Data processing system, data processing device, data processing method, and program

Abstract

In order to facilitate the removal of an abnormal value from training data of machine learning for predicting radio quality, this data processing system comprises: a determination unit that determines the occurrence of a handover on the basis of measurement data on radio quality; and a filtering unit that, when a handover occurs, outputs, as training data of machine learning for predicting the radio quality, measurement data obtained by removing data for a prescribed period after the handover from among the measurement data on the radio quality.

Description

Data processing system, data processing device, data processing method, and program

[0001] The present invention relates to a data processing system, a data processing device, a data processing method, and a program.

[0002] There are techniques for switching to another propagation path when wireless quality deteriorates. For example, a technique is known in which the movement of an obstacle is predicted and the switching of the propagation path is controlled based on the prediction result (see, for example, Patent Document 1).

[0003] Japanese Patent Application Laid-Open No. 2020-92386

[0004] Machine learning is used for a variety of purposes, including the prediction of wireless quality. However, if data containing outliers is used for machine learning, the accuracy of machine learning estimation may decrease. Therefore, it is desirable to remove outliers from the acquired data before using it as training data.

[0005] In conventional technology, this data removal was determined based on know-how or thresholds, etc. It is also possible to perform data removal itself using machine learning, but this requires a large amount of training data and manual labeling of the data itself.

[0006] As described above, conventional techniques have had difficulty in removing outliers from learning data for machine learning that predicts wireless quality.

[0007] An embodiment of the present invention has been made in consideration of the above-mentioned problems, and provides a data processing system that can easily remove outliers from learning data for machine learning that predicts wireless quality.

[0008] In order to solve the above problem, a data processing system according to an embodiment of the present invention includes a judgment unit that judges whether a handover will occur based on measurement data of wireless quality, and a filtering unit that, when the handover occurs, outputs measurement data from the measurement data of the wireless quality, from which data for a predetermined period after the handover has been removed, as learning data for machine learning that predicts the wireless quality.

[0009] According to an embodiment of the present invention, it is possible to provide a data processing system that can easily remove outliers from learning data for machine learning that predicts wireless quality.

[0010] FIG. 1 is a diagram illustrating an example of the configuration of a data processing system according to Example 1. FIG. 2 is a diagram illustrating an overview of data processing according to Example 1. FIG. 3 is a flowchart illustrating an example of data processing according to Example 1. FIG. 4 is a diagram illustrating filtering processing according to the present embodiment. FIG. 5 is a diagram illustrating an example of the configuration of a data processing system according to Example 2. FIG. 6 is a flowchart illustrating an example of data processing according to Example 2. FIG. 7 is a diagram illustrating abnormal value detection processing according to Example 2. FIG. 8 is a diagram illustrating an example of the hardware configuration of a computer.

[0011] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The embodiment described below is merely an example, and the embodiment to which the present invention is applied is not limited to the following embodiment.

[0012] 1 is a diagram showing an example of the configuration of a data processing system according to this embodiment. The data processing system 1 includes a data processing device 100 that removes outliers from learning data for machine learning that predicts wireless quality.

[0013] The data processing device 100 outputs, for example, measurement data of wireless quality acquired by the data acquisition unit 10, from which abnormal values ​​have been removed, to the machine learning unit 20 or the like as learning data for machine learning to predict wireless quality.

[0014] The data acquisition unit 10 executes a data acquisition process to acquire measurement data of wireless quality. The data acquisition unit 10 may be included in, for example, a wireless terminal that measures the wireless quality of wireless communication, or in a monitoring system that collects measurement data of wireless quality from wireless terminals. The data acquisition unit 10 may also be included in the data processing device 100.

[0015] The wireless quality measurement data acquired by the data acquisition unit 10 includes, for example, wireless quality data such as throughput and RSRP (Reference Signal Received Power), identification information of the wireless base station such as PCI (Physical Cell ID) and CID (Cell ID), and information on the measurement date and time.

[0016] The machine learning unit 20 executes machine learning processing to learn a prediction model for predicting wireless quality (e.g., throughput) by machine learning using learning data output by the data processing device 100. The machine learning unit 20 is included in, for example, an information processing device or a server device that can communicate with the data processing device 100. Note that the machine learning unit 20 may be provided outside the data processing system 1.

[0017] The data processing device 100 is an information processing device having a computer configuration, or a system including multiple computers. The data processing device 100 realizes each functional configuration, such as a determination unit 101 and a filtering unit 102, by executing a predetermined program on the computer included in the data processing device 100. Note that at least a part of each of the above functional configurations may be realized by hardware.

