WO2023149005A1 - Estimation apparatus and estimation method - Google Patents

Estimation apparatus and estimation method Download PDF

Info

Publication number
WO2023149005A1
WO2023149005A1 PCT/JP2022/033035 JP2022033035W WO2023149005A1 WO 2023149005 A1 WO2023149005 A1 WO 2023149005A1 JP 2022033035 W JP2022033035 W JP 2022033035W WO 2023149005 A1 WO2023149005 A1 WO 2023149005A1
Authority
WO
WIPO (PCT)
Prior art keywords
estimation
product
section
shelf
processor
Prior art date
Application number
PCT/JP2022/033035
Other languages
French (fr)
Japanese (ja)
Inventor
健一郎 山田
Original Assignee
株式会社日立製作所
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 株式会社日立製作所 filed Critical 株式会社日立製作所
Publication of WO2023149005A1 publication Critical patent/WO2023149005A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions

Definitions

  • the present invention relates to an estimation device and an estimation method for estimating data.
  • Patent Literature 1 describes a first input unit for inputting information indicating that a user reached out for a first product, an estimation unit for estimating attributes of the user, and a case where the user did not purchase the first product. and a storage unit that stores that the user reaches out for the first product in association with the user's attribute, and when the user purchases the second product in the same category as the first product, the storage unit: An electronic device that stores information of a second product is disclosed.
  • the above-described conventional technology does not take into account the length of time people stay in front of the shelves and the planogram setting based on the flow of people.
  • the purpose of the present invention is to facilitate estimation of surface stay time.
  • the technology disclosed herein is an estimating device having a processor that executes a program and a storage device that stores the program, wherein the processor is an area in which partitions having one or more surfaces facing a person are arranged. in the plane, based on an acquisition process for acquiring people flow direction data indicating the movement trajectory of the person in time series, the people flow direction data acquired by the acquisition process, and the position of the plane in the section; and an output process for outputting the stay time calculated by the calculation process.
  • FIG. 1 is an explanatory diagram of a system configuration example 1 of an analysis system according to a first embodiment.
  • FIG. 2 is a plan view showing compartments.
  • FIG. 3 is an explanatory diagram showing an example of a section peripheral environment data storage unit.
  • FIG. 4 is an explanatory diagram showing an example of a surface layout data storage unit.
  • FIG. 5 is a block diagram showing a hardware configuration example of a computer.
  • FIG. 6 is an explanatory diagram showing an example of a registration screen.
  • FIG. 7 is an explanatory diagram showing an example of a display screen displaying a surface stay time estimation mode.
  • FIG. 8 is an explanatory diagram showing an example of a display screen displaying the recommended planogram estimation mode.
  • FIG. 1 is an explanatory diagram of a system configuration example 1 of an analysis system according to a first embodiment.
  • FIG. 2 is a plan view showing compartments.
  • FIG. 3 is an explanatory diagram showing an example of a section peripheral environment data storage unit.
  • FIG. 9 is an explanatory diagram of system configuration example 2 of the analysis system according to the first embodiment.
  • FIG. 10 is an explanatory diagram of a system configuration example 3 of the analysis system according to the first embodiment.
  • FIG. 11 is a flow chart showing an example of a screen staying time estimation processing procedure by a cloud server.
  • FIG. 12 is a flowchart showing an example of a procedure for estimating the unit price per customer by the unit for estimating the unit price per customer.
  • FIG. 13 is a flowchart illustrating an example of a shelf allocation estimation processing procedure by a shelf allocation estimation unit.
  • FIG. 14 is an explanatory diagram showing an example of grouping product groups.
  • FIG. 15 is an explanatory diagram showing an example of product groups belonging to a group.
  • FIG. 16 is an explanatory diagram showing example 1 of product addition.
  • FIG. 17 is an explanatory diagram of product addition example 2. As shown in FIG.
  • FIG. 1 is an explanatory diagram of a system configuration example 1 of an analysis system according to a first embodiment.
  • Configuration example 1 is an example of a configuration for estimating a screen stay time using a screen stay time estimation model.
  • the surface stay time estimation model is a learning model for estimating the stay time of the person 134 in front of the shelf surface, which is the front surface of the shelf 132 .
  • the cloud system 100 includes a cloud server 101, commercial facilities 102A, 102B, 102C, . , 103B, 103C, . .
  • the commercial facility 102 has a section 131 and a people flow data acquisition unit 130 installed near the section 131 .
  • Compartment 131 has shelves 132 for displaying merchandise.
  • three shelves 132 are arranged in a U-shape.
  • a portion of the section 131 that connects with a passage outside the section 131 is referred to as an opening 133 .
  • the opening 133 is a portion through which a person 134 enters and exits.
  • the people flow data acquisition unit 130 acquires the people flow data 111 of the people 134 around and within the block 131 and transmits the acquired people flow data 111 to the cloud server 101 via the network 104 .
  • the people flow data acquisition unit 130 is, for example, LiDAR, and detects the shape of an object, the distance to the object, and the moving direction of the object. As a result, for example, the people flow data acquisition unit 130 identifies the shape of each person 134 and obtains the people flow data 111 indicating how far and in which direction the person 134 having the identified shape moves. That is, the people flow data 111 is time-series position information of the people 134 . Since LiDAR can also detect whether a person 134 has extended their arms in the XY plane direction, the people flow data 111 also includes the number of times each person 134 has extended their arms in the XY plane direction.
  • the information processing device 103 is a computer operated by a shelf allocation manager or a tenant manager (hereinafter simply referred to as manager) of the commercial facility 102, and inputs data to the cloud server 101 and receives data from the cloud server 101. output.
  • the information processing device 103A is a computer operated by the manager of the commercial facility 102A
  • the information processing device 103B is a computer operated by the manager of the commercial facility 102B
  • the information processing device 103C is a computer operated by the manager of the commercial facility 102B.
  • a computer operated by the manager of the commercial facility 102C a computer operated by the manager of the commercial facility 102C.
  • the cloud server 101 has a block surrounding environment data storage unit 113 , a surface layout data storage unit 114 , and a surface stay time estimation model storage unit 115 .
  • the compartment surrounding environment data storage unit 113 stores the compartment surrounding environment data.
  • the block surrounding environment data is data indicating the surrounding environment of the block 131. For example, the scale of the commercial facility 102 in which the block 131 is located, the construction (enclosed or open) of the commercial facility 102, the location, the average traffic volume of the people 134 indicates
  • the surface layout data storage unit 114 stores surface layout data.
  • the surface layout data is data that defines the layout of the shelf surface of the shelf 132 .
  • the page stay time estimation model storage unit 115 stores the page stay time estimation model 116 for each section 131 .
  • the plane stay time estimation model 116 exists for each section 131, and when the people flow direction data 112, the section surrounding environment data, and the plane arrangement data are input, the plane stay time 117, which is an estimated value, is calculated for each shelf surface in the section 131. output to
  • the cloud server 101 also has a people flow direction analysis unit 121 , a planogram estimation unit 122 , and a customer unit price estimation unit 123 .
  • the people flow direction analysis unit 121 analyzes the people flow data 111 and outputs the people flow direction data 112 of the person 134 .
  • the people flow direction data 112 is, for example, a time-series movement trajectory of a person 134 moving on the XY plane. LiDAR can also detect whether the person 134 has extended their arm in the XY plane direction.
  • the planogram of the product group is determined by the width direction and the height direction of the shelf surface S#, the people flow direction data 112 indicates the number of times each person 134 stretches his or her arm for each product on the shelf surface S#. Also includes
  • the planogram estimation unit 122 estimates the planogram based on the surface stay time 117 and outputs the planogram data 118 as an estimation result.
  • the customer unit price estimation unit 123 calculates the customer unit price in the section 131 based on the face stay time 117 .
  • the unit price per customer includes the amount of sales per customer based on the amount of sales of all products in section 131 within a certain period, and the amount of sales per customer based on the amount of sales for each item in section 131 within a certain period. There is a sales amount for each and At least one of these customer unit prices is referred to as customer unit price data 119 .
