly larger compared to the normal model 23B (step S16).

[0050] In the judgment of step S16, if there is a peak that is significantly larger than the normal model 23B (step S16; Yes), the data change classification function 23D classifies the time series data of the monitored object as abnormal data type 3 of time series data (step S19), and transitions the processing to step S21.

[0051] In the judgment of step S16, if there is no peak that is significantly larger than the normal model 23B (step S16; No), the data change classification function 23D classifies the abnormal data type of the time series data of the monitored object into other types (step S20). Next, the data change classification function 23D notifies the administrator of its own system 11 of the change type of the classified time-series data via the output unit 24 (step S21).

[0052] As described above, according to the second embodiment, in addition to the effects of the first embodiment, upon detection of a deterioration in data accuracy, the local system operation unit 21 of the local system 11 classifies and notifies the cause of the inaccurate linked data, making it possible to easily grasp and respond to abnormal conditions in the linked system. [3] Third embodiment FIG. 7 is a schematic block diagram of the system of the third embodiment. In FIG. 7, the same parts as those in FIG. 4 are denoted by the same reference numerals. The local system 11Y of the third embodiment includes a local system operation unit 21, a data accumulation unit 22, a data accuracy deterioration detection unit 23Y, an output unit 24, and an input unit 25.

[0053] The local system operation unit 21 and the data storage unit 22 are similar to those in the first embodiment, and therefore the detailed description thereof will be used.

[0054] The data accuracy deterioration detection unit 23Y monitors the time series data CDT10 and the time series data CDT20, detects deterioration of the linked data CD10 or the linked data CD20, and notifies the local system operation unit 21, or notifies the administrator of the local system 11 via the output unit 24.

[0055] Furthermore, the data with degraded accuracy detected is processed (corrected, etc.) based on instructions from the administrator of the local system 11 via the input unit 25, and the processed time-series data is provided to the local system operation unit.

[0056] Here, the functional configuration of the data accuracy deterioration detection unit 23Y will be described. The data accuracy degradation detection unit 23 has a model creation function 23A, a normal model 23B, an abnormality detection function 23C, a data change classification function 23D, an accuracy degradation section identification function 23E, and an accuracy degradation data processing function 23F.

[0057] At a predetermined timing, the model creation function 23A creates and updates the next normal model 23B based on multiple time series data CDT10, CDT20 stored in the data storage unit 22, based on which it has been detected that no deterioration in accuracy has occurred based on the previously created normal model 23B.

[0058] The anomaly detection function 23C detects a deterioration in accuracy of the time series data CDT10-Y based on the input time series data CDT10-Y and the normal model 23B as an anomaly, notifies the system operation unit 21, and also notifies the administrator of the system 11 via the output unit 24.

[0059] The data change classification function 23D classifies the type of data change (accuracy deterioration) for the time series data CDT10-Y in which accuracy deterioration has been detected, relative to normal data (e.g., normal model 23B), based on the detection result of the anomaly detection function 23C, and notifies the administrator of the system 11 via the output unit 24.

[0060] Based on the detection result of the anomaly detection function 23C, the accuracy degradation section identification function 23E identifies the accuracy degradation section where accuracy degradation has been detected and notifies the accuracy degradation data processing function 23F, and also notifies the administrator of the system 11 via the output unit 24.

[0061] The accuracy degradation data processing function 23F processes (corrects) the time series data CDT10-Y corresponding to the accuracy degradation section where accuracy degradation has been detected and which is based on instructions from the administrator of the local system 11 input via the input unit 25 and which is identified by the accuracy degradation section identification function 23E, and outputs the corrected time series data to the local system operation unit 21 for operation control.

[0062] Here, the operation of the data accuracy deterioration detection unit 23Y will be described. The data precision deterioration detection unit 23 generates a normal model 23B using a model creation function 23A based on the plurality of aligned time series data CDT10-1 to CDT10-3.

[0063] The abnormality detection function 23C of the data accuracy deterioration detection unit 23 compares the time series data CDT10-4 with the time series changes of the previously generated normal model 23B, and determines that the data accuracy has deteriorated if the similarity is below a certain value (if the similarity is lower than a predetermined standard similarity and can be considered to be dissimilar).

[0064] The result of the determination by the data accuracy deterioration detection unit 23 is notified to the administrator of the local system 11 via the output unit 24, and is also notified to the local system operation unit 21, the data change classification function 23D, and the accuracy deterioration section identification function 23E.

[0065] Therefore, since the obtained time series data CDT10-4 has low accuracy and is not suitable for use in controlling the local system 11, the local system operation unit 21 applies the time series data CDT10-4 with reduced accuracy to the future value prediction model, etc. 21A and does not use it for control.

