JP7267025B2 - Data processing method and program - Google Patents

Description

One embodiment of the present invention relates to a data processing method and a data processing apparatus.

Normally, when data is input to a system from the outside, it is necessary to process the data in a fixed order through a plurality of systems. For example, Patent Literature 1 discloses a data processing device that receives data transmitted from an external company and processes a series of data via a plurality of systems.

JP-A-2002-352076

As described above, when data is processed through a plurality of systems, each of the plurality of systems is also connected to a plurality of systems, which complicates system connections. Therefore, in order to perform a series of data processing, even if the order of passing through a plurality of systems is fixed, it is necessary to determine the system for performing the next processing in each of the plurality of systems. . Therefore, a lot of labor and time have been spent to process a series of data.

Accordingly, one object of an embodiment of the present invention is to reduce the user's processing burden in a series of data processing.

A data processing method according to an embodiment of the present invention determines the output destination of acquired data based on artificial intelligence that specifies the output destination of the data when arbitrary data is input, and outputs the data. output first.

In the above method, when data output to an output destination and processed by the output destination is input, the output destination of the data is determined, and the data is output to the output destination.

In the above method, determining the output destination of the acquired data includes updating the learning model used for artificial intelligence based on the arbitrary data and the output destination of the acquired data.

In the above method, when an error is acquired from the output destination, the data and the corresponding output destination are received and stored as error data.

In the above method, when an error is obtained from the output destination, updating the learning model used for artificial intelligence based on the data including the predetermined item and the value corresponding to the item, and the correct output destination.

The above method includes updating a learning model used for artificial intelligence based on the predetermined item, the value input to the item, and the added output destination.

According to one embodiment of the present invention, it is possible to reduce the user's processing burden in a series of data processing.

1 is a block diagram showing the configuration of a data processing system according to one embodiment of the present invention; FIG. 1 is a block diagram showing the hardware configuration of a data processing device according to one embodiment of the present invention; FIG. 1 is a functional block diagram of a data processing device according to one embodiment of the present invention; FIG. 1 is a functional block diagram of a system connected to a data processing device according to one embodiment of the present invention; FIG. 4 is a flowchart for explaining a data processing method according to one embodiment of the present invention; 1 is a data flow diagram of a data processing device according to an embodiment of the present invention; FIG. (A) An example of data acquired by a data processing device. (B) An example of data acquired by the data processing device. (A) An example of data acquired by a data processing device. (B) An example of data acquired by the data processing device. 4 is a flowchart for explaining a data processing method according to one embodiment of the present invention; (A) An example of error data acquired by the data processing device. (B) An example of correct data acquired by the data processing device. 1 is a data flow diagram of a data processing device according to an embodiment of the present invention; FIG. It is an example of a data flow processed by each of a plurality of systems. It is an example of data processed by each of a plurality of systems.

An embodiment of the present invention will be described in detail below with reference to the drawings. The embodiments shown below are examples of embodiments of the present invention, and the present invention should not be construed as being limited to these embodiments. That is, it is possible to apply a known technique to the multiple embodiments described below, modify them, and implement them in various modes.

(First embodiment)
In this embodiment, a data processing system 200 according to one embodiment of the present invention will be described with reference to FIGS. 1 to 13. FIG.

[overview]
12 and 13 are diagrams for explaining a series of data processing related to settlement business executed in a financial system. FIG. 12 is a diagram explaining a data flow when a series of data processing is executed via a plurality of systems. FIG. 13 shows an example of data processed by each of a plurality of systems. As shown in FIG. 12, data input to the front system 505 is processed through a plurality of systems up to an information system 509 .

As shown in FIG. 12, when data 511 is input to the front system 505, the front system 505 transmits the data 511 to the entry system A506. The entry system A 506 processes the data 511 and transmits the data 512 as data 512 to the external system 201 via the channel system 510 and the communication device 202 . The external system 201 processes the data 512 to generate the data 513 and transmits the AML sanction check 503 via the communication device 202 and the channel system 510 . Next, the entry system A 203 generates data 514 by processing the data 511 and the data 512 and transmits it to the information system 509 .

As explained above, even if the series of data processing flows and the order of the systems used are fixed, as in the case of settlement operations in the banking industry, in order to perform the following processing in each system system must be determined. For example, as shown in FIG. 12, it is necessary to determine to which of the channel system 510 or the information system 509 the data 512 processed by the bookkeeping system A203 should be sent.

