Program and data processing device

Abstract

The object is to provide a program and a data processing device that improve the accuracy of determining who is a visitor to a target facility. [Solution] The program causes a computer to execute a process of acquiring first GPS data representing the position of a person to be determined as a visitor based on a GPS (Global Positioning System) relative to a target facility, and determining whether the person to be determined as a visitor has visited the target facility based on the first GPS data and a trained model, and the trained model is generated by machine learning multiple training data that respectively define the relationship between beacon position data representing the position of the visitor estimated based on wireless communication between a beacon terminal installed in a facility other than the target facility and a mobile terminal carried by a visitor who actually visited the other facility, and second GPS data representing the position of the visitor based on the GPS.

Description

[Technical field]

[0001] The present invention relates to a program and a data processing device. [Background technology]

[0002] It is known that personal devices such as smartphones can estimate their location through communication with GPS (Global Positioning System) satellites (see, for example, Patent Documents 1 to 3). There is also known a technology that determines that a user of a mobile terminal has visited a facility when the current location of the mobile terminal is within the location range of the facility for a stay determination time or longer (see, for example, Patent Document 4). It is also known that in urban areas, GPS positioning data may have large errors due to the influence of densely packed buildings, etc. (see, for example, Patent Document 5). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] U.S. Pat. No. 1,097,2862 [Patent Document 2] U.S. Pat. No. 1,035,2718 [Patent Document 3] US Patent Application Publication No. 2013 / 0297198 [Patent Document 4] International Publication No. 2018 / 120361 [Patent Document 5] JP 2023-148599 A Summary of the Invention [Problem to be solved by the invention]

[0004] As described above, if the GPS positioning data has a large error, the current location of the mobile terminal equipped with the GPS may be outside the location range of the facility (hereinafter referred to as the target facility). As a result, even if the user carrying the mobile terminal actually visited the target facility as a visitor, it may be erroneously determined that the user did not visit the target facility. Also, even if the user did not actually visit the target facility as a visitor, it may be erroneously determined that the user visited the target facility.

[0005] Therefore, in one aspect, an object of the present invention is to provide a program and a data processing device that improve the accuracy of determining who is a subject of determination of a visit to a target facility. [Means for solving the problem]

[0006] In one embodiment, the program causes a computer to execute a process of acquiring first GPS data representing a GPS-based position of a person to be determined as a visitor relative to a target facility, and determining whether or not the person to be determined as a visitor has visited the target facility based on the first GPS data and a trained model, and the trained model is generated by machine learning multiple training data that respectively define the relationship between beacon position data representing the position of the visitor estimated based on wireless communication between a beacon terminal installed in a facility other than the target facility and a mobile terminal carried by a visitor who actually visited the other facility, and second GPS data representing the position of the visitor based on the GPS.

[0007] In the above configuration, the trained model can be generated by machine learning a plurality of teacher data each defining a relationship between the beacon position data, the second GPS data, and specified feature data, and the specified feature data can be configured to include attribute data of the visitor, facility characteristic data of the other facility, weather data on the visit day when the visitor visited the other facility, and GPS characteristic data including at least one or all of the visitor's movement speed, the distance from the visitor's position to the other facility, and the orientation of the visitor relative to the other facility, calculated based on the second GPS data.

[0008] In the above configuration, at least one of the attribute data, the facility characteristic data, the weather data, and the GPS characteristic data can be excluded from the specified feature data based on the accuracy of the visitor determination subject.

[0009] In the above configuration, a probability value representing the accuracy of determining whether the person to be determined as a visitor has visited the target facility can be calculated, a predetermined confusion matrix is ​​generated based on the probability value, and a first threshold value for determining whether the probability value represents a visit or a non-visit is determined by repeatedly comparing the confusion matrix with a time-series correlation coefficient between correct answers and predictions regarding the visit determination.

[0010] In the above configuration, the first threshold value may be displayed independently and adjustably on a display device connected to the computer.

[0011] In the above configuration, a plurality of second thresholds different from the first threshold are calculated based on the first threshold, and the first threshold and the plurality of second thresholds can be selectively displayed on a display device connected to the computer.

[0012] In the above configuration, the trained model may be generated by machine learning multiple training data that each define a relationship between the beacon position data, the second GPS data, and another wireless communication having a different wireless communication standard from the wireless communication.

[0013] In the above configuration, the beacon location data may include a time of the visitor's arrival at the other facility, and the trained model may be generated by machine learning multiple training data each of which defines a relationship between the beacon location data and the second GPS data from a first time before the visit time to a second time after the visit time.

[0014] In the above configuration, the beacon location data includes the visitor's visit time to the other facility, and the trained model can be generated by machine learning multiple training data that each define a relationship between the beacon location data and the second GPS data from a first time that is a first predetermined time before the visit time to a second time that is a second predetermined time after the visit time that is longer than the first predetermined time.

[0015] In the above configuration, the number of people can be corrected based on the number of people who are determined to have visited the target facility and demographic data including the population of the area including the target facility.