[0066] That is, since time-series data with low data accuracy due to deterioration of data accuracy is not used for control, control can always be performed using time-series data with high data accuracy (associated data).

[0067] Moreover, in the second embodiment, the data change classification function 23D performs classification according to the procedure shown in FIG. 6, and notifies the administrator of its own system 11Y via the output unit 24 of the type of abnormal data.

[0068] In parallel with the execution of the data change classification function, the accuracy deteriorating section identifying function 23E identifies an accuracy deteriorating section in which accuracy deterioration has been detected, based on the detection result of the abnormality detection function 23C.

[0069] Here, the process of the accuracy deteriorating section specifying function 23E will be described. FIG. 8 is an explanatory diagram of the accuracy deteriorating section identifying process. When the anomaly detection function 23C detects a deterioration in accuracy of the time series data CDT10-Y of the monitored object, the time series data CDT10-Y of the monitored object as abnormal data shown in FIG. 8(A) and the normal model 23B shown in FIG. 8(B) are divided into equally spaced blocks BL, and the similarity of the time series changes in each block BL is measured.

[0070] The measurement results are arranged in chronological order, and blocks BL where the similarity has significantly decreased (below the threshold value Th in FIG. 8(C)) are designated as sections of deteriorated accuracy BLL. In addition, for blocks BL at both ends of the accuracy degradation section BLL, by repeatedly dividing the blocks BL into smaller blocks and measuring the similarity, a more accurate accuracy degradation section BLL can be obtained.

[0071] Then, the accuracy deteriorating section specifying function 23E notifies the manager of its own system 11Y of the specified accuracy deteriorating section BLL via the output unit 24, and also notifies the accuracy deteriorating data processing function 23F.

[0072] As a result, when an instruction to process the accuracy-deteriorating data is given by the administrator of the local system 11Y to the accuracy-deteriorating data processing function 23F, the accuracy-deteriorating data processing function 23F processes the time-series data for the accuracy-deteriorating section.

[0073] Here, the process of the accuracy degradation data processing function 23F will be described. FIG. 9 is a diagram for explaining the accuracy degradation data processing in the first example. FIG. 9 shows an example of processing for the abnormal data type 1 described above.

[0074] When the abnormal data is abnormal data type 1 as shown in FIG. 9(A), the abnormal data is processed by sequentially connecting adjacent vertices with different values ​​from a certain vertex in the abnormal data CDTF before processing. As a result, as shown in FIG. 9B, it is possible to obtain processed time-series data CDTA whose values ​​change more gently than before processing, making it possible to obtain data with higher accuracy.

[0075] FIG. 10 is a diagram for explaining the accuracy degradation data processing in the second example. FIG. 10 shows an example of processing for the abnormal data type 3 described above.

[0076] When the abnormal data is abnormal data type 3 as shown in FIG. 10(A), the adjacent data difference is calculated for the abnormal data CDTF before processing as shown in FIG. 10(B), and the adjacent data difference data CDTD is calculated.

[0077] Then, the abnormal data is processed by deleting extremely large peaks and connecting the data before and after them. As a result, as shown in FIG. 10(C), it is possible to obtain processed time-series data CDTA whose values ​​change more gradually than before processing, making it possible to obtain data with higher accuracy.

[0078] As described above, according to the third embodiment, when a deterioration in data accuracy is detected, the degraded data is processed to produce processed data whose accuracy is closer to that of the original data, thereby preventing the local system operation unit 21 of the local system 11 from performing unintended operations by referring to the linked data with low accuracy.

[0079] Even in this case, by detecting deterioration in data accuracy based solely on the time series data received from the linked systems and processing the data based solely on the time series data received from the linked systems, it is possible to eliminate the constraints of linked systems 12-1, 12-2, such as the ability to provide specific data for detecting deterioration in data accuracy, and it is possible to detect deterioration in the accuracy of linked data provided from a larger number of linked systems and process it to deal with the problem.

[0080] In addition, since there is no need to provide additional linked data for detecting deterioration in data accuracy and processing data with deteriorating accuracy, there is no need to make major modifications to the linked systems 12-1, 12-2 and the system itself 11, and it is possible to retrofit existing systems with a function for detecting deterioration in accuracy and a function for processing data with deteriorating accuracy relatively easily.

[0081] [4] Modifications of the embodiment In the above explanation, when linked data with degraded accuracy is received, a request to resend the linked data is not made to the linked system, but it is also possible to configure the system to make a resend request so as to automatically re-acquire the time series data for the identified time series section with degraded accuracy. It is also possible to configure the system so that the time series data of the time series section with deteriorated accuracy is not used or is automatically discarded. Furthermore, when the accuracy of the data processed by the deteriorated accuracy data processing function is insufficient, the system can be configured so that the processed time series data is not used or is automatically discarded.