Also, as shown in FIG. 12, when the systems are connected in series, such as the front system 505, the entry system A506, and the channel system 510, when maintaining the entry system A203, the front system that cooperates with the entry system A506 The functions of system 505 and channel system 510 must be temporarily stopped. Therefore, the maintainability of the system is degraded. On the other hand, when systems are connected in parallel, when outputting to the next system, it is necessary to determine which system to output to, which consumes a lot of labor and time. There's a problem.

Therefore, in the data processing system according to one embodiment of the present invention, each of a plurality of systems through which a series of data processing passes is connected to a data processing device 210, as shown in FIG. In the data processing device 210, when data is input, the output destination system of the acquired data is determined based on artificial intelligence (AI) that identifies the system to be the output destination, and the data is output. output to the destination system. By repeatedly executing the above processing, even when executing a series of data processing, there is no need to specify the system for performing the next processing in each of the plurality of systems. Therefore, the user's processing load in a series of data processing can be reduced.

In addition, since multiple systems are connected in parallel to the data processing device, even when performing maintenance on one of the multiple systems, the functions of the system linked to that system will be stopped. you don't have to. Therefore, the maintainability of the system is improved.

[Configuration of data processing system]
FIG. 1 is a schematic diagram illustrating the configuration of a data processing system 200 according to one embodiment of the invention. Data processing system 200 includes communication device 202, data processing device 210, and a plurality of system A 203 through system F 208. FIG. A data processing system 200 is connected to an external system 201 via a network.

The external system 201 is, for example, a system of an external organization.

The communication device 202 is a device that relays or transfers data with the external system 201, and has a function of selecting or controlling a data transfer route.

The data processing device 210 is connected to each of a plurality of systems A203 to F208. The data processing device 210 has a function of analyzing data input from any one of the plurality of systems A203 to F208 by artificial intelligence and specifying the output destination system of the data based on the analysis result. Here, artificial intelligence includes an input layer that receives signals from the outside, an output layer that outputs calculation results to the outside, and performs predetermined calculations on the information received from the input layer and outputs the processing results to the output layer. It is a neural network composed of an intermediate layer for output.

In the following description, a system that transmits data to the data processing device 210 is called an input source system, and a system that outputs data is called an output destination system. Further, data acquired from the input source system by the data processing device 210 is called acquired data, and data outputted from the data processing device 210 to the output destination system is called output data.

[Hardware Configuration of Data Processor]
FIG. 2 is a block diagram illustrating the hardware configuration of the data processing device 210 according to one embodiment of the invention. As shown in FIG. 2, the data processing device 210 has a control unit 101, a recording medium 102, a RAM 103, a display unit 104, an operation unit 105, and a communication unit .

The control unit 101 is, for example, a CPU (Central Processing Unit) or the like, and controls operations of function blocks described later. The control unit 101, for example, reads out an operation program recorded in the recording medium 102, develops it in the RAM 103, and executes the operation program, thereby realizing a data processing function by each functional block.

The recording medium 102 is, for example, a recording device such as a non-volatile memory or HDD (Hard Disk Drive) capable of permanently retaining information and rewriting information. The recording medium 102 stores the operation programs of the functional blocks of the data processing device 210 and information such as parameters required for the operation of each block. The recording medium 102 also stores learning models based on data including items and values corresponding to the items, and output destination systems of the acquired data, as will be described in detail later. Note that the operating program of the functional block and the learning model may be stored in different recording media.

The RAM 103 is, for example, a rewritable storage device such as a volatile memory that is used to temporarily hold information. The RAM 103 is used as a development area for the operation programs of the functional blocks recorded on the recording medium 102 and as a storage area for storing intermediate data output in the operation of the functional blocks.

The display unit 104 is, for example, a display device such as a liquid crystal display panel or an EL display panel. The display unit 104 may be a display device that is integrally connected to a housing that constitutes the data processing device 210 . Alternatively, the display unit 104 may be an external display device connected to the data processing device 210 . The display unit 104 displays a display screen regarding an output unit, which is a functional block of the data processing device 210, which will be described later.

The operation unit 105 is a user interface such as buttons, a mouse, a keyboard, and a touch sensor. When the operation unit 105 detects that an operation input has been made to each user interface, the operation unit 105 outputs a control signal corresponding to the operation input made to the control unit 101 .

The communication unit 106 is a communication interface that the data processing device 210 has. Via the communication unit 106, the data processing device 210 can connect to another terminal device via a network and transmit and receive information to and from the other terminal device. For example, if system A 203 to system F 208 are different terminal devices, data processor 210 may be connected to system A 203 to system F 208 via communication unit 106 .

