Information processing device, information processing method, and information processing program

Abstract

To appropriately execute time window segmentation processing even when there is a missing part in a data set.SOLUTION: An information processing device 10 includes a complement part 12a, a creation part 12b, and a deletion part 12c. The complement part 12a complements a missing part of time-series data based on acquisition intervals of the time-series data. The creation part 12b executes segmentation processing on the time-series data including a part complemented by the complement part 12a, and creates a data pair consisting of learning data and prediction target data. In the segmentation processing, the deletion part 12c deletes the data pair created based on the data complemented by the complement part.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an information processing device, an information processing method, and an information processing program. [Background technology]

[0002] Conventionally, there are known devices that assume multivariate time series data, provide analysis and classification functions optimized for it, and create machine learning models for multivariate time series signals (for example, Non-Patent Document 1). This device overlaps the advanced procedures required for multivariate time series analysis and eliminates technical barriers, making it possible for even users with little prior knowledge to create and operate AI models.

[0003] One of the processes incorporated in the aforementioned device is a time window extraction process (for example, Patent Document 1). The time window extraction process is a process that adds constraints for multivariate time series data to the process of extracting obtained signal data at regular intervals, and by applying this process multiple times to a long signal, the aforementioned device can obtain pairs of training data and prediction data in the amount necessary for training an AI model. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2020-057290 [Non-patent literature]

[0005] [Non-Patent Document 1] “Introduction | User Manual”, [online], [Retrieved July 22, 2022], Internet〈https: / / manual.nodeai.io / 〉 Summary of the Invention [Problem to be solved by the invention]

[0006] However, in the above-mentioned conventional technology, when data is missing due to some influence such as the absence of data for a certain period or the inability to obtain data correctly, it is necessary to divide the data at the missing portion and perform time window extraction processing for each portion.In addition, there is also the problem that when there are multiple missing data in the time series data or when some edge cases cannot be extracted correctly when dividing the data, it is necessary to deal with them individually. [Means for solving the problem]

[0007] In order to solve the above-mentioned problems and achieve the object, the information processing device of the present invention is characterized by including a completion unit that completes missing portions of time series data based on an acquisition interval of the time series data, and a creation unit that performs a cutout process on the time series data including the portions completed by the completion unit and creates a data pair consisting of training data and prediction target data. [Effects of the Invention]

[0008] According to the present invention, it is possible to appropriately perform time window extraction processing even when there are missing portions in the time series data. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is a diagram showing an overview of information processing according to an embodiment. [Figure 2] FIG. 2 is a block diagram showing the configuration of the information processing device according to the embodiment. [Figure 3] FIG. 3 is a diagram illustrating an overview of the time window extraction process in the information processing according to the embodiment. [Figure 4] FIG. 4 is a diagram illustrating a specific example of information processing according to the embodiment. [Figure 5] FIG. 5 is a diagram showing a specific example of information processing according to the embodiment. [Figure 6] FIG. 6 is a diagram showing a specific example of information processing according to the embodiment. [Figure 7] FIG. 7 is a flowchart showing the flow of information processing in the information processing device according to the embodiment. [Figure 8] FIG. 8 is a diagram illustrating an example of a computer that executes an information processing program. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, embodiments of an information processing device, an information processing method, and an information processing program according to the present application will be described in detail with reference to the accompanying drawings. Note that the information processing device, the information processing method, and the information processing program according to the present application are not limited to these embodiments.

[0011] [1. System configuration example] FIG. 1 is a diagram illustrating an overview of information processing according to this embodiment. The information processing device 10 complements missing data portions based on the acquisition interval of time-series data. The information processing device 10 then performs a time window extraction process on the complemented time-series data to create data pairs each consisting of learning data and prediction target data. After that, the information processing device 10 deletes data pairs created based on the complemented data from among the created data pairs.

[0012] The information processing device 10 first complements missing portions of the time series data based on the acquisition interval of the time series data. For example, if there is an interval other than the acquisition interval of the time series data in the obtained time series data, the information processing device 10 determines that data is missing at that interval and discovers the missing portions of the data. Then, for the discovered missing portions of the data, the information processing device 10 complements the missing portions of the time series data by inserting values ​​obtained by linear interpolation into existing data for the extended period.

[0013] Next, the information processing device 10 performs a time window extraction process on the interpolated time series data. For example, the information processing device 10 performs a time window extraction process (described later) on the interpolated time series data to create a data pair consisting of learning data and prediction target data.