  • the cloud server 101 may also estimate the rent of the location within the commercial facility 102 where the parcel 131 is located based on the visit time 117 and a coefficient specific to the commercial facility 102 .
  • FIG. 2 is a plan view showing the partition 131. As shown in FIG. FIG. 2 shows three types of compartments 131A-131C. Sections 131A-131C are referred to as section 131 when not distinguished.
  • X is the direction orthogonal to the opening 133 and Y is the people flow orthogonal to X.
  • Section 131 has shelf surface S# and area A#. # is a number. Area A# corresponds to shelf surface S# with the same number #. Specifically, for example, person 134 faces shelf surface S# when staying in area A#.
  • the shelf surface S# is the front surface of the shelf 132 which is the frontage of the shelf 132 . Note that although the shelf surface S# is a surface in which the front surfaces of one or more shelves 132 are flush with each other, in this example, the shelf surface S# will be described as the front surface of one shelf 132 in order to simplify the description. .
  • FIG. 3 is an explanatory diagram showing an example of the partition surrounding environment data storage unit 113.
  • the partition surrounding environment data storage unit 113 holds the partition surrounding environment data in, for example, a table format.
  • the block peripheral environment data storage unit 113 has facility ID 301, scale 302, building 303, location 304, and average traffic volume 305 as fields.
  • a combination of field values in the same row constitutes an entry that defines block surrounding environment data for one commercial facility 102 .
  • the facility ID 301 is identification information that uniquely identifies the commercial facility 102 .
  • the scale 302 is information indicating whether the commercial facility 102 is a regional type, a community type, or a neighborhood type.
  • the building 303 is information indicating whether the structure of the commercial facility 102 is closed or open.
  • the location 304 is information indicating whether the installation location of the commercial facility 102 is close to a station, a city, or a suburb. Near the station means near the nearest station of the commercial facility 102 or underground of the nearest station.
  • the average traffic volume 305 is the traffic volume of the people 134 per unit time (person/hour) in the commercial facility 102 .
  • the scale 302, building 303 and location 304 are, for example, one-hot encoded.
  • FIG. 4 is an explanatory diagram showing an example of the surface arrangement data storage unit 114.
  • the surface layout data storage unit 114 holds the surface layout data 400 of the section 131 in, for example, a table format.
  • the surface arrangement data 400 has facility ID 301, section ID 402, shelf surface ID 403, and arrangement information 404 as fields.
  • the partition ID 402 is identification information that uniquely identifies the partition 131 .
  • the shelf surface ID 403 is identification information that uniquely identifies the shelf surface S#.
  • the arrangement information 404 is information that defines the position of the shelf surface S# within the section 131 and the normal direction of the shelf surface S#.
  • the arrangement information 404 also holds data such as the number of shelves of the shelf 132 on the shelf surface specified by the shelf surface ID 403 and the shape and volume of the frontage of each shelf.
  • FIG. 5 is a block diagram showing a hardware configuration example of the computer 500.
  • the computer 500 has a processor 501 , a storage device 502 , an input device 503 , an output device 504 and a communication interface (communication IF) 505 .
  • Processor 501 , storage device 502 , input device 503 , output device 504 and communication IF 505 are connected by bus 506 .
  • Processor 501 controls computer 500 .
  • a storage device 502 serves as a work area for the processor 501 .
  • the storage device 502 is a non-temporary or temporary recording medium that stores various programs and data.
  • Examples of the storage device 502 include ROM (Read Only Memory), RAM (Random Access Memory), HDD (Hard Disk Drive), and flash memory.
  • the input device 503 inputs data.
  • the input device 503 includes, for example, a keyboard, mouse, touch panel, numeric keypad, scanner, microphone, and sensor.
  • the output device 504 outputs data.
  • Output devices 504 include, for example, displays, printers, and speakers.
  • Communication IF 505 connects to a network and transmits and receives data.
  • the people flow direction analysis unit 121 the planogram estimation unit 122, and the customer unit price estimation unit 123 shown in FIG. It is a function to be realized.
  • FIG. 6 is an explanatory diagram showing an example of the registration screen.
  • the registration screen 600 is a screen for registering information about the section 131 managed by the information processing apparatus 103 and for correcting registered information.
  • the registration screen 600 includes a section ID input field 601, a section name selection pull-down 602, a section photo display area 603, section surrounding environment data 604, a surface arrangement layout 605, a registration button 606, and a correction button 607. have.
  • the section ID input field 601 is an area for the administrator to operate the input device 503 to input the section ID 402 .
  • the registration button 606 By pressing the registration button 606 , the entered character string is registered in the cloud server 101 as the partition ID 402 .
  • the correction button 607 By pressing the correction button 607 , the corrected character string is updated to the cloud server 101 as a new section ID 402 unless it overlaps with another section ID 402 .
  • the block name selection pull-down 602 is a user interface that displays a list of blocks 131 and accepts selection of the blocks 131 .
  • the section name selection pull-down 602 is an area for inputting the name (section name) of the section 131 whose section ID 402 is not registered.
  • the register button 606 By pressing the register button 606 , the entered character string is registered in the cloud server 101 as a partition name in association with the character string (partition ID 402 ) entered in the partition ID input field 601 .
  • the correction button 607 the corrected character string is associated with the section ID 402 as the section name, and the cloud server 101 is updated.
  • a section photo display area 603 is an area for displaying image data that is a picture of the section 131 (section picture).
  • the image data is displayed as a section photograph in the section photograph display area 603 .
  • the registration button 606 By pressing the registration button 606, the image data being displayed is registered in the cloud server 101 as a section photograph in association with the character string (section ID 402) entered in the section ID input field 601.
  • the correction button 607 By pressing the correction button 607, the image data after replacement is updated in the cloud server 101 in association with the section ID 402 as the section photograph after replacement.
  • the block surrounding environment data 604 is the block surrounding environment data related to the commercial facility 102 in which the block 131 exists. Which section 131 exists in which commercial facility 102 is managed by the surface layout data 400 of the surface layout data storage unit 114 .
  • the information processing device 103 By requesting the cloud server 101 with the facility ID 301 of the corresponding commercial facility 102 , the information processing device 103 acquires the section surrounding environment data 604 of the commercial facility 102 and displays it on the registration screen 600 .
  • the surface arrangement layout 605 is, for example, figure data obtained by planarly viewing the section 131A, and is created from the surface arrangement data 400.
  • FIG. 1 A perspective view of the section 131A.
  • a registration button 606 is a user interface for registering the parcel name, parcel photograph, parcel surrounding environment data 604 (or its facility ID 301), and surface layout layout 605 in the cloud server 101 in association with the parcel ID 402.
  • a correction button 607 is a user interface for correcting the section ID 402 , section name, section photograph, and surface layout 605 and updating the cloud server 101 .
  • FIG. 7 is an explanatory diagram showing an example of a display screen displaying the time-of-visit estimation mode.
  • the display screen 700 includes an estimation target section ID input field 701, an estimation target section name selection pull-down 702, a first estimation condition selection pull-down 703, a second estimation condition selection pull-down 704, a surface stay time heat map 705, a surface It has a stay time list 706 , an execution button 707 and a mode change button 708 .
  • the estimated target section ID input field 701 is an area for the administrator to operate the input device 503 to input the estimated target section ID.
  • the estimation target section ID is identification information that uniquely identifies the estimation target section.
  • the estimation target section is the section 131 that is the target for estimating the surface stay time.
  • the estimation target block name selection pull-down 702 is a user interface that displays a list of blocks 131 and accepts the selection of an estimation target block.
  • the first estimation condition selection pulldown 703 is a user interface that displays a list of first estimation conditions and accepts selection of the first estimation condition.
  • the first estimation conditions are weekdays, holidays, and days of the week such as Sunday to Monday.
  • the first estimation condition can also be set without condition setting.
  • the second estimation condition selection pulldown 704 is a user interface that displays a list of second estimation conditions and accepts selection of the second estimation condition.
  • the second estimation condition is a period such as per day, per week, or per month.
  • the second estimation condition can also be set without condition setting.
  • the face staying time heat map 705 is a heat map of the face staying time 117 in the estimation target section that satisfies the first estimation condition and the second estimation condition. That is, the length of the surface staying time 117 is displayed by the shade of color.