In this embodiment, to simplify the explanation, the hardware configuration of the data processing device 210 is described as having various blocks, but one embodiment of the present invention is not limited to this. For example, at least one of various blocks may be configured as a device separate from the data processing device 210 . Specifically, the recording medium 102 and the display unit 104 may be configured as separate devices. Also, the recording medium 102 may be divided into a plurality of recording media.

[Functional block diagram of data processor]
FIG. 3 is a functional block diagram of the data processing device 210. As shown in FIG. The data processing device 210 includes an acquisition unit 211 , a determination unit 212 , an AI 213 , a learning model 214 , teacher data 215 , an output unit 216 and error data 217 .

The acquisition unit 211 acquires data output from the input destination system 221 . It is preferable that the data acquired by the acquisition unit 211 have a common format even when the data is output from any one of the plurality of systems A203 to F208 shown in FIG. However, the items included in the data and the values corresponding to the items differ depending on the input destination system 221 to which the data is output.

The determination unit 212 analyzes the content of the data acquired by the acquisition unit 211 by the AI 213, generates an analysis result, and outputs the analysis result. The AI 213 analyzes data including items and values corresponding to the items by text mining or image recognition, inputs the data to the input layer, and infers a system that should output the data using a learning model. The AI 213 outputs analysis results including the inferred destination system and accuracy. Accuracy indicates the accuracy of the delivery destination inferred by the AI 213 . For example, accuracy may be the probability that the destination system inferred by AI 213 is correct.

Also, the AI 213 re-learns every predetermined period to update the learning model 214 . The learning model 214 is generated based on the characteristics of the data including the items and the values corresponding to the items based on the past results and the output destination system of the data. AI 213 re-learns (e.g., deep learning) based on teacher data 215 including new learning data (e.g., output destination system of data including items and values corresponding to items) accumulated in a predetermined period. Update the learning model 214 (neural network). This allows the AI 213 to correctly infer new destination systems with the updated learning model 214 that the previous learning model 214 cannot correctly infer. Note that the timing of updating the learning model 214 is not limited to every predetermined period, and may be executed at the timing when a predetermined number of teaching data 215 are accumulated.

The output unit 216 outputs the data to the output destination system 222 specified by the determination unit 212 .

When the data processing device 210 acquires an error from the output destination system 222 , the data processing device 210 receives the data and the corresponding output destination system 222 and accumulates it as error data 217 . The error data 217 accumulates data including items and values corresponding to the items, and correct learning data after correction to the correct output destination system as teacher data 215 . The AI 213 updates the learning model 214 by re-learning based on teacher data 215 including correct learning data accumulated during a predetermined period. This allows the AI 213 to correctly infer the output destination system with the updated learning model 214, which could not be inferred accurately with the previous learning model 214. The timing of updating the learning model 214 with the error data 217 as the correct learning data is not limited to every predetermined period, and may be performed at the timing when a predetermined number of data are accumulated.

[System functional blocks]
Next, functional blocks of the system connected to the data processor 210 will be described with reference to FIG. Although the functional blocks of system A 203 are described in FIG. 4, system B 204 to system F 208 also have similar functional blocks.

System A 203 includes a data receiver 241 , a processor 242 , a data output unit 243 and an error output unit 244 .

The data receiving unit 241 receives data output from the data processing device 210 or another system.

The processing unit 242 processes the data received by the data receiving unit 241 . Values included in the data are converted or extracted by the processing unit 242 .

The data output unit 243 outputs the data processed by the processing unit 242 to the data processing device 210 or another system.

The error output unit 244 outputs an error to the data processing device 210 when the data receiving unit 241 cannot receive data, or when the data receiving unit 241 receives data but the processing unit 242 cannot process the data.

[Data processing method]
Next, a data processing method according to one embodiment of the present invention will be described with reference to FIG. The data processing method has steps S301 to S304. The control unit 101 causes the data processing device 210 to execute each process.

First, the acquisition unit 211 acquires data output from the input source system (step S301). Next, the determination unit 212 analyzes the acquired data by the AI 213 (step S302). Next, based on the analysis result, the output destination system of the acquired data is specified (step S303). Next, the acquired data is output to the specified output destination system (step S304). As described above, the data acquired by the data processing device 210 can be output to the output destination system. When the data processed by the output destination system is input to the data output to the output destination system, the data processing device 210 executes the processing from step S301 to step S304 to output the data to the output destination. Determine the system and output to the output destination system of the data.