[0016] In one embodiment, the data processing device has an acquisition unit that acquires first GPS data representing a GPS-based position of a person to be determined as a visitor relative to a target facility, and a judgment unit that judges whether the person to be determined as a visitor has visited the target facility based on the first GPS data and a learned model, and is characterized in that the learned model is generated by machine learning multiple training data that each define the relationship between beacon position data representing the position of the visitor estimated based on wireless communication between a beacon terminal installed in another facility other than the target facility and a mobile terminal carried by a visitor who actually visited the other facility, and second GPS data representing the position of the visitor based on the GPS. Effect of the Invention

[0017] According to the present invention, it is possible to improve the accuracy of determining who is a visitor to a target facility. [Brief description of the drawings]

[0018] [Figure 1] 1 is an example of a data processing system. [Diagram 2] 2 is an example of a hardware configuration of a data processing server. [Diagram 3] 2 is an example of a functional configuration of a data processing server. [Figure 4] 13 is an example of beacon log data. [Diagram 5] 1 is an example of learning target GPS log data. [Figure 6] 1 is an example of POI data. [Figure 7] 13 is an example of GPS log data to be determined. [Figure 8] FIG. 11 is a diagram illustrating an example of determination target GPS log data. [Figure 9] 1 is an example of area population data. [Figure 10] FIG. 11 is a diagram for explaining an example of acquisition of learning target GPS log data. [Figure 11] 1 is an example of feature amount data. [Figure 12] FIG. 13 is a diagram illustrating an example of how a determination threshold is determined. [Figure 13] FIG. 13 is a diagram illustrating an example of correcting a provisional number of people to an actual number of people. [Figure 14] 1A is an example of feature amount data in case 1. 1B is an example of feature amount data in case 2. 1C is an example of feature amount data in case 3. 1D is an example of feature amount data in case 4. [Figure 15] This is an example of performance evaluation of a trained model. [Figure 16] 13 is a flowchart illustrating an example of a process executed by the data processing server. [Figure 17] 13A is an example of a visit determination result according to a comparative example, and FIG. 13B is an example of a visit determination result according to an embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0020] As shown in Fig. 1, the data processing system ST is a computer system including a terminal device 10 and a data processing server 100. The terminal device 10 and the data processing server 100 are connected via a communication network NW. The communication network NW includes either or both of a LAN (Local Area Network) and the Internet. The data processing server 100 is an example of a data processing device.

[0021] In FIG. 1, a PC (Personal Computer) is shown as an example of the terminal device 10, but the terminal device 10 is not limited to a PC. The terminal device 10 may be a smart terminal such as a smartphone or a tablet terminal. In addition, in FIG. 1, a physical server device is shown as an example of the data processing server 100, but the data processing server 100 may be a virtual server device. Furthermore, in FIG. 1, one data processing server 100 is shown as an example, but multiple data processing servers 100 may be provided in the data processing system ST, and various data processing may be distributed among the multiple data processing servers 100.

[0022] The data processing system ST is used by a user 11 belonging to a business company. The business company may be a manufacturer that produces products, or a non-manufacturer that provides services. Manufacturers include, but are not limited to, food manufacturers, cosmetics manufacturers, and shoe manufacturers, for example. Non-manufacturers include, but are not limited to, restaurants and retail stores, for example. For example, the business company may include a company that conducts market research for non-manufacturers, or a company that provides sales support such as consulting.

[0023] A user 11 can use the data processing system ST by operating an input device 12 provided in the terminal device 10 and accessing the data processing server 100. For example, when the user 11 performs a predetermined operation on the input device 12, the control device 13 of the terminal device 10 transmits an instruction corresponding to the predetermined operation to the data processing server 100. Upon receiving the instruction, the data processing server 100 executes various data processing operations based on the received instruction and transmits the processing results to the control device 13.

[0024] Although details will be described later, for example, when the data processing server 100 receives an instruction corresponding to a predetermined operation, the data processing server 100 judges whether or not the person to be determined as a visitor to the target facility has visited the target facility based on the GPS data of the person to be determined as a visitor to the target facility and a trained model generated in advance. When the data processing server 100 judges whether or not the person to be determined as a visitor has visited the target facility, the data processing server 100 transmits the judgment result to the control device 13. When the control device 13 receives the judgment result, the control device 13 displays a predetermined confirmation screen including the judgment result on the display device 14 provided in the terminal device 10. As a result, the predetermined confirmation screen appears on the display device 14. By viewing the predetermined confirmation screen, the user 11 can grasp with high accuracy whether or not the person to be determined as a visitor has actually visited the target facility.

[0025] The hardware configuration of the data processing server 100 will be described with reference to Fig. 2. Note that the above-mentioned terminal device 10 basically has the same hardware configuration as the data processing server 100, and therefore a detailed description thereof will be omitted.

[0026] The data processing server 100 includes a CPU (Central Processing Unit) 100A as a processor, and a RAM (Random Access Memory) 100B and a ROM (Read Only Memory) 100C as memories. The data processing server 100 includes a network I / F (Interface) 100D and a HDD (Hard Disk Drive) 100E. Instead of the HDD (Hard Disk Drive) 100E, a SSD (Solid State Drive) may be used.