  • the face staying time heat map 705 is created by the cloud server 101 superimposing the color gradation indicating the face staying time 117 on the face arrangement layout 605 using the people flow direction data 112 .
  • a surface staying time list 706 is an estimation result showing an estimated value of the surface staying time for each shelf surface S#.
  • the surface staying time list 706 has a shelf surface ID 403 and a surface staying time 117 .
  • the estimation target section ID input field 701 or the estimation target section name selection that satisfies the first estimation condition and the second estimation condition of the first estimation condition selection pull-down menu 703 and the second estimation condition selection pull-down menu 704 is performed.
  • This is a user interface for calculating surface stay time 117 for each shelf surface S# for the estimation target section specified by pull-down 702 and displaying surface stay time heat map 705 and surface stay time list 706 .
  • a mode change button 708 is a user interface for changing the display contents of the display screen 700 from the surface stay time estimation mode to the recommended planogram estimation mode, and from the recommended planogram estimation mode to the surface stay time estimation mode. is.
  • FIG. 8 is an explanatory diagram showing an example of a display screen 700 displaying the recommended planogram estimation mode.
  • the display screen 700 includes an estimation target section ID input field 701, an estimation target section name selection pull-down 702, a shelf type selection pull-down 801, a shelf number selection pull-down 802, and a first estimation condition selection pull-down. 703 , estimation target section plane layout data 804 , planogram layout 805 , execution button 807 , and mode change button 708 .
  • the shelf type selection pull-down 801 is a user interface that displays a list of shelf types and accepts selection of a shelf type.
  • the shelf type is the type of shelf.
  • a shelf level selection pull-down 802 is a user interface for displaying a list of the number of shelves and accepting a selection of the number of shelves.
  • the estimation target block surface layout data 804 is the surface layout layout 605 of the estimation target block.
  • the planogram layout 805 is graphic data indicating the planogram of each shelf surface S# of the estimation target section.
  • the estimation target section ID input field 701 that satisfies the shelf type, the number of shelves, and the first estimation condition of the shelf type selection pull-down 801, the number of shelves selection pull-down 802, and the first estimation condition selection pull-down 703 is displayed.
  • it is a user interface for calculating the planogram for each shelf surface S# for the estimation target section specified by the estimation target section name selection pull-down 702 and displaying the planogram layout 805 as the calculation result.
  • FIG. 9 is an explanatory diagram of a system configuration example 2 of the cloud system 100 according to the first embodiment.
  • Configuration example 2 is an example of a configuration for learning a surface stay time estimation model.
  • the cloud server 101 has a people flow direction data storage unit 901 , a face stay time storage unit 902 , and a face stay time estimation model learning unit 903 .
  • the people flow direction data storage unit 901 stores the people flow direction data 112 .
  • the page stay time storage unit 902 stores the page stay time.
  • the face stay time is, for example, the actually measured value of the stay time of the person 134 for each shelf face S# of each section 131 .
  • the surface stay time estimation model learning unit 903 uses the people flow direction data 112, the section surrounding environment data, and the arrangement information 404 (position and normal direction) of the shelf surface S# in the surface arrangement data 400 as explanatory variables, and calculates the surface stay time. As an objective variable, the surface stay time estimation model 116 is learned by machine learning.
  • the surface stay time estimation model learning unit 903 shown in FIG. 9 is, specifically, a function realized by causing the processor 501 to execute a program stored in the storage device 502, for example.
  • the people flow direction data storage unit 901 and the surface stay time storage unit 902 shown in FIG. 9 are stored in the storage device 502 .
  • FIG. 10 is an explanatory diagram of a system configuration example 3 of the cloud system 100 according to the first embodiment.
  • Configuration example 3 is an example of a configuration for learning a planogram model.
  • the cloud server 101 includes a learning planogram data 1001, a learning planogram data storage unit 1002, a customer unit price storage unit 1003, a planogram estimation model 1004, a planogram estimation model storage unit 1005, and a planogram estimation model. and a learning unit 1010 .
  • the planogram data for learning 1001 is planogram data for learning.
  • the shelving allocation data is data indicating what is arranged on which shelf of which shelf surface S#.
  • the planogram data storage unit 1002 for learning stores the planogram data 1001 for learning.
  • the customer unit price storage unit 1003 stores the customer unit price for each section 131 .
  • the planogram estimation model 1004 is a learning model for estimating the planogram.
  • the planogram estimation model storage unit 1005 stores the planogram estimation model 1004 .
  • the planogram estimation model learning unit 1010 uses the people flow direction data 112, the section surrounding environment data, the arrangement information 404 (position and normal direction) of the shelf surface S# in the surface arrangement data 400, and the customer unit price as explanatory variables, and uses them as explanatory variables for learning.
  • a planogram estimation model 1004 is learned by machine learning using the planogram data 1001 as an objective variable.
  • planogram estimation model learning unit 1010 shown in FIG. 9 is, specifically, a function realized by causing the processor 501 to execute a program stored in the storage device 502, for example.
  • the learning planogram data storage unit 1002 and the customer unit price storage unit 1003 shown in FIG. 9 are stored in the storage device 502 .
  • FIG. 11 is a flow chart showing an example of a page stay time estimation processing procedure by the cloud server 101 .
  • the cloud server 101 receives from the information processing device 103 the section ID 402 of the section to be estimated input by the information processing device 103 (hereinafter referred to as the section ID 402 to be estimated), the first estimation condition, and the second estimation condition. (Step S1101).
  • the cloud server 101 uses the people flow direction analysis unit 121 to acquire the people flow direction data 112 corresponding to the first estimation condition and the second estimation condition for the estimation target section (step S1102).
  • the cloud server 101 inputs the people flow direction data 112, the block surrounding environment data, and the surface arrangement data acquired in step S1102 to the surface stay time estimation model 116 of the estimation target block, and calculates the surface stay time 117, which is an estimated value. (step S1103).
  • the cloud server 101 generates a screen stay time heat map 705 and a screen stay time list 706 from the screen stay time 117 (step S1104), and transmits them to the information processing device 103 (step S1105). As a result, a display screen 700 as shown in FIG. 7 is displayed on the information processing apparatus 103 .
  • FIG. 12 is a flowchart showing an example of a procedure for estimating the unit price per customer by the unit price estimating unit 123 for customers.
  • the cloud server 101 acquires the estimation target section ID 402 and the estimation target period of the customer unit price data 119 from the information processing device 103 (step S1201).
  • the cloud server 101 obtains the number of times that the person 134 stretched his or her arm toward the shelf surface S# within the estimation target period based on the people flow direction data 112 for the estimation target section selected in step S1201 (step S1202). . Specifically, for example, cloud server 101 acquires the number of times that person 134 has extended his/her arm to shelf surface S# for each product display position on shelf surface S#.
  • the cloud server 101 acquires from the information processing device 103 the sales amount of the products displayed in the estimation target section within the estimation target period (step S1203).
  • the cloud server 101 divides the sales amount of each product displayed in the estimation target section within the estimation target period acquired in step S1203 by the number of times the arm is stretched to the display position of the product. A customer unit price for each product is calculated (step S1204).
  • the cloud server 101 transmits the customer unit price calculated in step S1204 to the information processing device 103 (step S1205). Thereby, the customer unit price data 119 is displayed on the information processing device 103 .
  • FIG. 13 is a flow chart showing an example of the shelf allocation estimation processing procedure by the shelf allocation estimation unit 122 .
  • the cloud server 101 acquires the number of products sold in the estimation target section in the estimation target period and the customer unit price data 119 (step S1301). Specifically, for example, when a POS (Point Of Sales) system is introduced in the information processing apparatus 103 and the commercial facility 102, the cloud server 101 performs estimation in the estimation target period from the POS data of the information processing apparatus 103. Acquire the number of products sold and the unit price per customer in the target section.
  • POS Point Of Sales
  • the cloud server 101 calculates the number of sales of the product group in the estimation target section during the estimation target period entered by the administrator in the information processing device 103. is acquired from the information processing apparatus 103 .
  • the cloud server 101 also acquires the customer unit price data 119 in the estimation target period from the customer unit price estimation unit 123 .