[0018] The determination unit 101 executes a determination process to determine whether a handover has occurred based on the measurement data of wireless quality acquired by the data acquisition unit 10. For example, the determination unit determines that a handover has occurred when the PCI, CID, or the like included in the measurement data has changed.

[0019] The filtering unit 102 performs a filtering process to remove data for a predetermined period after handover from the wireless quality measurement data acquired by the data acquisition unit 10 and output the resulting measurement data as learning data for machine learning to predict wireless quality.

[0020] (Processing Overview) Fig. 2 is a diagram for explaining an overview of data processing according to Example 1. Fig. 2 shows an example of measurement data 210 that the data processing device 100 acquires from the data acquisition unit 10, and an example of training data 220 that the data processing device 100 outputs.

[0021] In the example of FIG. 2, measurement data 210 indicates the relationship between RSRP and throughput of data 201 immediately after handover and data 202 other than immediately after handover, among measurement data of wireless quality.

[0022] In machine learning, if data containing outliers is used for learning, the accuracy of machine learning estimation may decrease.

[0023] Therefore, the data processing device 100 according to the first embodiment focuses on the cause of deterioration in wireless quality immediately after handover, and removes the data 201 immediately after handover as an abnormal value from the measurement data 210, and outputs the measurement data as the learning data 220. This enables the data processing device 100 to improve the accuracy of machine learning by the machine learning unit 20.

[0024] <Processing Flow> Next, the processing flow of the data processing method according to the first embodiment will be described.

[0025] 3 is a flowchart illustrating an example of data processing according to the embodiment 1. This processing shows an example of data processing executed by the data processing device 100 described with reference to FIG.

[0026] In step S301, the data processing device 100 acquires measurement data of wireless quality acquired by the data acquisition unit 10.

[0027] In step S302, the determination unit 101 determines whether a handover has occurred based on the acquired measurement data of wireless quality. For example, the determination unit determines that a handover has occurred when information about the wireless base station, such as PCI (or CID), included in the measurement data has changed. If a handover has occurred, the determination unit 101 proceeds to step S303. On the other hand, if a handover has not occurred, the determination unit 101 proceeds to step S304.

[0028] In step S303, the filtering unit 102 removes data for a predetermined period after the handover from the acquired measurement data of the wireless quality.

[0029] 4 is a diagram for explaining the filtering process according to this embodiment, in which the horizontal axis represents time and the vertical axis represents throughput (an example of wireless quality), and the acquired wireless quality measurement data is displayed in chronological order.

[0030] For example, suppose that a handover occurs at time t1 in Fig. 4. In this case, the filtering unit 102 removes data from the measurement data from time t1 when the handover occurs to time t2 after a predetermined period T has elapsed, as data 201 immediately after the handover.

[0031] In step S304, the filtering unit 102 outputs the measurement data as learning data. For example, if a handover has occurred, the filtering unit 102 removes data for a predetermined period T after the handover from the acquired measurement data of wireless quality, and outputs the resulting measurement data as learning data to the machine learning unit 20, etc. On the other hand, if a handover has not occurred, the filtering unit 102 outputs the acquired measurement data of wireless quality as learning data to the machine learning unit 20, etc.

[0032] By the process of FIG. 3, the data processing device 100 can easily remove abnormal values ​​that occur immediately after a handover from the learning data of the machine learning that predicts the wireless quality.

[0033] [Example 2] Fig. 5 is a diagram illustrating an example of the configuration of a data processing system according to Example 2. The data processing system 1 according to Example 2 includes a data acquisition unit 10, a data processing device 100, and a machine learning unit 20, similar to the data processing system 1 according to Example 1 described with reference to Fig. 1.

[0034] The data processing device 100 according to the second embodiment includes an actual value database (DB) 501 in addition to the functional components of the data processing device 100 according to the first embodiment described with reference to Fig. 1. The actual value DB 501 may be provided outside the data processing device 100.

[0035] The performance value DB 501 is a database that stores in advance information such as correlations between normal parameters and / or abnormal values, which are necessary for determining the deterioration of wireless quality immediately after a handover.

[0036] The filtering unit 102 according to the second embodiment detects an abnormal value (deterioration of wireless quality) by comparing the measurement data after handover among the measurement data of wireless quality acquired by the data acquiring unit 10 with the values ​​in the performance value DB 501. For example, the filtering unit 102 determines that the measurement data after handover is an abnormal value when the measurement data is a value outside the upper limit or lower limit stored in the performance value DB 501 or is a value outside the correlation.