[0027] The data processing server 100 may include at least one of an input I / F 100F, an output I / F 100G, an input / output I / F 100H, and a drive device 100I, as necessary. The CPU 100A to the drive device 100I are connected to each other by an internal bus 100J. In other words, the data processing server 100 can be realized by a computer.

[0028] An input device 710 is connected to the input I / F 100F. Examples of the input device 710 include a keyboard, a mouse, and a touch panel. A display device 720 is connected to the output I / F 100G. Examples of the display device 720 include a liquid crystal display. A semiconductor memory 730 is connected to the input / output I / F 100H. Examples of the semiconductor memory 730 include a USB (Universal Serial Bus) memory and a flash memory. The input / output I / F 100H reads a program stored in the semiconductor memory 730. The input I / F 100F and the input / output I / F 100H each include a USB port, for example. The output I / F 100G includes a display port, for example.

[0029] A portable recording medium 740 is inserted into the drive device 100I. The portable recording medium 740 may be a removable disk such as a CD (Compact Disc)-ROM or a DVD (Digital Versatile Disc). The drive device 100I reads a program recorded on the portable recording medium 740. The network I / F 100D includes, for example, a LAN port and a communication circuit. The communication circuit includes either or both of a wired communication circuit and a wireless communication circuit. The network I / F 100D is connected to a communication network NW.

[0030] A program stored in at least one of ROM 100C, HDD 100E, and semiconductor memory 730 is temporarily stored in RAM 100B by CPU 100A. A program recorded in portable recording medium 740 is temporarily stored in RAM 100B by CPU 100A. By executing the stored program, CPU 100A realizes various functions described later, and also executes a data processing method including various processes described later. The program may be one according to a flowchart described later.

[0031] The functional configuration of the data processing server 100 will be described with reference to Figures 3 to 15. Note that Figure 3 shows the main functions of the data processing server 100.

[0032] 3, the data processing server 100 includes a storage unit 110, a processing unit 120, and a communication unit 130. The storage unit 110 can be realized by either or both of the RAM 100B and the HDD 100E described above. The processing unit 120 can be realized by the CPU 100A described above. The communication unit 130 can be realized by the network I / F 100D described above.

[0033] The storage unit 110, the processing unit 120, and the communication unit 130 are connected to each other. The storage unit 110 includes a beacon log storage unit 111, a GPS log storage unit 112, a POI (Point Of Interest) information storage unit 113, a determination target storage unit 114, and an area population storage unit 115. The storage unit 110 uses the beacon log storage unit 111, the GPS log storage unit 112, the POI information storage unit 113, the determination target storage unit 114, and the area population storage unit 115 to store various data.

[0034] In addition, a POI is an example of a facility, and includes, for example, a commercial facility such as a store. The store includes restaurants and retail stores. The POI may include public facilities such as parks, libraries, and stations, sports facilities such as baseball fields and soccer stadiums, medical facilities such as hospitals and clinics, and roads such as sidewalks and roadways. The road may be a motorway (for example, a national expressway or an urban expressway) including SA (Service Area) and PA (Parking Area), or may be a general road other than a motorway including intersections and T-junctions. In addition, the POI is not limited to such man-made objects, and may include natural objects such as mountains, rivers, and lakes. In this way, the POI corresponds to a specific feature on the map information, such as a man-made object or a natural object.

[0035] The processing unit 120 includes a teacher data generation unit 121, a feature amount data generation unit 122, a machine learning unit 123, and a visit determination unit 124. The processing unit 120 also includes a judgment threshold determination unit 125, a correction unit 126, and a characteristic data selection unit 127. The processing unit 120 processes various data using the teacher data generation unit 121, the feature amount data generation unit 122, the machine learning unit 123, the visit determination unit 124, the judgment threshold determination unit 125, the correction unit 126, and the characteristic data selection unit 127.

[0036] The beacon log storage unit 111 stores beacon log data. The beacon log data is an example of beacon position data. The beacon log data represents the position of a visitor who actually visited a second POI different from the first POI. The first POI according to this embodiment is an example of a target facility, such as a restaurant. A beacon terminal is not installed in the first POI. On the other hand, the second POI according to this embodiment is an example of another facility, such as a retail store. A beacon terminal is installed in the second POI. Thus, the first POI and the second POI differ in terms of the business, the presence or absence of a beacon terminal, the position of the POI identified by latitude and longitude, etc.

[0037] The location of the visitor is estimated based on BLE (Bluetooth (registered trademark) Low Energy) communication between a beacon terminal installed at the second POI and a mobile terminal carried by a visitor who actually visits the second POI. The mobile terminal may be a smartphone, a tablet terminal, a smart watch, a game terminal, or smart glasses. BLE communication is an example of wireless communication. In this way, the location of the visitor is estimated based on BLE communication (specifically, the radio wave strength of the BLE communication), so the location of the visitor is estimated with higher accuracy than in the case of GPS.