[0014] Thereafter, the information processing device 10 deletes the data pairs created based on the interpolated data from the created data pairs. For example, the information processing device 10 performs gradation on the values ​​of the flag series assigned to the interpolated data using a step function, and deletes the data pairs according to the values ​​of the flag series after gradation.

[0015] In this way, the information processing device 10 complements missing portions of the time series data, performs time window extraction processing, and then deletes data pairs created based on the complemented data. As a result, even if there are missing portions in the time series data, the information processing device 10 can appropriately perform time window extraction processing and create a data set that includes only normal data pairs.

[0016] 2. Configuration of Information Processing Device 10 Next, the configuration of the information processing device 10 shown in Fig. 1 will be described with reference to Fig. 2. Fig. 2 is a diagram showing an example of the configuration of the information processing device according to the embodiment. As shown in Fig. 2, the information processing device 10 according to the embodiment has a communication unit 11, a control unit 12, and a storage unit 13.

[0017] The communication unit 11 is realized by, for example, a network interface card (NIC), etc. The communication unit 11 is connected to an external terminal via a wired or wireless connection, and transmits and receives information.

[0018] The storage unit 13 is realized by a storage device such as a RAM (Random Access Memory), a hard disk, etc. The storage unit 13 stores data and programs necessary for various processes by the control unit 12, and includes a time-series data storage unit 13a, a post-interpolation data storage unit 13b, a parameter storage unit 13c, a data pair storage unit 13d, and a post-deletion data pair storage unit 13e as units closely related to the present invention.

[0019] The time series data storage unit 13a stores input time series data. For example, the time series data storage unit 13a stores multivariate time series data input from the outside via the communication unit 11. Here, the multivariate time series data refers to a data format in which an arbitrary multivariate signal is converted into discrete values ​​by sampling and associated with time series information.

[0020] The post-complementation data storage unit 13b stores time series data in which missing data portions have been complemented by the complementing unit 12a (described later). For example, the post-complementation data storage unit 13b stores time series data in which the time series has been unified to the acquisition interval by the complementing unit 12a (described later) and missing data portions have been complemented.

[0021] The parameter storage unit 13c stores parameters used in a time window extraction process, which will be described later. For example, the parameter storage unit 13c stores four parameters N (prediction destination), M (window width), L (rounding width), and S (stride width), which are used in the time window extraction process, which will be described later. It is assumed that the four parameters are stored in advance before the information processing according to this embodiment is performed.

[0022] The data pair storage unit 13d stores data pairs created by the creation unit 12b (described later). For example, the data pair storage unit 13d stores data pairs each consisting of learning data and prediction target data, which are created by the creation unit 12b (described later) performing a time window extraction process on time-series data.

[0023] The post-deletion data pair storage unit 13e stores the remaining data pairs after the deletion process is performed by the deletion unit 12c (described later). For example, the post-deletion data pair storage unit 13e stores the remaining data pairs after the deletion unit 12c (described later) deletes the data pairs created based on the complemented data from the data pairs stored in the data pair storage unit 13d.

[0024] The control unit 12 is realized by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like executing various programs stored in a storage device inside the information processing device 10 using RAM as a work area. The control unit 12 is also realized by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). The control unit 12 has a complementing unit 12a, a creating unit 12b, and a deleting unit 12c, and may also have a calculating unit 12d and a setting unit 12e as necessary.

[0025] The complementing unit 12a complements missing portions of the time series data based on the acquisition interval of the time series data. Then, the complementing unit 12a stores the complemented time series data in the complemented data storage unit 13b. For example, the complementing unit 12a finds missing portions of data from the obtained time series data, and for the missing portions of the found data, complements the missing portions of the time series data by inserting values ​​obtained by linear interpolation into existing data for the extended period, and stores the completed data in the complemented data storage unit 13b.

[0026] The acquisition interval of the time series data is the monitoring interval of the target device when creating the time series data, and may be set in advance, or may be notified from the monitoring device of the target device together with the input of the time series data, or the most frequent value calculated by the calculation unit 12d described later may be used as the acquisition interval of the time series data.

[0027] Here, for a section that is an interval other than the acquisition interval of the time series data and is shorter than the acquisition interval, the complementing unit 12a complements the time series data by downsampling so that the time series data is unified to the acquisition interval. Note that the complementing unit 12a uses a method such as linear interpolation or dropout as downsampling.