[0082] In addition, when presenting time series data information to an administrator of the system, it is also possible to present time series data with a predetermined accuracy, time series data with reduced accuracy, and processed time series data in a visually distinguishable manner (for example, by adding labels, changing display colors, or changing the display mode). Similarly, in order to enable the system operation unit to distinguish between time series data with a predetermined accuracy, time series data with reduced accuracy, and processed time series data, it is also possible to assign a label (identification data) to each data, enabling automatic processing.

[0083] The system of this embodiment functions as a data processing device, and can be configured as hardware using a normal computer equipped with, for example, a control device such as a CPU, a storage device such as a ROM (Read Only Memory) or RAM, an external storage device such as an HDD or SSD, a display device such as a display device, and input devices such as a keyboard and a mouse.

[0084] The programs executed by the system of this embodiment are provided as files in an installable or executable format recorded on a computer-readable recording medium such as a USB memory, a semiconductor memory device such as an SSD, or a DVD (Digital Versatile Disk).

[0085] The program executed by the system of this embodiment may be stored on a computer connected to a network such as the Internet and provided by downloading it via the network. The program executed by the device of this embodiment may be provided or distributed via a network such as the Internet. The program for the system of this embodiment may be provided in a state where it is pre-installed in a ROM or the like.

[0086] Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and spirit of the invention, and are included in the scope of the invention and its equivalents described in the claims. [Explanation of symbols]

[0087] 10. Linked Systems 11, 11X, 11Y Own system 12-1 First partner system 12-2 Second Collaborative System 12 Partner Systems 21 System operation unit 21A Future value prediction models, etc. 22 Data Storage Section 23, 23X, 23Y Data accuracy deterioration detection unit 23A Model Creation Function 23B Normal model 23C Anomaly detection function (anomaly detection section) 23D Data change classification function (Data change classification part) 23E Accuracy deterioration section identification function (Accuracy deterioration section identification part) 23F Accuracy degradation data processing function (Accuracy degradation data processing part) 24 Output section 25 Input section BL Block BLL Accuracy deterioration section CDT10-X Collaborative Data CDT10-Y time series data CDTA time series processed data CDTD Adjacent data difference data CDTF Anomaly Data

Claims

1. a data storage unit that receives and stores time-series data corresponding to the linked data from a linked system, the time-series data being changed with a predetermined periodicity; an anomaly detection unit that determines whether or not the accuracy of the received time series data has deteriorated based on whether or not a time series change in the received time series data is similar to a time series change in time series data corresponding to a normal data state; A data processing device comprising:

2. A data change classification unit that classifies a type of data change based on a pattern of the time series change of the received time series data when the accuracy of the received time series data deteriorates.

2. A data processing apparatus according to claim 1.

3. When the data change classification unit classifies the type of data change, an output unit presents candidates for causes of the deterioration of the accuracy.

3. A data processing apparatus according to claim 2.

4. an accuracy degradation section identifying unit that identifies an accuracy degradation section of the received time series data based on a pattern of the time series change of the received time series data; 2. A data processing apparatus according to claim 1.

5. 3. The data processing device according to claim 2, further comprising a degraded accuracy data processing unit that processes the time series data with degraded accuracy in accordance with a classification of the type of data change based on a pattern of the time series change in the received time series data.

6. When presenting information about the time series data, the time series data having a predetermined accuracy, the time series data having a deteriorated accuracy, and the processed time series data are presented in a visually distinguishable manner.

6. A data processing apparatus according to claim 5.

7. A system for linking systems that is composed of multiple independent systems, The independent system includes a data storage unit that receives and stores time-series data corresponding to linked data whose values ​​change with a predetermined periodicity input from another system; an anomaly detection unit that determines whether or not the accuracy of the received time series data has deteriorated based on whether or not a time series change in the received time series data is similar to a time series change in time series data corresponding to normal data; An inter-system collaboration system equipped with the above.

8. A data processing method executed in a data processing device that processes linked data from a linked system, comprising: the data processing device receiving and storing time series data corresponding to the linked data whose value changes with a predetermined periodicity; determining whether the accuracy of the received time series data has deteriorated based on whether a time series change in the received time series data is similar to a time series change in time series data corresponding to normal data; A data processing method comprising:

9. A program for controlling a data processing device that processes linked data from a linked system by a computer, The computer, A means for receiving and storing time series data corresponding to the linked data whose value changes with a predetermined periodicity; a means for determining whether or not the accuracy of the received time series data has deteriorated based on whether or not a time series change in the received time series data is similar to a time series change in time series data corresponding to a normal data state; A program that makes it work.

Images