[0038] An access point that provides Wi-Fi (registered trademark) may be installed at the second POI. Wi-Fi (registered trademark) is an example of another wireless communication. Wi-Fi (registered trademark) has a different wireless communication standard from BLE communication. In this case, by using BLE communication and Wi-Fi (registered trademark) in combination, the location of the visitor can be estimated with higher accuracy than when BLE communication is used alone. Thus, the use of BLE communication alone or the use of BLE communication and Wi-Fi (registered trademark) in combination has been described as an example, but as long as it can be estimated that the visitor has come to the location, the estimation of the location of the visitor is not limited to these methods. For example, the location of the visitor may be estimated by using Wi-Fi (registered trademark) alone. In addition, the location of the visitor may be estimated based on the active action of the user, such as check-in using a QR (Quick Response) code (registered trademark) (so-called QR check-in), use of application software, or answering a questionnaire.

[0039] Here, as shown in FIG. 4, the beacon log data includes a number of items such as a device ID (identifier), a beacon terminal ID, latitude, longitude, and detection date and time. In the device ID item, a unique identifier for identifying a mobile terminal such as a smartphone or tablet is registered. In the beacon terminal ID item, a unique identifier for identifying a beacon terminal is registered. In each of the latitude and longitude items, for example, the latitude and longitude of the mobile terminal estimated based on BLE communication between the beacon terminal and the mobile terminal are registered. In the detection date and time item, the detection date and time of the mobile terminal detected by the beacon terminal based on BLE communication is registered at a predetermined detection time interval.

[0040] Therefore, if the visitor carries a mobile terminal, the visitor's current location can be identified with high accuracy. In this way, if the beacon terminal is installed at the second POI, it is assumed that the visitor will stay inside the second POI or near the outside of the second POI.

[0041] Returning to FIG. 3, the GPS log storage unit 112 stores the GPS log data to be learned. The GPS log data to be learned is an example of second GPS data. The GPS log data to be learned represents a position based on the GPS function of the mobile terminal. Therefore, if a visitor carries a mobile terminal, the position of the visitor is determined based on the GPS function.

[0042] As shown in FIG. 5, the GPS log data to be learned includes multiple items such as a device ID, latitude, longitude, and positioning date and time. In the device ID item, a unique identifier for identifying a mobile terminal such as a smartphone or tablet is registered. In the beacon terminal ID item, a unique identifier for identifying a beacon terminal is registered. In the latitude and longitude items, the latitude and longitude of the mobile terminal measured based on the GPS function of the mobile terminal are registered. In the positioning date and time item, except for the part shown in parentheses, the positioning date and time of the latitude and longitude are registered at a predetermined positioning time interval different from the detection time interval described above. The positioning time interval may be longer or shorter than the detection time interval. Details of the part shown in parentheses will be described later. In this way, the latitude and longitude included in the learning target GPS log data and the beacon log data are close to each other, but often do not match.

[0043] Returning to Fig. 3, the POI information storage unit 113 stores POI data related to POIs including the first POI and second POI described above. The POI data is an example of facility characteristic data. As shown in Fig. 6, the POI data includes multiple items such as POI-ID, POI name, business type, central latitude, central longitude, size, and floor. Although not shown, the POI data may also include items such as site area and location.

[0044] The POI-ID field registers a unique identifier that identifies the POI. The POI name field registers the name of the POI. The business type field registers the business model of the POI, such as food and beverage or retail. The central latitude field registers the latitude at the center of the area occupied by the POI. The central longitude field registers the longitude at the center of the area occupied by the POI. The size field registers the physical size of the POI, such as small or medium. The floor field registers the hierarchical level of the POI.

[0045] In this way, the POI data has the center position of the area occupied by the POI. Therefore, it is possible to determine whether or not a visitor or a person to be determined as a visitor has visited a POI based on the center position of the POI and the current position of the visitor or the person to be determined as a visitor based on the GPS. For example, if the current position of the visitor is included in an area corresponding to the vicinity of the center position of the POI, the visitor is determined to have visited the POI. Conversely, if the current position of the visitor is included in an area far away from the center position of the POI, the visitor is determined to have not visited the POI.

[0046] The area occupied by the POI may be a circular area of ​​a predetermined radius based on the central position, a rectangular area based on the central position, or a polygonal area made up of a combination of polygonal areas.

[0047] Returning to FIG. 3, the judgment target storage unit 114 stores judgment target GPS log data. The judgment target GPS log data is an example of first GPS data. The judgment target GPS log data represents a position based on the GPS function of a mobile terminal. Therefore, if the person to be visited to be judged carries a mobile terminal, the position of the person to be visited to be judged is determined based on the GPS function.

[0048] As shown in FIG. 7, the GPS log data to be judged includes multiple items such as a device ID, latitude, longitude, and positioning date and time. A unique identifier for identifying a mobile device such as a smartphone or tablet is registered in the device ID item. A unique identifier for identifying a beacon terminal is registered in the beacon terminal ID item. The latitude and longitude items are registered with the latitude and longitude of the mobile terminal that is positioned based on the GPS function of the mobile terminal carried by the person to be judged for visitation. The positioning date and time item is registered with the date and time of latitude and longitude measurement at a specified positioning time interval.