[0037] Furthermore, when an abnormal value is detected, filtering section 102 removes data from the measurement data of wireless quality during a predetermined period after handover in which the abnormal value was detected, and outputs the resulting measurement data as learning data.

[0038] The processing contents of the determining unit 101 according to the second embodiment are the same as the processing contents of the determining unit 101 according to the first embodiment.

[0039] <Processing Flow> Next, the processing flow of the data processing method according to the second embodiment will be described.

[0040] Fig. 6 is a flowchart showing an example of data processing according to the second embodiment. This processing shows an example of processing executed by the data processing device 100 described in Fig. 5. Note that detailed description of processing content similar to the data processing according to the first embodiment described in Fig. 3 will be omitted here.

[0041] In step S601, the data processing device 100 acquires the measurement data of the wireless quality acquired by the data acquisition unit 10.

[0042] In step S602, the determination unit 101 determines whether a handover has occurred based on the acquired wireless quality measurement data. If a handover has occurred, the determination unit 101 shifts the process to step S603. On the other hand, if a handover has not occurred, the determination unit 101 shifts the process to step S606.

[0043] In step S603, the filtering unit 102 detects an abnormal value immediately after the handover from the measurement data.

[0044] Fig. 7 is a diagram for explaining the abnormal value detection process according to the second embodiment. Fig. 7 shows an example of measurement data 210 acquired by the data processing device 100 from the data acquisition unit 10. In the example of Fig. 7, the measurement data 210 shows the relationship between RSRP and throughput of data 201 immediately after handover and data 202 other than immediately after handover, among the measurement data of wireless quality.

[0045] The filtering unit 102, for example, determines whether each piece of data 201 immediately after handover is a value that deviates from the correlation stored in the actual value DB 501, and detects data that deviates from the correlation (for example, data 701) as an abnormal value.

[0046] In step S604, the filtering unit 102 branches the process depending on whether an abnormal value is detected. If an abnormal value is detected, the filtering unit 102 shifts the process to step S605. On the other hand, if an abnormal value is not detected, the filtering unit 102 shifts the process to step S606.

[0047] In step S605, the filtering unit 102 removes data for a predetermined period T after the occurrence of handover in which an abnormal value was detected from the acquired measurement data of wireless quality, for example, as described with reference to FIG.

[0048] In step S606, the filtering unit 102 outputs the measurement data as learning data.

[0049] 6, when an abnormal value is detected, filtering unit 102 removes data for a predetermined period after handover in which the abnormal value was detected from the acquired measurement data of wireless quality and outputs the resulting measurement data as learning data. On the other hand, when no abnormal value is detected, filtering unit 102 outputs the acquired measurement data of wireless quality as learning data.

[0050] According to the data processing device 100 of Example 2, it is possible to remove abnormal values ​​from the measurement data by comparing them with normal data for an event in which wireless quality is likely to deteriorate, such as immediately after handover, thereby improving the estimation accuracy by machine learning.

[0051] <Hardware Configuration> The data processing device 100 has, for example, the hardware configuration of a computer 800 as shown in Fig. 8. Alternatively, the data processing device 100 is realized by a plurality of computers 800.

[0052] 8 is a diagram showing the hardware configuration of a computer. In the example of Fig. 8, a computer 800 includes a processor 801, a memory 802, a storage device 803, a communication device 804, an input device 805, an output device 806, a bus B, and the like.

[0053] The processor 801 is, for example, an arithmetic unit such as a CPU (Central Processing Unit) that executes predetermined programs to realize various functions. The memory 802 is a storage medium readable by the computer 800, and includes, for example, a RAM (Random Access Memory) and a ROM (Read Only Memory). The storage device 803 is a computer-readable storage medium, and may include, for example, a HDD (Hard Disk Drive), an SSD (Solid State Drive), various optical disks, and magneto-optical disks.

[0054] The communication device 804 includes one or more communication devices for communicating with other devices via a wireless or wired network. The input device 805 is an input device (e.g., a keyboard, a mouse, a microphone, a switch, a button, a sensor, etc.) that accepts input from the outside. The output device 806 is an output device (e.g., a display, a speaker, an LED lamp, etc.) that outputs to the outside.