[Specific example of a series of data processing in the data processing device]
Next, a specific example in which the data processing device 210 performs a series of data processing will be described with reference to FIGS. 6 to 8. FIG. FIG. 6 is a data processing system 200 in accordance with one embodiment of the present invention, demonstrating the data flow for performing a sequence of data processing. FIG. 8 is an example of data acquired by the acquisition unit 211 of the data processing device 210 . Also, in the following description, the external system 201 is the SWIFT net system, the system A 205 is the AML sanction check, the system C 205 is the front system, and the system F 208 is the information system. As used herein, the SWIFT net system is a network system that provides financial communication messages relating to the processing of international financial transactions between banks.

In this embodiment, as a series of data processing executed by the data processing apparatus 210, a case will be described in which system C 205, system D 206, external system 201, system A 203, and system F 208 are executed in that order.

First, when data is input to system C 205 , data 231 is transmitted from system C 205 to data processor 210 . The data processing device 210 acquires the data 231 in the acquisition unit 211 . Next, the determination unit 212 analyzes the items included in the data 231 and the values corresponding to the items using the AI 213 . In the data 231, the item is, for example, [Date], and the value corresponding to the item is, for example, yyyy/mm/dd. By analyzing the multiple items included in the data 231 and the values corresponding to each of the multiple items by the AI 213, the system to which the data 231 should be output is determined to be the system D206. Next, the data 231 is output from the output unit 216 to the system D206. System D 206 receives data 231 by data receiving section 241 . Next, the data 231 is processed by the processing unit 242 to generate the data 232 . The processing unit 242 may perform data conversion using, for example, the items of the data 231 and the values corresponding to the items. For example, data conversion converts [date] yyyy/mm/dd, [currency] JPY, and [amount] xxxxx yen in the data 231 to [code 1] yyyymmddJPYxxxxx. Next, the data 232 is output from the data output unit 243 to the data processing device 210 .

Next, the data processing device 210 acquires the data 232 by the acquisition unit 211 . Next, the determination unit 212 determines that the system to which the data 232 should be output is the external system 201 by analyzing the items included in the data 232 and the values corresponding to the items using the AI 213 . Next, data 232 is output from the output unit 216 to the external system 201 via the communication device 202 . The external system 201 receives the data 232 through the data receiving section 241 . Next, the processing unit 242 processes the data 232 to generate the data 233 . The processing unit 242 performs data conversion using, for example, the items of the data 232 and the values corresponding to the items. Next, the data 233 is output from the data output unit 243 to the data processing device 210 .

Next, the data processing device 210 acquires the data 233 by the acquisition unit 211 . Next, the determination unit 212 determines that the system to which the data 233 should be output is the system A 203 by analyzing the items included in the data 233 and the values corresponding to the items using the AI 213 . Next, data 233 is output from the output unit 216 to the system A 203 . System A 203 receives data 233 by data receiving section 241 . Next, the processing unit 242 processes the data 233 to generate the data 234 . Next, the data 234 is output from the data output unit 243 to the data processing device 210 .

Next, the data processing device 210 acquires the data 234 by the acquisition unit 211 . Next, the determination unit 212 analyzes the items included in the data 234 and the values corresponding to the items using the AI 213, thereby determining that the system to which the data 234 should be output is the system F208. Next, the data 234 is output from the output unit 216 to the system F208. The system F 208 receives the data 234 by the data receiving section 241 . Thus, a series of data processing executed by the data processing device 210 is completed.

According to the data processing device or data processing method according to one embodiment of the present invention, even if a series of data processing is executed via a plurality of systems, the data acquired by the data processing device 210 can be processed by AI. By analyzing, it is possible to identify and output the output destination system to which the data should be output. Therefore, it is not necessary for each of the plurality of systems to determine the system for performing the next processing. Therefore, the user's processing burden in a series of data processing can be reduced.

The layout of data acquired by the data processing device 210 is not limited to the layouts shown in FIGS. 7 and 8, and may be other items, and the value input format may be different.

(Modification 1)
Although not described in detail in the above embodiment, a case where the data processing device 210 receives an error from the output destination system will be described with reference to FIG. FIG. 9 is a diagram explaining a data processing method when the data processing device 210 receives an error from the output destination system. FIG. 10(A) is the data including the error, and FIG. 10(B) is the correct data with the error corrected. The data processing method has steps S301 to S307. The control unit 101 causes the data processing device 210 to execute each process.