[0049] According to the GPS log data to be determined, for example, as shown in Fig. 8, various positions 22, 23, 24, etc. of the visit determination subject near a POI (specifically, a first POI) 21 are represented on map information MP. For example, a plurality of chronological positions 22 of the visit determination subject identified by a device ID "D051" is represented on map information MP. Also, a plurality of chronological positions 23 of the visit determination subject identified by a device ID "D052" is represented on map information MP.

[0050] Returning to FIG. 3, the area population storage unit 115 stores area population data. The area population data is an example of demographic data. The area population data is obtained by dividing the above-mentioned map information MP into a plurality of areas (hereinafter referred to as area meshes) represented by polygons, and includes the actual mobile population of each area mesh. In this embodiment, a square is used as an example of the shape of the area mesh, but the shape of the area mesh may be a rectangle, a hexagon, or the like.

[0051] As shown in Fig. 9, the area population data includes multiple items such as area ID, latitude range, longitude range, counting date, population, installation POI #1, and installation POI #2. Although not shown, the area population data also includes items such as population by age group and male / female ratio. A unique identifier that identifies the area mesh is registered in the area ID item. The latitude range item is registered with the latitude range that defines the area mesh. The longitude range item is registered with the longitude range that defines the area mesh.

[0052] The date of aggregation, along with the month and year, is registered in the aggregation date field. The population field registers the floating population of all age groups. The floating population is the total number of people temporarily staying in the target area mesh. Fields such as Installation POI #1 and Installation POI #2 register the identifiers of POIs installed in the area mesh.

[0053] Returning to FIG. 3, the teacher data generating unit 121 acquires beacon log data (see FIG. 4) from the beacon log storage unit 111. The teacher data generating unit 121 also acquires a portion of the GPS log data (see FIG. 5) included in a radius of several kilometers (e.g., 1 km) around the POI (specifically, the second POI) from the GPS log storage unit 112. More specifically, as shown in FIG. 10, the teacher data generating unit 121 acquires GPS log data to be learned from the visit date and time "t0" of the second POI to the first positioning date and time "T1" which is a first predetermined time Δt1 (e.g., several minutes) before. The teacher data generating unit 121 also acquires GPS log data to be learned from the visit date and time "t0" to the second positioning date and time "T5" which is a second predetermined time Δt2 (e.g., several minutes) after the visit date and time. The visit date and time is an example of the visit time and corresponds to the detection date and time described above. The first positioning date and time is an example of the first time, and the second positioning date and time is an example of the second time.

[0054] In this way, the second predetermined time Δt2 is longer than the first predetermined time Δt1. The reason for this is to reliably secure the GPS log data to be learned. In other words, there is a time lag between the response date and time of the BLE communication and the detection date and time (i.e., the visit date and time), and it is highly likely that the generation of the beacon log data is not complete and no data remains. In order to reliably acquire the GPS log data to be learned based on the beacon log data, the second predetermined time Δt2, which is longer than the first predetermined time Δt1, is adopted in this embodiment.

[0055] When the teacher data generating unit 121 acquires a part of the GPS log data to be learned, the teacher data generating unit 121 determines the part of the GPS log data to be learned as the GPS log data of a visitor who actually visited the second POI. In other words, the teacher data generating unit 121 determines the remaining part of the GPS log data to be learned as the GPS log data of a non-visitor who did not visit the second POI. When the teacher data generating unit 121 determines the GPS log data of the visitor, the teacher data generating unit 121 determines the optimal first predetermined time Δt1 and second predetermined time Δt2 for each second POI for each of the first predetermined time Δt1 and the second predetermined time Δt2. For example, the teacher data generating unit 121 evaluates the correlation between the beacon log data and the GPS log of the visitor in a time series, and determines the first predetermined time Δt1 and the second predetermined time Δt2 at which the correlation coefficient is maximized.

[0056] Returning to Fig. 3, the feature amount data generating unit 122 generates predetermined feature amount data. More specifically, the feature amount data generating unit 122 calculates the movement speed of the visitor based on a part of the GPS log data of the visitor described above. In addition, the feature amount data generating unit 122 acquires POI data (see Fig. 6) from the POI information storage unit 113, and calculates the distance from the position of the visitor to the second POI and the orientation of the visitor with respect to the second POI based on the part of the GPS log data of the visitor and the POI data.

[0057] When the feature data generating unit 122 calculates the moving speed, distance, and direction, it generates GPS characteristic data including at least one of the moving speed, distance, and direction, and adds it to the predetermined feature data. It is desirable that the GPS characteristic data includes all of the moving speed, distance, and direction. This is likely to improve the accuracy of the visit determination for the visit determination target person. On the other hand, if the GPS characteristic data includes any one of the moving speed, distance, and direction, it is likely that the judgment load of the visit determination is reduced. In this case, there is a possibility that power consumption is reduced. In addition, although details will be described later, the feature data generating unit 122 adds weather data, attribute data, and the like selected as characteristic data by the characteristic data selecting unit 127 to the predetermined feature data. As a result, as shown in FIG. 11, the feature data generating unit 122 can generate feature data including various characteristic data such as GPS characteristic data and weather data.