[0055] The bus B is commonly connected to the above components and transmits, for example, address signals, data signals, and various control signals. The processor 801 is not limited to a CPU, and may be, for example, a DSP (Digital Signal Processor), a PLD (Programmable Logic Device), or an FPGA (Field Programmable Gate Array).

[0056] (Supplementary Note) The data processing device 100 in this embodiment is not limited to being realized by a dedicated device, but may also be realized by a general-purpose computer. In this case, a program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be read into a computer system and executed to realize the function. Note that the term "computer system" here includes hardware such as an OS and peripheral devices.

[0057] Furthermore, "computer-readable recording media" includes various storage devices such as portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, as well as hard disks built into computer systems. Furthermore, "computer-readable recording media" may also include devices that dynamically store programs for a short period of time, such as communication lines when transmitting programs via networks such as the Internet or communication lines such as telephone lines, and devices that store programs for a certain period of time, such as volatile memory within computer systems that serve as servers or clients in such cases.

[0058] Furthermore, the above program may be one that realizes part of the above-mentioned functions, or may be one that can realize the above-mentioned functions in combination with a program already recorded in a computer system, or may be one that is realized using hardware such as a PLD or FPGA.

[0059] According to the present embodiment, it is possible to provide a data processing system 1 that can easily remove outliers from learning data of machine learning that predicts wireless quality. As a result, the data processing system 1 can improve the accuracy of prediction of wireless quality by machine learning.

[0060] Summary of Embodiments This specification discloses at least the following data processing systems. A data processing device, a data processing method, and a program. (Item 1) A data processing system comprising: a determination unit that determines the occurrence of a handover based on measurement data of wireless quality; and a filtering unit that, when the handover occurs, outputs measurement data from which data for a predetermined period after the handover has been removed from the measurement data of the wireless quality as learning data for machine learning to predict the wireless quality. (Item 2) The filtering unit detects an abnormal value after the handover from the measurement data of the wireless quality, and, when the abnormal value is detected, outputs measurement data from which data for the predetermined period after the handover in which the abnormal value was detected has been removed as the learning data, and when the abnormal value is not detected, outputs the measurement data of the wireless quality as the learning data. The data processing device described in Item 1. (Clause 3) A data processing device comprising: a determination unit that determines the occurrence of a handover based on measurement data of wireless quality; and a filtering unit that, when the handover has occurred, outputs the measurement data of the wireless quality from which data for a predetermined period after the handover has been removed as learning data for machine learning to predict the wireless quality. (Clause 4) A data processing method in which a computer executes: a process of determining the occurrence of a handover based on the measurement data of wireless quality; and a process of, when the handover has occurred, outputting the measurement data of the wireless quality from which data for a predetermined period after the handover has been removed as learning data for machine learning to predict the wireless quality. (Clause 5) A program, or a storage medium storing a program, that causes a computer to execute: a process of determining the occurrence of a handover based on the measurement data of wireless quality; and a process of, when the handover has occurred, outputting the measurement data of the wireless quality from which data for a predetermined period after the handover has been removed as learning data for machine learning to predict the wireless quality.

[0061] Although the present embodiment has been described above, the present invention is not limited to such a specific embodiment, and various modifications and changes are possible within the scope of the gist of the present invention described in the claims.

[0062] REFERENCE SIGNS LIST 1 Data processing system 10 Data acquisition unit 20 Machine learning unit 100 Data processing device 101 Determination unit 102 Filtering unit 501 Performance value DB 800 Computer

Claims

1. A data processing system comprising: a judgment unit that judges whether a handover will occur based on measurement data of wireless quality; and a filtering unit that, when the handover occurs, outputs measurement data from which data for a predetermined period after the handover has been removed from the measurement data of the wireless quality as learning data for machine learning that predicts the wireless quality.

2. A data processing device having: a judgment unit that judges whether a handover will occur based on measurement data of wireless quality; and a filtering unit that, when the handover occurs, outputs measurement data from which data for a predetermined period after the handover has been removed from the measurement data of the wireless quality as learning data for machine learning that predicts the wireless quality.

3. A data processing method in which a computer executes the following processes: determining the occurrence of a handover based on measurement data of wireless quality; and, if the handover occurs, outputting the measurement data of the wireless quality from which data for a predetermined period after the handover has been removed as learning data for machine learning to predict the wireless quality.

4. A program that causes a computer to execute the following processes: determining the occurrence of a handover based on measurement data of wireless quality; and, when the handover occurs, outputting the measurement data of the wireless quality from which data for a predetermined period after the handover has been removed as learning data for machine learning that predicts the wireless quality.

Images