[0028] Furthermore, when an interval other than the acquisition interval exists in the time series data, the complementing unit 12a may extend or contract the period of the portion other than the acquisition interval so that the interval other than the acquisition interval becomes the acquisition interval, and insert a value obtained by linear interpolation into the data of the extended or contracted period. Then, the complementing unit 12a stores the complemented time series data in the complemented data storage unit 13b.

[0029] For example, if there is an interval other than the acquisition interval of the time series data in the obtained time series data, the completion unit 12a determines that data loss has occurred at that interval and discovers the missing portion of the data.The completion unit 12a then expands or contracts the period of the missing portion of the time series data stored in the time series data storage unit 13a so that the interval other than the acquisition interval becomes the acquisition interval, and inserts a value obtained by linear interpolation into the data of the expanded or contracted period, thereby completing the missing portion of the data.

[0030] As a result, the time series of the obtained time series data is unified to the acquisition interval, so that time series data without any missing parts is created, and the complementing unit 12a stores the complemented time series data in the complemented data storage unit 13b.

[0031] The creation unit 12b performs an extraction process on the time series data including the portion complemented by the complementing unit 12a to create a data pair consisting of learning data and prediction target data. The creation unit 12b then stores the created data pair in the data pair storage unit 13d. For example, the creation unit 12b creates a data pair consisting of learning data and prediction target data from the complemented time series data stored in the complemented data storage unit 13b and the parameters stored in the parameter storage unit 13c by a time window extraction process (to be described later), and stores the data pair in the data pair storage unit 13d.

[0032] Here, the time window cutting process performed by the creation unit 12b will be described with reference to Fig. 3. Fig. 3 is a diagram showing an overview of the time window cutting process in the information processing according to the embodiment. In the time window cutting process, four parameters N (prediction destination), M (window width), L (rounding width), and S (stride width) are used, and as described above, the four parameters are stored in advance in the parameter storage unit 13c before the information processing according to the embodiment is performed.

[0033] The parameter N is a parameter that indicates the prediction destination, and a value N minutes later is estimated. The parameter M is a parameter that indicates the window width, and data for M consecutive minutes is used as the explanatory variable. The parameter L is a parameter that indicates the rounding width, and data with a window width of M is divided into L minutes, and the average value of the data within the divided intervals is calculated. Note that L is set to be a divisor of M. The parameter S is a parameter that indicates the stride width, and a data pair is created every S minutes.

[0034] In the example of Figure 3, in multivariate time series data, one row represents data for each minute, and corresponding data is inserted into each of columns A to E, with the following settings: N (prediction destination) = 2, M (window width) = 8, L (rounding width) = 2, and S (stride width) = 3.

[0035] For example, the creation unit 12b uses data from the past eight minutes from a reference time as explanatory variables as learning data, divides the eight minutes of data into two-minute intervals, and calculates the average value to obtain four values ​​for each column. The creation unit 12b also obtains the data inserted into each column two minutes later as prediction destination data. This creates a data pair consisting of learning data and prediction destination data for the reference time. The creation unit 12b then creates the aforementioned data pair every three minutes.

[0036] The deletion unit 12c deletes data pairs created based on the data complemented by the complementing unit 12a in the cut-out process. Thereafter, the deletion unit 12c stores the remaining data pairs after the deletion in the post-deletion data pair storage unit 13e. For example, the deletion unit 12c determines the data pairs to be deleted from the data pairs stored in the data pair storage unit 13d based on the flag series set in the complemented data, thereby deleting the data pairs created using the data complemented by the complementing unit 12a, and stores the remaining data pairs after the deletion in the post-deletion data pair storage unit 13e.

[0037] Furthermore, the deletion unit 12c may select a data pair to be deleted based on the flag series set by the setting unit 12e. For example, the deletion unit 12c performs gradation on each data of the data pair stored in the data pair storage unit 13d using a step function in which values ​​of the flag series set to the time-series data by the setting unit 12e (described later) are set to 0 for values ​​less than or equal to 0 and set to 1 for values ​​other than 0. Then, if there is any data included in the data pair in which the value of the flag series after gradation is "1", the deletion unit 12c deletes the data pair.

[0038] The calculation unit 12d calculates the most frequent value of the time-series arithmetic progression assigned to the time-series data as the acquisition interval. Then, the calculation unit 12d notifies the completion unit 12a of the calculated most frequent value as the acquisition interval. For example, the calculation unit 12d calculates the time-series arithmetic progression of the time-series data stored in the time-series data storage unit 13a, and calculates the most frequent value of the obtained arithmetic progression. Then, the calculation unit 12d notifies the completion unit 12a of the obtained most frequent value as the acquisition interval.