  • the cloud server 101 groups the products in the section 131 by the number of products sold and the unit price per customer (step S1302).
  • FIG. 14 is an explanatory diagram showing an example of grouping product groups.
  • the horizontal axis is the customer unit price of the product, and the threshold is Th1.
  • the vertical axis is the number of products sold, and the threshold is Th2.
  • Group G1 is a set of products whose unit price per customer is equal to or greater than threshold Th1 and whose number of sales is equal to or greater than threshold Th2.
  • a group G2 is defined as a set of products whose unit price per customer is smaller than the threshold value Th1 and whose number of sales is equal to or greater than the threshold value Th2.
  • a group G3 is a set of products whose unit price per customer is equal to or greater than the threshold value Th1 and whose number of sales is smaller than the threshold value Th2.
  • a group G4 is a set of products whose unit price per customer is smaller than the threshold Th1 and whose number of sales is smaller than the threshold Th2.
  • the cloud server 101 classified the display target product groups in the section 131 into four groups G1 to G4, but the number of groups is not limited to four as long as there are two or more.
  • the cloud server 101 sets the product display priority of the group (step S1303).
  • the cloud server 101 sets in order of group G1 ⁇ group G2 ⁇ group G3 ⁇ group G4. That is, the groups G1 and G2 whose number of sales is Th2 or more are given the highest priority, and among them, the group G1 whose unit price per customer is Th1 is given priority. After the groups G1 and G2, the group G3 having a customer unit price of Th2 or higher is given priority over the group G4.
  • the cloud server 101 sets in the order of group G1 ⁇ group G4 ⁇ group G3 ⁇ group G2 according to user settings. Note that the cloud server 101 may set the priority order of product display based only on the number of sales, or may set the priority order of product display based only on the unit price per customer.
  • the cloud server 101 determines whether or not there is an unselected shelf surface S# for the estimation target section (step S1304). If there is an unselected shelf surface S# (step S1304: Yes), the cloud server 101 selects the shelf surface S# with the longest surface stay time 117 among the unselected shelf surface S#s (step S1305).
  • the cloud server 101 determines whether or not there is an unselected group (step S1306). If there are unselected groups (step S1306: Yes), the cloud server 101 selects one unselected group with the highest priority set in step S1303 (step S1307).
  • the cloud server 101 determines whether or not the product remains in the selected group (step S1308). If the product remains (step S1308: Yes), the cloud server 101 selects the product from the selection group (step S1309). The selected items are removed from the selection group. If no product remains in the selected group (step S1308: No), the process returns to step S1306.
  • FIG. 15 is an explanatory diagram showing an example of product groups belonging to a group.
  • group G1 includes "toy A", “educational toy A”, “mischievous toy A”, and "wisdom ring A”.
  • the order in which the products in the group are selected is not limited. Specifically, for example, the order of selection may be random, name order, sales amount order, or customer unit price order.
  • the cloud server 101 determines whether the selected product can be added to the selected shelf surface S# (step S1310). Specifically, for example, the product has shape data and volume data, and the shelf surface S# also has residual storage shape and residual volume data for each shelf. The cloud server 101 determines whether or not the selected product can be stored on any shelf based on the shape data and volume data of the selected product and the remaining storage shape and remaining volume data of each shelf of the selected shelf surface S#. to judge.
  • step S13101310 No
  • the cloud server 101 adds the selected product to the shelf level of the selected shelf face S#.
  • FIG. 16 is an explanatory diagram showing example 1 of product addition.
  • shelf 132 has shelves T1-T7 in the height direction.
  • the racks T1 to T7 are referred to as the racks T when not distinguished.
  • a range within which the shelf T can be reached by the person 134's line of sight or in a standing posture is referred to as a golden zone 1600 .
  • the golden zone 1600 is trays T3-T5.
  • the shelves T3 to T5 in the golden zone 1600 are preferentially selected as the display destination of the selected product.
  • Outer trays T1, T2, T6, and T7 are selected in descending order of distance from the golden zone 1600 .
  • FIG. 17 is an explanatory diagram showing example 2 of product addition.
  • Tray T7 may also be included in golden zone 1600, as shown in FIG. Which shelf T is included in the golden zone 1600 and how the shelf T is selected can be set in the cloud server 101 from the information processing device 103 in advance by the administrator.
  • step S1311 return to step S1308. If there is no unselected shelf surface S# in step S1304 (step S1304: No), or if there is no unselected group in step S1306 (step S1306: No), the cloud server 101 sends the planogram result to the information processing apparatus 103. output to As a result, the display screen 700 shown in FIG. 8 can be displayed on the information processing apparatus 103 .
  • the planogram estimation process is performed based on rules, but the planogram estimation process may be performed by machine learning.
  • the cloud server 101 uses the number of products sold, the unit price per customer of products, and the surface stay time 117 for each shelf surface S# as explanatory variables, and the product display position on the shelf 132 as an objective variable.
  • a split learning model is generated for each section 131 and stored in the storage device 502 .
  • the cloud server 101 outputs the product display position on the shelf 132 of the prediction target product by inputting the number of sales of the prediction target product, the customer unit price, and the surface stay time 117 for each shelf surface S# of the estimation target section. do. In this way, the cloud server 101 can output the display position of each product to the information processing device 103 as a planogram result.
  • the present embodiment it is possible to facilitate estimation of the page stay time 117, and to set an efficient planogram in consideration of the page stay time 117, thereby reducing the burden on the administrator. can be planned.
  • the rent of the section 131 can be optimized.
  • the present invention is not limited to the above-described embodiments, and includes various modifications and equivalent configurations within the scope of the attached claims.
  • the above-described embodiments have been described in detail in order to explain the present invention in an easy-to-understand manner, and the present invention is not necessarily limited to those having all the described configurations.
  • part of the configuration of one embodiment may be replaced with the configuration of another embodiment.
  • the configuration of another embodiment may be added to the configuration of one embodiment.
  • other configurations may be added, deleted, or replaced with respect to a part of the configuration of each embodiment.
  • each configuration, function, processing unit, processing means, etc. described above may be implemented in hardware, for example, by designing a part or all of them with an integrated circuit, and the processor implements each function. It may be realized by software by interpreting and executing a program to execute.
  • Storage devices such as memory, hard disk, SSD (Solid State Drive), or IC (Integrated Circuit) card, SD card, DVD (Digital Versatile Disc) Can be stored on media.
  • control lines and information lines indicate those that are considered necessary for explanation, and do not necessarily indicate all the control lines and information lines necessary for implementation. In practice, it can be considered that almost all configurations are interconnected.

Landscapes

  • Business, Economics & Management (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Engineering & Computer Science (AREA)
  • Development Economics (AREA)
  • Strategic Management (AREA)
  • General Business, Economics & Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Game Theory and Decision Science (AREA)
  • Data Mining & Analysis (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

An estimation apparatus, which has a processor that executes programs and storage devices that stores the programs, is characterized by executing an acquisition process of acquiring person-flow-direction data indicating a time-series movement trajectory of a person in an area containing a section having one or more surfaces the person would face, a calculation process of calculating the stay time of the person at the surfaces on the basis of the person-flow-direction data acquired in the acquisition process and the locations of the surfaces in the section, and an output process of outputting the stay time calculated by the calculation process.

Description

推定装置および推定方法Estimation device and estimation method 参照による取り込みImport by reference
　本出願は、令和４年（２０２２年）２月３日に出願された日本出願である特願２０２２-０１５６０９の優先権を主張し、その内容を参照することにより、本出願に取り込む。 This application claims the priority of Japanese Patent Application No. 2022-015609 filed on February 3, 2022, and incorporates the contents thereof into the present application by reference.
　本発明は、データを推定する推定装置および推定方法に関する。 The present invention relates to an estimation device and an estimation method for estimating data.