[0058] Returning to FIG. 3, the machine learning unit 123 generates a trained model by machine learning the multiple teacher data generated by the teacher data generation unit 121. The machine learning unit 123 can also generate a trained model by machine learning multiple teacher data with feature amounts, each of which defines the relationship between multiple teacher data and feature amount data. By using a trained model that has been machine-learned from multiple teacher data with feature amounts, it is highly likely that the accuracy of visit determination will be improved compared to when feature amount data is not used. Note that in this embodiment, the machine learning method is not particularly limited. For example, the machine learning method may be gradient boosting, a deep neural network, or a transformer.

[0059] The visit determination unit 124 acquires the determination target GPS log data (see FIG. 7) for the first POI from the determination target storage unit 114. When the visit determination unit 124 acquires the determination target GPS log data, the visit determination unit 124 judges whether or not the visit determination target has visited the first POI based on the determination target GPS log data and the learned model. This improves the accuracy of judging the visit determination target for the first POI. The visit determination unit 124 may also judge whether or not the visit determination target has visited the first POI based on the determination target GPS log data and the learned model with feature amounts. This further improves the accuracy of judging the visit determination target for the first POI.

[0060] When determining whether or not the visit determination subject has visited the first POI, the visit determination unit 124 calculates a probability value representing the accuracy of the visit determination for each position (i.e., a combination of latitude and longitude) of the visit determination subject included in the determination subject GPS log data. If the determination accuracy of the visit determination is high, a high probability value is calculated. Conversely, if the determination accuracy of the visit determination is low, a low probability value is calculated.

[0061] The judgment threshold determination unit 125 determines a judgment threshold for the probability value calculated by the visit determination unit 124. The judgment threshold is a threshold for determining whether the probability value indicates a visit or a non-visit. The judgment threshold is in the range of a minimum threshold "0 (zero)" to a maximum threshold "1". Even if a high probability value is calculated, depending on the judgment threshold, there is a possibility that the visit determination subject may be erroneously determined to not have visited the first POI, so the judgment threshold determination unit 125 determines an optimal judgment threshold. Specifically, as shown in the upper part of FIG. 12, when the positions 22, 23, 24, etc. of various visit determination subjects are classified into visits and non-visits, the judgment threshold determination unit 125 generates a confusion matrix.

[0062] As shown in the middle of Figure 12, the confusion matrix contains correct answers "1" representing visits and incorrect answers "0" representing non-visits as correct label items in the column direction. In addition, the confusion matrix contains positive judgments "1" representing visits and negative judgments "0" representing non-visits as predicted label items in the row direction. The combination of the correct answer "1" and the positive judgment "1" corresponds to a true positive. The combination of the correct answer "1" and the negative judgment "0" corresponds to a false negative. The combination of the incorrect answer "0" and the positive judgment "1" corresponds to a false positive. The combination of the incorrect answer "0" and the negative judgment "0" corresponds to a true negative.

[0063] When the judgment threshold determination unit 125 generates a confusion matrix, it repeatedly compares the confusion matrix with the time-series correlation coefficients of the correct answers and predictions regarding visit judgment within a preset range of allowable values ​​in the confusion matrix, and determines the judgment threshold of the probability value at which the correlation coefficient is maximized as the optimal judgment threshold, as shown in the lower part of Fig. 12. The optimal judgment threshold is an example of a first threshold. By using the optimal judgment threshold, the accuracy of judgment between visits and non-visits to the target facility is improved.

[0064] The judgment threshold determination unit 125 may independently display the determined optimum judgment threshold in an adjustable manner on the display device 720. This allows the user 11 to operate the input device 710 to adjust the optimum judgment threshold to a desired judgment threshold desired by the user 11.

[0065] On the other hand, the judgment threshold determination unit 125 may determine an optimal judgment threshold, and then calculate multiple preliminary judgment thresholds different from the optimal judgment threshold based on the determined optimal judgment threshold. The preliminary judgment threshold is an example of a second threshold. For example, the judgment threshold determination unit 125 calculates multiple preliminary judgment thresholds that are spaced apart by a unit increment above and below the optimal judgment threshold. After calculating the multiple preliminary judgment thresholds, the judgment threshold determination unit 125 selectively displays the optimal judgment threshold and the multiple preliminary judgment thresholds on the display device 720. This allows the user 11 to operate the input device 710 to select one of the optimal judgment threshold and the multiple preliminary judgment thresholds.

[0066] Returning to FIG. 3, the correction unit 126 acquires the determination target GPS log data (see FIG. 7) from the determination target storage unit 114. In addition, the correction unit 126 acquires the area population data (see FIG. 9) from the area population storage unit 115. When the correction unit 126 acquires the determination target GPS log data and the area population data, the correction unit 126 calculates a correction value between the count value and the population based on the count value of the number of visit determination targets included in the determination target GPS log data and the population included in the area population data. When the correction unit 126 calculates the correction value, the correction unit 126 multiplies the number of visit determination targets determined to have visited the first POI by the correction value, and executes an expanded estimation to the actual number of visit determination targets who visited the first POI. In addition, when the correction unit 126 executes the expanded estimation, it corrects the demographic composition ratio. That is, the correction unit 126 corrects each composition ratio of the age group population and the male / female ratio included in the area population data based on the actual number of people who visited the first POI.