[0039] The setting unit 12e sets a flag sequence for the data complemented by the complementing unit 12a. For example, the setting unit 12e adds a flag sequence, which is a column that sets "1" to data that has been complemented due to missing data and sets "0" to other data, to the complemented time-series data stored in the complemented data storage unit 13b.

[0040] [3. Specific examples of information processing] 4 to 6, a specific example of information processing including a process of calculating a mode, which is an acquisition interval of time series data, a process of complementing time series data, and a process of setting a flag series, performed by the information processing device 10 will be described. Figures 4 to 6 are diagrams showing a specific example of information processing according to an embodiment.

[0041] First, the calculation unit 12d extracts only the time series of the input time series data and calculates the difference between the values ​​of adjacent time series to calculate an arithmetic progression. In the example of Figure 4, the difference between [0:00:00] and the adjacent [0:00:01] is "1", and the calculation unit 12d similarly calculates an arithmetic progression for each time series. As a result, there are four "1"s and one "6", so the calculation unit 12d calculates the most frequent value to be "1" and notifies the completion unit 12a of this as the acquisition interval.

[0042] Next, the completion unit 12a completes the missing portion of the time series data based on the acquisition interval notified by the calculation unit 12d. In the example of Fig. 5, the part "6" of the arithmetic progression calculated in Fig. 4 is outside the acquisition interval "1", so the completion unit 12a determines that there is missing data between [0:00:01] and [0:00:07].

[0043] Then, the complementing unit 12a extends the period between [0:00:01] and [0:00:07] for the missing portion so that the portion of the arithmetic progression "6" becomes the acquisition interval "1." The complementing unit 12a inserts values ​​calculated by linear interpolation based on the data of [0:00:01] and [0:00:07] into each piece of data from [0:00:02] to [0:00:06] that has been complemented by the extension.

[0044] Thereafter, the setting unit 12e sets a flag sequence for the interpolated data. In the example of Fig. 5, the setting unit 12e sets a flag sequence "1" for the data from [0:00:02] to [0:00:06] interpolated by the interpolation unit 12a, and sets a flag sequence "0" for the other data from [0:00:00] to [0:00:01] and from [0:00:07] to [0:00:10]. Through the processing by the interpolation unit 12a and the setting unit 12e, the time series data in the upper part of Fig. 5 becomes the time series data in the lower part of Fig. 5, and is stored in the interpolated data storage unit 13b.

[0045] Next, the creation unit 12b performs a time window extraction process on the interpolated time series data stored in the interpolated data storage unit 13b to create a data pair. In the example of Fig. 6, the creation unit 12b performs a time window extraction process on the interpolated time series data shown in the upper part of Fig. 6 under the conditions of reference time

[07] , parameter M (window width) = 2, parameter L (rounding width) = 2, and parameter N (prediction destination) = 2. In particular, a change in the value of the flag series related to the subsequent deletion process will be described.

[0046] Since M (window width) = 2, the creation unit 12b uses the data for 8 minutes from the reference time

[07] to

[00] , and since L (rounding width) = 2, divides the 8 minutes of data into 2-minute intervals and calculates the average value for each interval. Through this process, the flag series from

[00] to

[01] is calculated as "0," the flag series from

[02] to

[03] and

[04] to

[05] are calculated as "1," and the flag series from

[06] to

[07] is calculated as "0.5."

[0047] The creation unit 12b performs the same process as described above on the values ​​inserted into each column to create data consisting of the average values ​​of the values ​​in each column for every two minutes. That is, in the example of Fig. 6, the creation unit 12b calculates the average value for each of the values ​​inserted into the four columns, including the flag series, for every two minutes to create data consisting of four average values ​​for every two minutes. As a result, four data sets are created for the eight-minute data.

[0048] Then, the creation unit 12b creates a data pair consisting of the learning data and the predicted data, using the four pieces of data mentioned above as learning data and, since N (prediction destination) = 2, each numerical value inserted in the row

[09] two minutes after the reference time

[07] as the predicted data.

[0049] Thereafter, the deletion unit 12c performs gradation on the created data pairs using a step function on the flag series set by the setting unit 12e, such that values ​​less than or equal to 0 become 0 and other values ​​become 1, and deletes the data pairs based on the gradated flag series. In the example of Fig. 6, the flag series of the data from

[06] to

[07] , which is one of the learning data of the created data pair, is "0.5", a value other than 0, and therefore the flag series is converted to "1" by applying the step function described above.