　下記特許文献１は、ユーザが第１商品に手を伸ばしたことを示す情報を入力する第１入力部と、ユーザの属性を推定する推定部と、ユーザが第１商品を購入しなかった場合に、ユーザの属性に対応付けて第１商品に手を伸ばしたことを記憶する記憶部と、を備え、ユーザが第１商品と同一カテゴリの第２商品を購入した場合に、記憶部は、第２商品の情報を記憶する電子機器を開示する。 Patent Literature 1 below describes a first input unit for inputting information indicating that a user reached out for a first product, an estimation unit for estimating attributes of the user, and a case where the user did not purchase the first product. and a storage unit that stores that the user reaches out for the first product in association with the user's attribute, and when the user purchases the second product in the same category as the first product, the storage unit: An electronic device that stores information of a second product is disclosed.
特開２０１４－０２１７９５号公報JP 2014-021795 A
　しかしながら、上述した従来技術では、人流に基づく棚面前の人の滞在時間や棚割設定については、考慮されていない。 However, the above-described conventional technology does not take into account the length of time people stay in front of the shelves and the planogram setting based on the flow of people.
　本発明は、面滞在時間の推定容易化を図ることを目的とする。 The purpose of the present invention is to facilitate estimation of surface stay time.
　本開示技術は、プログラムを実行するプロセッサと、前記プログラムを記憶する記憶デバイスと、を有する推定装置であって、前記プロセッサは、人が対面する１以上の面を有する区画が配置されている領域において、時系列な前記人の移動軌跡を示す人流方向データを取得する取得処理と、前記取得処理によって取得された人流方向データと、前記区画内における前記面の位置と、に基づいて、前記面における前記人の滞在時間を算出する算出処理と、前記算出処理によって算出された滞在時間を出力する出力処理と、を実行することを特徴とする。 The technology disclosed herein is an estimating device having a processor that executes a program and a storage device that stores the program, wherein the processor is an area in which partitions having one or more surfaces facing a person are arranged. in the plane, based on an acquisition process for acquiring people flow direction data indicating the movement trajectory of the person in time series, the people flow direction data acquired by the acquisition process, and the position of the plane in the section; and an output process for outputting the stay time calculated by the calculation process.
　本発明の代表的な実施の形態によれば、面滞在時間の推定容易化を図ることができる。前述した以外の課題、構成及び効果は、以下の実施例の説明により明らかにされる。 According to the representative embodiment of the present invention, it is possible to facilitate estimation of face stay time. Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments.
図１は、実施例１にかかる分析システムのシステム構成例１を示す説明図である。FIG. 1 is an explanatory diagram of a system configuration example 1 of an analysis system according to a first embodiment. 図２は、区画を示す平面図である。FIG. 2 is a plan view showing compartments. 図３は、区画周辺環境データ記憶部の一例を示す説明図である。FIG. 3 is an explanatory diagram showing an example of a section peripheral environment data storage unit. 図４は、面配置データ記憶部の一例を示す説明図である。FIG. 4 is an explanatory diagram showing an example of a surface layout data storage unit. 図５は、コンピュータのハードウェア構成例を示すブロック図である。FIG. 5 is a block diagram showing a hardware configuration example of a computer. 図６は、登録画面の一例を示す説明図である。FIG. 6 is an explanatory diagram showing an example of a registration screen. 図７は、面滞在時間推定モードを表示する表示画面の一例を示す説明図である。FIG. 7 is an explanatory diagram showing an example of a display screen displaying a surface stay time estimation mode. 図８は、推奨棚割推定モードを表示する表示画面の一例を示す説明図である。FIG. 8 is an explanatory diagram showing an example of a display screen displaying the recommended planogram estimation mode. 図９は、実施例１にかかる分析システムのシステム構成例２を示す説明図である。FIG. 9 is an explanatory diagram of system configuration example 2 of the analysis system according to the first embodiment. 図１０は、実施例１にかかる分析システムのシステム構成例３を示す説明図である。FIG. 10 is an explanatory diagram of a system configuration example 3 of the analysis system according to the first embodiment. 図１１は、クラウドサーバによる面滞在時間推定処理手順例を示すフローチャートである。FIG. 11 is a flow chart showing an example of a screen staying time estimation processing procedure by a cloud server. 図１２は、客単価推定部による客単価推定処理手順例を示すフローチャートである。FIG. 12 is a flowchart showing an example of a procedure for estimating the unit price per customer by the unit for estimating the unit price per customer. 図１３は、棚割推定部による棚割推定処理手順例を示すフローチャートである。FIG. 13 is a flowchart illustrating an example of a shelf allocation estimation processing procedure by a shelf allocation estimation unit. 図１４は、商品群のグルーピング例を示す説明図である。FIG. 14 is an explanatory diagram showing an example of grouping product groups. 図１５は、グループに所属する商品群の一例を示す説明図である。FIG. 15 is an explanatory diagram showing an example of product groups belonging to a group. 図１６は、商品追加例１を示す説明図である。FIG. 16 is an explanatory diagram showing example 1 of product addition. 図１７は、商品追加例２を示す説明図である。FIG. 17 is an explanatory diagram of product addition example 2. As shown in FIG.
　＜分析システムのシステム構成例１＞
　図１は、実施例１にかかる分析システムのシステム構成例１を示す説明図である。構成例１は、面滞在時間推定モデルを用いて面滞在時間を推定する場合の構成の一例である。面滞在時間推定モデルとは、棚１３２の正面である棚面前での人１３４の滞在時間を推定する学習モデルである。 <System configuration example 1 of analysis system>
FIG. 1 is an explanatory diagram of a system configuration example 1 of an analysis system according to a first embodiment. Configuration example 1 is an example of a configuration for estimating a screen stay time using a screen stay time estimation model. The surface stay time estimation model is a learning model for estimating the stay time of the person 134 in front of the shelf surface, which is the front surface of the shelf 132 .
　クラウドシステム１００は、クラウドサーバ１０１と、商業施設１０２Ａ、１０２Ｂ、１０２Ｃ、…（これらを区別しない場合は、単に商業施設１０２）と、商業施設１０２Ａ、１０２Ｂ、１０２Ｃ、…の各々の情報処理装置１０３Ａ、１０３Ｂ、１０３Ｃ、…（これらを区別しない場合は、単に情報処理装置１０３）と、インターネット、ＬＡＮ（Ｌｏｃａｌ　Ａｒｅａ　Ｎｅｔｗｏｒｋ）、ＷＡＮ（Ｗｉｄｅ　Ａｒｅａ　Ｎｅｔｗｏｒｋ）などのネットワーク１０４を介して通信可能に接続される。 The cloud system 100 includes a cloud server 101, commercial facilities 102A, 102B, 102C, . , 103B, 103C, . .
　商業施設１０２は、区画１３１と、区画１３１近傍に設置された人流データ取得部１３０とを有する。区画１３１は、商品を陳列する棚１３２を有する。図１の区画１３１では、コの字形状に３つの棚１３２が配置されている。また、区画１３１において区画１３１外の通路と連結する部分を開口１３３と称す。開口１３３は、人１３４が出入りする部分である。 The commercial facility 102 has a section 131 and a people flow data acquisition unit 130 installed near the section 131 . Compartment 131 has shelves 132 for displaying merchandise. In the section 131 of FIG. 1, three shelves 132 are arranged in a U-shape. A portion of the section 131 that connects with a passage outside the section 131 is referred to as an opening 133 . The opening 133 is a portion through which a person 134 enters and exits.
　人流データ取得部１３０は、区画１３１周辺および区画１３１内の人１３４の人流データ１１１を取得し、取得した人流データ１１１をネットワーク１０４を介してクラウドサーバ１０１に送信する。人流データ取得部１３０は、たとえば、ＬｉＤＡＲであり、物体の形状や物体までの距離、物体の移動方向を検出する。これにより、たとえば、人流データ取得部１３０は、人１３４ごとに形状を特定し、特定した形状の人１３４がどの方向にどれくらい移動したかを示す人流データ１１１を取得する。すなわち、人流データ１１１は、人１３４の時系列な位置情報である。ＬｉＤＡＲは、人１３４がＸＹ平面方向に腕を伸ばしたか否かを検出することもできるため、人流データ１１１は、各々の人１３４がＸＹ平面方向に腕を伸ばした回数も含む。 The people flow data acquisition unit 130 acquires the people flow data 111 of the people 134 around and within the block 131 and transmits the acquired people flow data 111 to the cloud server 101 via the network 104 . The people flow data acquisition unit 130 is, for example, LiDAR, and detects the shape of an object, the distance to the object, and the moving direction of the object. As a result, for example, the people flow data acquisition unit 130 identifies the shape of each person 134 and obtains the people flow data 111 indicating how far and in which direction the person 134 having the identified shape moves. That is, the people flow data 111 is time-series position information of the people 134 . Since LiDAR can also detect whether a person 134 has extended their arms in the XY plane direction, the people flow data 111 also includes the number of times each person 134 has extended their arms in the XY plane direction.