[0067] Even if visit determination is performed on the determination target GPS log data, the determination target GPS log data is a sample that represents a portion of the area population data, so even if the number of visit determination target persons who visited the first POI is tallied, the tallied result only represents a provisional number of people. Therefore, the correction unit 126 executes an expanded estimation and correction to the demographic composition ratio. As a result, the provisional number of people is corrected to the actual number of people, as shown in FIG. 13.

[0068] Returning to FIG. 3, the characteristic data selection unit 127 selects characteristic data that leads to improvement of the visit determination accuracy of the visit determination subject. The characteristic data selection unit 127 is provided with analysis information for analyzing the determination result of the visit determination in advance, and when it is determined that the determination accuracy of the visit determination subject is low based on the analysis information, it selects characteristic data such as weather data, visitor attribute data, and traffic volume data. For example, the characteristic data selection unit 127 can acquire weather data and attribute data from outside the data processing server 100. The characteristic data selection unit 127 may acquire weather data and attribute data from a characteristic storage unit (not shown) provided in the storage unit 110.

[0069] Weather data includes weather and temperature. Visitor attribute data includes attributes such as the gender and age of the visitor, as well as the behavioral preferences of the visitor. Behavioral preferences indicate, for example, the characteristic that visitors to a beauty salon are often women in their 20s, or that when different retail stores such as a drugstore and a supermarket are adjacent to each other, visitors tend to visit one of the stores. Traffic volume data includes time-series road traffic volume.

[0070] When the characteristic data selection unit 127 acquires the characteristic data, it outputs the acquired characteristic data to the feature data generation unit 122. As a result, the feature data generation unit 122 adds the characteristic data to the feature data. For example, as shown in FIGS. 14(a) and 14(b), the feature data generation unit 122 adds weather data, which is one of the characteristic data, to the feature data. As a result, a new trained model is generated, and visit judgment is performed for the visit judgment target person.

[0071] If it leads to an improvement in the accuracy of visit determination, the characteristic data selection unit 127 may exclude at least one characteristic data from the feature data. For example, as shown in Figs. 14(c) and (d), the feature data generation unit 122 excludes visitor attribute data, which is one of the characteristic data, from the feature data. As a result, another new trained model is generated, and visit determination is performed for the visit determination target person. As a result, as shown in Fig. 15, an embodiment having a correlation coefficient closer to the correct example is obtained in comparison with the comparative example.

[0072] The operation of the data processing server 100 will be described with reference to FIG.

[0073] First, the teacher data generating unit 121 generates teacher data (step S1). For example, when the teacher data generating unit 121 receives an instruction corresponding to a predetermined operation, the teacher data generating unit 121 acquires beacon log data and a portion of the learning target GPS log data, and generates a plurality of teacher data each defining a relationship between the beacon log data and a portion of the learning target GPS log data.

[0074] After the teacher data generating unit 121 generates the teacher data, the feature data generating unit 122 then generates feature data (step S2). For example, the feature data generating unit 122 generates predetermined feature data including GPS characteristic data based on a part of the learning target GPS log data and POI data. After the feature data generating unit 122 generates the feature data, the machine learning unit 123 generates a trained model by machine learning (step S3). For example, the machine learning unit 123 generates a trained model by machine learning a plurality of teacher data and feature data. After the machine learning unit 123 generates the trained model, the visit determination unit 124 then acquires the determination target GPS log data from the determination target storage unit 114 (step S4).

[0075] When the GPS log data to be determined is acquired, the visit determination unit 124 determines whether or not the visit determination subject has visited the first POI based on the GPS log data to be determined and the learned model (step S5). After the visit determination is completed, the characteristic data selection unit 127 then determines whether or not all characteristic data has been selected (step S6).

[0076] If all characteristic data have not been selected (step S6: NO), the characteristic data selection unit 127 selects characteristic data (step S7). For example, the characteristic data selection unit 127 selects weather data as characteristic data. When the characteristic data selection unit 127 selects characteristic data, the feature data generation unit 122 executes the process of step S2. Therefore, when the characteristic data selection unit 127 selects weather data, the feature data generation unit 122 generates feature data by adding the weather data to the GPS characteristic data.

[0077] On the other hand, if all the characteristic data have been selected (step S6: YES), the judgment threshold determination unit 125 determines the judgment threshold (step S8). More specifically, as described above, the judgment threshold determination unit 125 determines the judgment threshold for the probability value calculated by the visit determination unit 124. When the judgment threshold determination unit 125 determines the judgment threshold, the correction unit 126 executes a correction process (step S9) and ends the process. Specifically, the correction unit 126 executes an expanded estimation as the correction process, and then corrects the demographic composition ratio. Note that the characteristic data selection unit 127 may additionally execute the processes of steps S6 and S7 after the process of step S9. This may improve visit determination for visit determination targets.

[0078] The present embodiment will be described in comparison with a comparative example with reference to FIGS.