[0050] Then, since the data pair created by the creating unit 12b includes three pieces of data whose post-gradation flag series is "1", the deleting unit 12c deletes the data pair. By the series of processes described above, even if there is a missing portion in the time-series data, the information processing device 10 can appropriately perform the time window extraction process and delete, from the created data pairs, the data pairs created based on the interpolated data by the flag series set in the interpolated data.

[0051] 4. An example of information processing by an information processing device Next, information processing of the information processing device 10 will be described with reference to Fig. 7. Fig. 7 is a flowchart showing the flow of information processing in the information processing device according to the embodiment. The information processing device 10, for example, accepts input of time-series data (step S101).

[0052] When the information processing device 10 receives the input of time series data (step S101; Yes), the complementing unit 12a complements the missing part of the time series data based on the acquisition interval of the time series data (step S102). On the other hand, when the information processing device 10 does not receive the input of time series data (step S101; No), the information processing device 10 waits until it receives the input of time series data.

[0053] Then, the creating unit 12b performs a cutout process on the interpolated time-series data to create a data pair (step S103). After that, the deleting unit 12c deletes the data pair created based on the interpolated data pair (step S104).

[0054] 5. Effects of the embodiment As described above, the information processing device 10 according to this embodiment includes the completion unit 12a that completes missing portions of time-series data based on the acquisition interval of the time-series data, the creation unit 12b that performs an extraction process on the time-series data including the portions completed by the completion unit 12a and creates data pairs each consisting of learning data and prediction target data, and the deletion unit 12c that deletes the data pairs created based on the data completed by the completion unit 12a in the extraction process.

[0055] The complementing unit 12a finds missing parts of the time series data based on the acquisition interval of the input time series data and complements the missing parts. The creating unit 12b performs a time window extraction process on the complemented time series data to create data pairs consisting of learning data and prediction target data. The deleting unit 12c deletes data pairs created based on the complemented data from the created data pairs.

[0056] As a result, the information processing device 10 can appropriately perform time window extraction processing even when there are missing parts in the time series data by complementing the missing parts of the time series data, and can create a data set containing only normal data by deleting inappropriate data pairs created using the complemented data.

[0057] Furthermore, the calculation unit 12d of the information processing device 10 calculates the most frequent value of the time-series arithmetic progression assigned to the time-series data as the acquisition interval. This allows the information processing device 10 to calculate the acquisition interval of the time-series data from the input time-series data, and even if the acquisition interval of the time-series data is not set in advance, it is possible to complement missing parts of the time-series data using the calculated acquisition interval (mode).

[0058] Furthermore, when an interval other than the acquisition interval exists in the time series data, the complementing unit 12a of the information processing device 10 expands or contracts the period for the part other than the acquisition interval so that the interval other than the acquisition interval becomes the acquisition interval, and inserts a value obtained by linear interpolation into the data of the expanded or contracted period.

[0059] This enables the information processing device 10 to discover missing data portions from the data acquisition interval, and by inserting values ​​obtained by linear interpolation into the missing portions, it is possible to complement the data so that the missing portions match the data acquisition interval.

[0060] Furthermore, the setting unit 12e of the information processing device 10 sets a flag sequence for the data complemented by the complementing unit, and the deleting unit 12c selects the data pair to be deleted based on the flag sequence set by the setting unit. This allows the information processing device 10 to distinguish between complemented data and non-complemented data based on the set flag sequence, thereby achieving the effect of being able to delete a data pair including inappropriate complemented data.

[0061] [6. System Configuration, etc.] Of the processes described in the above embodiments, some of the processes described as being performed automatically can also be performed manually. Alternatively, all or some of the processes described as being performed manually can be performed automatically using known methods. In addition, the information including the processing procedures, specific names, various data, and parameters shown in the above documents and drawings can be changed as desired unless otherwise specified. For example, the various information shown in each drawing is not limited to the information shown in the drawings.

[0062] Furthermore, the components of each device shown in the figure are functional concepts and do not necessarily have to be physically configured as shown. In other words, the specific form of distribution and integration of each device is not limited to that shown, and all or part of them can be functionally or physically distributed and integrated in any unit depending on various loads and usage conditions. Furthermore, all or any part of the processing functions performed by each device can be realized by a CPU and a program analyzed and executed by the CPU, or can be realized as hardware using wired logic.