　情報処理装置１０３は、商業施設１０２の棚割管理者またはテナント管理者（以下、単に、管理者）が操作するコンピュータであり、クラウドサーバ１０１にデータを入力したり、クラウドサーバ１０１からのデータを出力したりする。具体的には、たとえば、情報処理装置１０３Ａは、商業施設１０２Ａの管理者が操作するコンピュータであり、情報処理装置１０３Ｂは、商業施設１０２Ｂの管理者が操作するコンピュータであり、情報処理装置１０３Ｃは、商業施設１０２Ｃの管理者が操作するコンピュータである。 The information processing device 103 is a computer operated by a shelf allocation manager or a tenant manager (hereinafter simply referred to as manager) of the commercial facility 102, and inputs data to the cloud server 101 and receives data from the cloud server 101. output. Specifically, for example, the information processing device 103A is a computer operated by the manager of the commercial facility 102A, the information processing device 103B is a computer operated by the manager of the commercial facility 102B, and the information processing device 103C is a computer operated by the manager of the commercial facility 102B. , a computer operated by the manager of the commercial facility 102C.
　クラウドサーバ１０１は、区画周辺環境データ記憶部１１３と、面配置データ記憶部１１４と、面滞在時間推定モデル記憶部１１５と、を有する。区画周辺環境データ記憶部１１３は、区画周辺環境データを記憶する。区画周辺環境データとは、区画１３１の周辺環境を示すデータであり、たとえば、区画１３１がある商業施設１０２の規模、商業施設１０２の建造（エンクローズドまたはオープン）、立地、人１３４の平均通行量を示す。 The cloud server 101 has a block surrounding environment data storage unit 113 , a surface layout data storage unit 114 , and a surface stay time estimation model storage unit 115 . The compartment surrounding environment data storage unit 113 stores the compartment surrounding environment data. The block surrounding environment data is data indicating the surrounding environment of the block 131. For example, the scale of the commercial facility 102 in which the block 131 is located, the construction (enclosed or open) of the commercial facility 102, the location, the average traffic volume of the people 134 indicates
　面配置データ記憶部１１４は、面配置データを記憶する。面配置データとは、棚１３２の棚面の配置を規定するデータである。面滞在時間推定モデル記憶部１１５は、面滞在時間推定モデル１１６を区画１３１ごとに記憶する。面滞在時間推定モデル１１６は、区画１３１ごとに存在し、人流方向データ１１２、区画周辺環境データおよび面配置データが入力されると、推定値である面滞在時間１１７を区画１３１内の棚面ごとに出力する。 The surface layout data storage unit 114 stores surface layout data. The surface layout data is data that defines the layout of the shelf surface of the shelf 132 . The page stay time estimation model storage unit 115 stores the page stay time estimation model 116 for each section 131 . The plane stay time estimation model 116 exists for each section 131, and when the people flow direction data 112, the section surrounding environment data, and the plane arrangement data are input, the plane stay time 117, which is an estimated value, is calculated for each shelf surface in the section 131. output to
　また、クラウドサーバ１０１は、人流方向分析部１２１と、棚割推定部１２２と、客単価推定部１２３と、を有する。人流方向分析部１２１は、人流データ１１１を分析して人１３４の人流方向データ１１２を出力する。人流方向データ１１２は、たとえば、ＸＹ平面上を移動する人１３４の時系列な移動軌跡である。また、ＬｉＤＡＲは、人１３４がＸＹ平面方向に腕を伸ばしたか否かを検出することもできる。また、棚面Ｓ＃の幅方向および高さ方向により、商品群の棚割が決まっているため、人流方向データ１１２は、各々の人１３４が棚面Ｓ＃の商品ごとに腕を伸ばした回数も含む。 The cloud server 101 also has a people flow direction analysis unit 121 , a planogram estimation unit 122 , and a customer unit price estimation unit 123 . The people flow direction analysis unit 121 analyzes the people flow data 111 and outputs the people flow direction data 112 of the person 134 . The people flow direction data 112 is, for example, a time-series movement trajectory of a person 134 moving on the XY plane. LiDAR can also detect whether the person 134 has extended their arm in the XY plane direction. In addition, since the planogram of the product group is determined by the width direction and the height direction of the shelf surface S#, the people flow direction data 112 indicates the number of times each person 134 stretches his or her arm for each product on the shelf surface S#. Also includes
　棚割推定部１２２は、面滞在時間１１７に基づいて、棚割を推定し、推定結果として棚割データ１１８を出力する。客単価推定部１２３は、面滞在時間１１７に基づいて、区画１３１における客単価を算出する。なお、客単価には、ある期間内における区画１３１での全商品の売上金額に基づく客一人当たりの売上金額と、ある期間内における区画１３１での商品ごとの売上金額に基づく客一人当たりの商品ごとの売上金額と、がある。これらの少なくとも１つの客単価を客単価データ１１９と称す。また、クラウドサーバ１０１は、面滞在時間１１７と商業施設１０２固有の係数とに基づいて、区画１３１が配置される商業施設１０２内の場所の賃料を推定してもよい。 The planogram estimation unit 122 estimates the planogram based on the surface stay time 117 and outputs the planogram data 118 as an estimation result. The customer unit price estimation unit 123 calculates the customer unit price in the section 131 based on the face stay time 117 . The unit price per customer includes the amount of sales per customer based on the amount of sales of all products in section 131 within a certain period, and the amount of sales per customer based on the amount of sales for each item in section 131 within a certain period. There is a sales amount for each and At least one of these customer unit prices is referred to as customer unit price data 119 . The cloud server 101 may also estimate the rent of the location within the commercial facility 102 where the parcel 131 is located based on the visit time 117 and a coefficient specific to the commercial facility 102 .
　＜棚面＞
　図２は、区画１３１を示す平面図である。図２では、３種類の区画１３１Ａ～１３１Ｃを示している。区画１３１Ａ～１３１Ｃを区別しない場合は、区画１３１と称す。Ｘは、開口１３３に対して直交する方向であり、Ｙは、Ｘに直交する人流である。区画１３１は、棚面Ｓ＃と領域Ａ＃とを有する。＃は番号である。領域Ａ＃は同一番号＃の棚面Ｓ＃に対応する。具体的には、たとえば、人１３４は、領域Ａ＃に滞在するときに棚面Ｓ＃と対面する。棚面Ｓ＃は、棚１３２の間口となる棚１３２の正面である。なお、棚面Ｓ＃は、１以上の棚１３２の正面が面一になった面であるが、本例では説明を単純化するため、棚面Ｓ＃は１つの棚１３２の正面として説明する。 <Shelf surface>
FIG. 2 is a plan view showing the partition 131. As shown in FIG. FIG. 2 shows three types of compartments 131A-131C. Sections 131A-131C are referred to as section 131 when not distinguished. X is the direction orthogonal to the opening 133 and Y is the people flow orthogonal to X. Section 131 has shelf surface S# and area A#. # is a number. Area A# corresponds to shelf surface S# with the same number #. Specifically, for example, person 134 faces shelf surface S# when staying in area A#. The shelf surface S# is the front surface of the shelf 132 which is the frontage of the shelf 132 . Note that although the shelf surface S# is a surface in which the front surfaces of one or more shelves 132 are flush with each other, in this example, the shelf surface S# will be described as the front surface of one shelf 132 in order to simplify the description. .