First, the acquisition unit 211 acquires data output from the input source system (step S301). Next, the determination unit 212 analyzes the acquired data by the AI 213 (step S302). Next, based on the analysis result, the output destination system of the acquired data is specified (step S303). Next, the acquired data is output to the specified output destination system (step S304). Next, it is determined whether or not an error has been received from the output destination system (step S305). For example, when an error is received, an error is received from the output destination system within a predetermined time after outputting acquired data to the output destination system, or when the data cannot be received by the output destination system. . If an error is received from the output destination system (step S305; Yes), the error data 217 stores data including the item in error, the value corresponding to the item, and the output destination system. For example, assume that the AI 213 of the determination unit 212 identifies the system F 208 as the output destination system, outputs the data 234A shown in FIG. . The data processor 210 accumulates the data 234A that received the error and the system F 208 as the error data 217. FIG. After that, the error data is corrected to correct data, and the correct data is accumulated as teacher data (step S307). For example, data 234A containing error data shown in FIG. 10A is corrected to correct data 234 shown in FIG. 10B, and the data 234B and system F 208 are stored as teacher data. . If no error is received from the output system (step S305; No), the subsequent processing is terminated.

As described above, even if the data processing device 210 transmits data to the wrong output destination system, the error output from the output destination system is accumulated as the error data 217, and the error data is corrected. By re-learning correct data as teacher data, the accuracy of the decision logic can be improved.

(Modification 2)
Although not described in detail in the above embodiment, a data processing method when a new system is connected to the data processing device 210 will be described with reference to FIG. FIG. 11 is a diagram showing an example of the data flow when the new system G209 is connected to the data processing device 210. As shown in FIG.

As shown in FIG. 11, when a new system is connected to the data processing device 210, the AI 213 of the determination unit 212 needs to perform relearning and update the teacher data DB 213. FIG. In this case, as teacher data 215, data including the system G209, items to be output to the system G209, and values corresponding to the items are used as teacher data 215 for re-learning. As a result, the AI 213 in the determination unit 212 can accurately infer a new output destination system, which could not be inferred accurately with the previous learning model, using the updated learning model 214 .

(Modification 3)
In the embodiment of the present invention, the output destination of acquired data is the system, but the output destination is not limited to the system in the present invention. The output destination of the acquired data is not particularly limited as long as the data can be output, and may be a program, URL, node, e-mail, server, or the like. For example, the learning model is updated by re-learning the acquired data based on the teacher data whose output destination is the server. As a result, the data output destination can be a server as well as the system.

101: control unit, 102: recording medium, 104: display unit, 105: operation unit, 106: communication unit, 200: data processing system, 201: external system, 202: communication device, 203: system A, 204: system B , 205: System C, 206: System D, 207: System E, 208: System F, 209: System G, 210: Data processing device, 211: Acquisition unit, 212: Determination unit, 214: Learning model, 215: Teacher Data, 216: Output unit, 217: Error data, 221: Input destination system, 222: Output destination system, 231: Data, 232: Data, 233: Data, 234: Data, 234A: Data, 234B: Data, 241: Data receiving unit, 242: processing unit, 243: data output unit, 244: error output unit, 503: sanction check, 505: front system, 509: information system, 510: channel system, 511: data, 512: data, 513: data, 514: data

Claims (6)

A computer implemented method comprising:
For a learning model that has learned the relationship between input data in a data format corresponding to an input source system and having a value data format corresponding to a predetermined item, and an output destination system determined in accordance with the data format obtaining information identifying a second system from the learning model by providing first data obtained from the first system, and outputting the first data to the second system;
Obtaining information identifying a third system from the learning model by providing second data converted based on the first data obtained from the second system, and providing the second data to the third system Output , data processing method. After determining the second system based on the first data,
2. The data processing method according to claim 1 , comprising updating said learning model based on said first data and information identifying said second system obtained from said learning model. If an error is obtained from the second system,
2. The method according to claim 1, wherein said first data having said data format corresponding to said first system and said second system determined corresponding to said data format are received and stored as error data. Data processing method. If an error is obtained from the second system,
Updating the learning model using the relationship between the first data having the data format corresponding to the first system and the correct fourth system determined corresponding to the data format. Item 4. The data processing method according to item 3 . When adding a fifth system that cooperates with the fourth system,
Using the relationship between the fourth data having a data format corresponding to the fourth system and having a value data format corresponding to a predetermined item, and the added fifth system determined corresponding to the data format 5. The data processing method of claim 4 , comprising updating the learning model. The data processing method according to any one of claims 1 to 5 ,
A program that makes a computer run.

Images