[0079] First, as shown in Fig. 17(a), in the comparative example, if the position of the person to be determined for visit, which is measured using only GPS, is included within a range of a circle C1 of a predetermined radius from the center position of the first POI 21, the person to be determined for visit is determined to have visited the first POI 21. In other words, if the position of the person to be determined for visit is not included within the range of the circle C1, the person to be determined for visit is determined not to have visited the first POI 21. The predetermined radius may be several tens of meters or several kilometers.

[0080] On the other hand, as shown in FIG. 17(b), in the embodiment, by using GPS and the learned model in combination, it may be determined that the visit determination subject has not visited the first POI 21 even if the subject is within the range of the circle C1. Also, by using GPS and the learned model in combination, it may be determined that the visit determination subject has visited the first POI 21 even if the subject is outside the range of the circle C1. Thus, according to the present embodiment, by using GPS and the learned model in combination, the accuracy of determining the visit determination subject for the first POI 21 is improved. Note that the data processing server 100 can display a predetermined confirmation screen including the map information MP according to such an embodiment on the display device 14.

[0081] Although the preferred embodiment of the present invention has been described in detail above, the present invention is not limited to the specific embodiment, and various modifications and variations are possible within the scope of the gist of the present invention described in the claims. [Explanation of symbols]

[0082] ST Data Processing System 100 Data Processing Servers 110 Storage section 111 Beacon log storage unit 112 GPS log storage section 113 POI information storage section 114 Judgment target memory unit 115 Area Population Memory Department 120 Processing section 121 Teacher Data Generation Unit 122 Feature data generation unit 123 Machine Learning Department 124 Visitor Determination Department 125 Judgment threshold determination unit 126 Correction Unit 127 Characteristic data selection section

Claims

1. Acquire first GPS data representing a location of the person to be determined as visiting the facility based on a Global Positioning System (GPS); determining whether the visit determination subject has visited the target facility based on the first GPS data and the learned model; The process is executed by a computer, The trained model is generated by machine learning a plurality of teacher data each defining a relationship between beacon position data representing a position of a visitor estimated based on wireless communication between a beacon terminal installed in a facility different from the target facility and a mobile terminal carried by a visitor who actually visited the facility, and second GPS data representing a position of the visitor based on the GPS. A program characterized by:

2. The trained model is generated by machine learning a plurality of teacher data each defining a relationship between the beacon position data, the second GPS data, and predetermined feature amount data, the predetermined feature amount data includes attribute data of the visitor, facility characteristic data of the other facility, weather data on a visit day when the visitor visited the other facility, and GPS characteristic data including at least one or all of a moving speed of the visitor calculated based on the second GPS data, a distance from the position of the visitor to the other facility, and an orientation of the visitor with respect to the other facility; 2. The program according to claim 1 .

3. excluding at least one of the attribute data, the facility characteristic data, the weather data, and the GPS characteristic data from the predetermined feature amount data based on the determination accuracy of the visitor determination subject; 3. The program according to claim 2, which causes the computer to execute a process.

4. calculating a probability value representing the accuracy of determining whether the visit determination subject has visited the target facility, generating a predetermined confusion matrix based on the probability value, and repeatedly comparing the confusion matrix with a time-series correlation coefficient between a correct answer and a prediction regarding the visit determination, to determine a first threshold value for determining whether the probability value represents a visit or a non-visit; 3. The program according to claim 1 or 2.

5. The first threshold value is independently displayed in an adjustable manner on a display device connected to the computer.

5. The program according to claim 4.

6. calculating a plurality of second thresholds different from the first threshold based on the first threshold; selectively displaying the first threshold value and the plurality of second threshold values ​​on a display device connected to the computer; 5. The program according to claim 4.

7. The trained model is generated by machine learning a plurality of teacher data each defining a relationship between the beacon position data, the second GPS data, and another wireless communication having a different wireless communication standard from the wireless communication.

3. The program according to claim 1 or 2.

8. the beacon location data includes a time of arrival of the visitor at the other facility; The trained model is generated by machine learning a plurality of teacher data each defining a relationship between the beacon position data and the second GPS data from a first time before the visit time to a second time after the visit time.

3. The program according to claim 1 or 2.

9. the beacon location data includes a time of arrival of the visitor at the other facility; The trained model is generated by machine learning a plurality of teacher data each defining a relationship between the beacon position data and the second GPS data from a first time point that is a first predetermined time before the visit time to a second time point that is a second predetermined time after the visit time that is longer than the first predetermined time.

3. The program according to claim 1 or 2.

10. correcting the number of people based on the number of people who are determined to have visited the target facility and demographic data including the population of the area including the target facility; 3. The program according to claim 1, wherein the program causes the computer to execute a process.

11. an acquisition unit that acquires first GPS data representing a position of the person to be determined as visiting the facility based on a Global Positioning System (GPS); A determination unit that determines whether the visit determination subject has visited the target facility based on the first GPS data and a learned model; A data processing device comprising: The trained model is generated by machine learning a plurality of teacher data each defining a relationship between beacon position data representing a position of a visitor estimated based on wireless communication between a beacon terminal installed in a facility different from the target facility and a mobile terminal carried by a visitor who actually visited the facility, and second GPS data representing a position of the visitor based on the GPS.

23. A data processing device comprising:

Images