　＜区画周辺環境データ記憶部１１３＞
　図３は、区画周辺環境データ記憶部１１３の一例を示す説明図である。区画周辺環境データ記憶部１１３は、区画周辺環境データを、たとえば、テーブル形式で保持する。区画周辺環境データ記憶部１１３は、フィールドとして、施設ＩＤ３０１と、規模３０２と、建屋３０３と、立地３０４と、平均通行量３０５と、を有する。同一行の各フィールドの値の組み合わせが、１つの商業施設１０２に関する区画周辺環境データを規定するエントリとなる。 <Section Surrounding Environment Data Storage Unit 113>
FIG. 3 is an explanatory diagram showing an example of the partition surrounding environment data storage unit 113. As shown in FIG. The partition surrounding environment data storage unit 113 holds the partition surrounding environment data in, for example, a table format. The block peripheral environment data storage unit 113 has facility ID 301, scale 302, building 303, location 304, and average traffic volume 305 as fields. A combination of field values in the same row constitutes an entry that defines block surrounding environment data for one commercial facility 102 .
　施設ＩＤ３０１は、商業施設１０２を一意に特定する識別情報である。規模３０２は、商業施設１０２がリージョナル型、コミュニティ型、およびネイバーフッド型のいずれかに該当するかを示す情報である。建屋３０３は、商業施設１０２の構造がエンクローズドおよびオープンのいずれに該当するかを示す情報である。立地３０４は、商業施設１０２の設置場所が駅チカ、都市、および郊外のいずれに該当するかを示す情報である。駅チカとは、商業施設１０２の最寄り駅の近くまたは最寄り駅の地下を意味する。平均通行量３０５は、商業施設１０２における単位時間あたりの人１３４の通行量（人／時）である。なお、規模３０２、建屋３０３および立地３０４は、たとえば、ワンホットエンコーディングされている。 The facility ID 301 is identification information that uniquely identifies the commercial facility 102 . The scale 302 is information indicating whether the commercial facility 102 is a regional type, a community type, or a neighborhood type. The building 303 is information indicating whether the structure of the commercial facility 102 is closed or open. The location 304 is information indicating whether the installation location of the commercial facility 102 is close to a station, a city, or a suburb. Near the station means near the nearest station of the commercial facility 102 or underground of the nearest station. The average traffic volume 305 is the traffic volume of the people 134 per unit time (person/hour) in the commercial facility 102 . Note that the scale 302, building 303 and location 304 are, for example, one-hot encoded.
　＜面配置データ記憶部１１４＞
　図４は、面配置データ記憶部１１４の一例を示す説明図である。面配置データ記憶部１１４は、区画１３１の面配置データ４００を、たとえば、テーブル形式で保持する。面配置データ４００は、フィールドとして、施設ＩＤ３０１と、区画ＩＤ４０２と、棚面ＩＤ４０３と、配置情報４０４と、を有する。 <Surface layout data storage unit 114>
FIG. 4 is an explanatory diagram showing an example of the surface arrangement data storage unit 114. As shown in FIG. The surface layout data storage unit 114 holds the surface layout data 400 of the section 131 in, for example, a table format. The surface arrangement data 400 has facility ID 301, section ID 402, shelf surface ID 403, and arrangement information 404 as fields.
　区画ＩＤ４０２は、区画１３１を一意に特定する識別情報である。棚面ＩＤ４０３は、棚面Ｓ＃を一意に特定する識別情報である。配置情報４０４は、棚面Ｓ＃の区画１３１内の位置および棚面Ｓ＃の法線方向を規定する情報である。また、配置情報４０４は、棚面ＩＤ４０３で特定される棚面における棚１３２の棚段の段数や各棚段の間口の形状や容積といったデータも保持する。 The partition ID 402 is identification information that uniquely identifies the partition 131 . The shelf surface ID 403 is identification information that uniquely identifies the shelf surface S#. The arrangement information 404 is information that defines the position of the shelf surface S# within the section 131 and the normal direction of the shelf surface S#. The arrangement information 404 also holds data such as the number of shelves of the shelf 132 on the shelf surface specified by the shelf surface ID 403 and the shape and volume of the frontage of each shelf.
　＜コンピュータのハードウェア構成＞
　つぎに、クラウドサーバ１０１および情報処理装置１０３（以下、総称して、コンピュータ５００）のハードウェア構成について説明する。 <Computer hardware configuration>
Next, hardware configurations of the cloud server 101 and the information processing device 103 (hereinafter collectively referred to as the computer 500) will be described.
　図５は、コンピュータ５００のハードウェア構成例を示すブロック図である。コンピュータ５００は、プロセッサ５０１と、記憶デバイス５０２と、入力デバイス５０３と、出力デバイス５０４と、通信インターフェース（通信ＩＦ）５０５と、を有する。プロセッサ５０１、記憶デバイス５０２、入力デバイス５０３、出力デバイス５０４、および通信ＩＦ５０５は、バス５０６により接続される。プロセッサ５０１は、コンピュータ５００を制御する。記憶デバイス５０２は、プロセッサ５０１の作業エリアとなる。また、記憶デバイス５０２は、各種プログラムやデータを記憶する非一時的なまたは一時的な記録媒体である。記憶デバイス５０２としては、たとえば、ＲＯＭ（Ｒｅａｄ　Ｏｎｌｙ　Ｍｅｍｏｒｙ）、ＲＡＭ（Ｒａｎｄｏｍ　Ａｃｃｅｓｓ　Ｍｅｍｏｒｙ）、ＨＤＤ（Ｈａｒｄ　Ｄｉｓｋ　Ｄｒｉｖｅ）、フラッシュメモリがある。入力デバイス５０３は、データを入力する。入力デバイス５０３としては、たとえば、キーボード、マウス、タッチパネル、テンキー、スキャナ、マイク、センサがある。出力デバイス５０４は、データを出力する。出力デバイス５０４としては、たとえば、ディスプレイ、プリンタ、スピーカがある。通信ＩＦ５０５は、ネットワークと接続し、データを送受信する。 FIG. 5 is a block diagram showing a hardware configuration example of the computer 500. As shown in FIG. The computer 500 has a processor 501 , a storage device 502 , an input device 503 , an output device 504 and a communication interface (communication IF) 505 . Processor 501 , storage device 502 , input device 503 , output device 504 and communication IF 505 are connected by bus 506 . Processor 501 controls computer 500 . A storage device 502 serves as a work area for the processor 501 . Also, the storage device 502 is a non-temporary or temporary recording medium that stores various programs and data. Examples of the storage device 502 include ROM (Read Only Memory), RAM (Random Access Memory), HDD (Hard Disk Drive), and flash memory. The input device 503 inputs data. The input device 503 includes, for example, a keyboard, mouse, touch panel, numeric keypad, scanner, microphone, and sensor. The output device 504 outputs data. Output devices 504 include, for example, displays, printers, and speakers. Communication IF 505 connects to a network and transmits and receives data.
　なお、図１に示した人流方向分析部１２１、棚割推定部１２２、および客単価推定部１２３は、具体的には、たとえば、記憶デバイス５０２に記憶されたプログラムをプロセッサ５０１に実行させることにより実現される機能である。また、図１に示した区画周辺環境データ記憶部１１３、面配置データ記憶部１１４、および面滞在時間推定モデル記憶部１１５は、記憶デバイス５０２に記憶される。 Specifically, the people flow direction analysis unit 121, the planogram estimation unit 122, and the customer unit price estimation unit 123 shown in FIG. It is a function to be realized. Also, the block surrounding environment data storage unit 113, the surface arrangement data storage unit 114, and the surface stay time estimation model storage unit 115 shown in FIG.
　＜表示画面例＞
　つぎに、情報処理装置１０３に表示される表示画面例について説明する。 <Display screen example>
Next, an example of a display screen displayed on the information processing apparatus 103 will be described.
　図６は、登録画面の一例を示す説明図である。登録画面６００は、情報処理装置１０３が担当する区画１３１に関する情報を登録したり、登録済みの情報を修正したりするための画面である。登録画面６００は、区画ＩＤ入力欄６０１と、区画名選択プルダウン６０２と、区画写真表示領域６０３と、区画周辺環境データ６０４と、面配置レイアウト６０５と、登録ボタン６０６と、修正ボタン６０７と、を有する。 FIG. 6 is an explanatory diagram showing an example of the registration screen. The registration screen 600 is a screen for registering information about the section 131 managed by the information processing apparatus 103 and for correcting registered information. The registration screen 600 includes a section ID input field 601, a section name selection pull-down 602, a section photo display area 603, section surrounding environment data 604, a surface arrangement layout 605, a registration button 606, and a correction button 607. have.