[0063] 2 may be held in a storage server or the like, rather than being held by the information processing device 10. In this case, the information processing device 10 acquires various pieces of information by accessing the storage server.

[0064] [7. Hardware Configuration] 8 is a diagram showing an example of a hardware configuration. The information processing device 10 according to the embodiment described above is realized by a computer 1000 having a configuration as shown in FIG.

[0065] 8 is a diagram showing an example of a computer that executes a test program. The computer 1000 includes, for example, a memory 1010 and a CPU 1020. The computer 1000 also includes a hard disk drive interface 1030, a disk drive interface 1040, a serial port interface 1050, a video adapter 1060, and a network interface 1070. These components are connected by a bus 1080.

[0066] The memory 1010 includes a ROM (Read Only Memory) 1011 and a RAM 1012. The ROM 1011 stores a boot program such as a BIOS (Basic Input Output System). The hard disk drive interface 1030 is connected to a hard disk drive 1090. The disk drive interface 1040 is connected to a disk drive 1100. A removable storage medium such as a magnetic disk or optical disk is inserted into the disk drive 1100. The serial port interface 1050 is connected to a mouse 1110 and a keyboard 1120, for example. The video adapter 1060 is connected to a display 1130, for example.

[0067] The hard disk drive 1090 stores, for example, an OS (Operating System) 1091, an application program 1092, a program module 1093, and program data 1094. That is, a program that defines each process of the information processing device 10 is implemented as a program module 1093 in which code executable by the computer 1000 is written. The program module 1093 is stored, for example, in the hard disk drive 1090. For example, a program module 1093 for executing the same process as the functional configuration of the information processing device 10 is stored in the hard disk drive 1090. Note that the hard disk drive 1090 may be replaced with an SSD (Solid State Drive).

[0068] Furthermore, setting data used in the processing of the above-described embodiment is stored as program data 1094, for example, in memory 1010 or hard disk drive 1090. Then, CPU 1020 reads program module 1093 and program data 1094 stored in memory 1010 or hard disk drive 1090 into RAM 1012 as necessary and executes them.

[0069] The program module 1093 and program data 1094 are not limited to being stored in the hard disk drive 1090, but may also be stored in, for example, a removable storage medium and read by the CPU 1020 via the disk drive 1100 or the like. Alternatively, the program module 1093 and program data 1094 may be stored in another computer connected via a network (LAN, WAN, etc.). The program module 1093 and program data 1094 may then be read by the CPU 1020 from the other computer via the network interface 1070. [Explanation of symbols]

[0070] 10. Information processing equipment 11 Communications Department 12 Control Unit 12a Complement part 12b Creation Department 12c Deletion 12d Calculation part 12e Settings 13 Storage section 13a Time series data storage unit 13b Complemented data storage section 13c Parameter storage section 13d Data pair storage section 13e Post-deletion data pair storage section

Claims

1. a complementing unit that complements missing portions of the time series data based on an acquisition interval of the time series data; a creating unit that performs a clipping process on the time series data including the portion complemented by the complementing unit, and creates a data pair including learning data including data calculated from the time series data before a reference time and prediction data including the time series data after the reference time; An information processing device comprising:

2. The apparatus further includes a calculation unit that calculates the most frequent value of the time-series arithmetic progression assigned to the time-series data as the acquisition interval.

2. The information processing apparatus according to claim 1, wherein:

3. When an interval other than the acquisition interval exists in the time series data, the complementing unit expands or contracts a period of the portion other than the acquisition interval so that the interval other than the acquisition interval becomes the acquisition interval, and inserts a value obtained by linear interpolation into data of the expanded or contracted period.

2. The information processing apparatus according to claim 1, wherein:

4. The data acquisition device further includes a setting unit that sets a flag sequence for the data complemented by the complementing unit. R 2. The information processing apparatus according to claim 1, wherein:

5. An information processing process executed by an information processing device, a complementing step of complementing missing portions of the time series data based on an acquisition interval of the time series data; a creating step of performing a cutout process on the time series data including the portion complemented by the complementing step to create a data pair including learning data including data calculated from the time series data before a reference time and prediction data including the time series data after the reference time; An information processing method comprising:

6. a complementation step of complementing missing portions of the time series data based on an acquisition interval of the time series data; a creation step of performing a cutout process on the time series data including the portion complemented by the complementation step, and creating data pairs including learning data including data calculated from the time series data before a reference time and prediction data including the time series data after the reference time; An information processing program that causes a computer to execute the above.

Images