　区画ＩＤ入力欄６０１は、管理者が入力デバイス５０３を操作して区画ＩＤ４０２を入力するための領域である。登録ボタン６０６の押下により、入力された文字列が区画ＩＤ４０２としてクラウドサーバ１０１に登録される。修正ボタン６０７の押下により、修正後の文字列が、他の区画ＩＤ４０２と重複しない限り、新たな区画ＩＤ４０２としてクラウドサーバ１０１に更新される。 The section ID input field 601 is an area for the administrator to operate the input device 503 to input the section ID 402 . By pressing the registration button 606 , the entered character string is registered in the cloud server 101 as the partition ID 402 . By pressing the correction button 607 , the corrected character string is updated to the cloud server 101 as a new section ID 402 unless it overlaps with another section ID 402 .
　区画名選択プルダウン６０２は、区画１３１の一覧を表示して区画１３１の選択を受け付けるユーザインタフェースである。登録画面６００では、区画名選択プルダウン６０２は、区画ＩＤ４０２が未登録の区画１３１の名称（区画名）が入力される領域となる。登録ボタン６０６の押下により、入力された文字列が区画名として、区画ＩＤ入力欄６０１に入力された文字列（区画ＩＤ４０２）と関連付けて、クラウドサーバ１０１に登録される。修正ボタン６０７の押下により、修正後の文字列が、区画名として区画ＩＤ４０２と関連付けてクラウドサーバ１０１に更新される。 The block name selection pull-down 602 is a user interface that displays a list of blocks 131 and accepts selection of the blocks 131 . In the registration screen 600, the section name selection pull-down 602 is an area for inputting the name (section name) of the section 131 whose section ID 402 is not registered. By pressing the register button 606 , the entered character string is registered in the cloud server 101 as a partition name in association with the character string (partition ID 402 ) entered in the partition ID input field 601 . By pressing the correction button 607, the corrected character string is associated with the section ID 402 as the section name, and the cloud server 101 is updated.
　区画写真表示領域６０３は、区画１３１の写真（区画写真）である画像データを表示する領域である。情報処理装置１０３内の画像データを、入力デバイス５０３で操作（ドラッグアンドドロップ）することにより、当該画像データが、区画写真として区画写真表示領域６０３に表示される。登録ボタン６０６の押下により、表示中の画像データが区画写真として、区画ＩＤ入力欄６０１に入力された文字列（区画ＩＤ４０２）と関連付けて、クラウドサーバ１０１に登録される。修正ボタン６０７の押下により、差替え後の画像データが、差替え後の区画写真として区画ＩＤ４０２と関連付けてクラウドサーバ１０１に更新される。 A section photo display area 603 is an area for displaying image data that is a picture of the section 131 (section picture). By operating (drag-and-drop) image data in the information processing apparatus 103 with the input device 503 , the image data is displayed as a section photograph in the section photograph display area 603 . By pressing the registration button 606, the image data being displayed is registered in the cloud server 101 as a section photograph in association with the character string (section ID 402) entered in the section ID input field 601. FIG. By pressing the correction button 607, the image data after replacement is updated in the cloud server 101 in association with the section ID 402 as the section photograph after replacement.
　区画周辺環境データ６０４は、区画１３１が存在する商業施設１０２に関する区画周辺環境データである。なお、どの区画１３１がどの商業施設１０２に存在するかについては、面配置データ記憶部１１４の面配置データ４００で管理されている。 The block surrounding environment data 604 is the block surrounding environment data related to the commercial facility 102 in which the block 131 exists. Which section 131 exists in which commercial facility 102 is managed by the surface layout data 400 of the surface layout data storage unit 114 .
　情報処理装置１０３は、対応する商業施設１０２の施設ＩＤ３０１でクラウドサーバ１０１に要求することにより、当該商業施設１０２の区画周辺環境データ６０４を取得して、登録画面６００に表示する。 By requesting the cloud server 101 with the facility ID 301 of the corresponding commercial facility 102 , the information processing device 103 acquires the section surrounding environment data 604 of the commercial facility 102 and displays it on the registration screen 600 .
　面配置レイアウト６０５は、例として区画１３１Ａを平面視した図形データであり、面配置データ４００により作成される。 The surface arrangement layout 605 is, for example, figure data obtained by planarly viewing the section 131A, and is created from the surface arrangement data 400. FIG.
　登録ボタン６０６は、区画名、区画写真、区画周辺環境データ６０４（またはその施設ＩＤ３０１でもよい）、および面配置レイアウト６０５を区画ＩＤ４０２と関連付けてクラウドサーバ１０１に登録するためのユーザインタフェースである。 A registration button 606 is a user interface for registering the parcel name, parcel photograph, parcel surrounding environment data 604 (or its facility ID 301), and surface layout layout 605 in the cloud server 101 in association with the parcel ID 402.
　修正ボタン６０７は、区画ＩＤ４０２、区画名、区画写真、および面配置レイアウト６０５を修正してクラウドサーバ１０１に更新するためのユーザインタフェースである。 A correction button 607 is a user interface for correcting the section ID 402 , section name, section photograph, and surface layout 605 and updating the cloud server 101 .
　図７は、面滞在時間推定モードを表示する表示画面の一例を示す説明図である。表示画面７００は、推定対象区画ＩＤ入力欄７０１と、推定対象区画名選択プルダウン７０２と、第１推定条件選択プルダウン７０３と、第２推定条件選択プルダウン７０４と、面滞在時間ヒートマップ７０５と、面滞在時間一覧７０６と、実行ボタン７０７と、モード変更ボタン７０８と、を有する。 FIG. 7 is an explanatory diagram showing an example of a display screen displaying the time-of-visit estimation mode. The display screen 700 includes an estimation target section ID input field 701, an estimation target section name selection pull-down 702, a first estimation condition selection pull-down 703, a second estimation condition selection pull-down 704, a surface stay time heat map 705, a surface It has a stay time list 706 , an execution button 707 and a mode change button 708 .
　推定対象区画ＩＤ入力欄７０１は、管理者が入力デバイス５０３を操作して推定対象区画ＩＤを入力するための領域である。推定対象区画ＩＤは、推定対象区画を一意に特定する識別情報である。推定対象区画は、面滞在時間の推定対象となる区画１３１である。 The estimated target section ID input field 701 is an area for the administrator to operate the input device 503 to input the estimated target section ID. The estimation target section ID is identification information that uniquely identifies the estimation target section. The estimation target section is the section 131 that is the target for estimating the surface stay time.
　推定対象区画名選択プルダウン７０２は、区画１３１の一覧を表示して推定対象区画の選択を受け付けるユーザインタフェースである。 The estimation target block name selection pull-down 702 is a user interface that displays a list of blocks 131 and accepts the selection of an estimation target block.
　第１推定条件選択プルダウン７０３は、第１推定条件の一覧を表示して第１推定条件の選択を受け付けるユーザインタフェースである。第１推定条件は、平日、休日、日曜日～月曜日といった曜日である。第１推定条件は、条件設定なしに設定することもできる。 The first estimation condition selection pulldown 703 is a user interface that displays a list of first estimation conditions and accepts selection of the first estimation condition. The first estimation conditions are weekdays, holidays, and days of the week such as Sunday to Monday. The first estimation condition can also be set without condition setting.
　第２推定条件選択プルダウン７０４は、第２推定条件の一覧を表示して第２推定条件の選択を受け付けるユーザインタフェースである。第２推定条件は、1日当たり、１週間あたり、１か月あたりといった期間である。第２推定条件は、条件設定なしに設定することもできる。 The second estimation condition selection pulldown 704 is a user interface that displays a list of second estimation conditions and accepts selection of the second estimation condition. The second estimation condition is a period such as per day, per week, or per month. The second estimation condition can also be set without condition setting.
　面滞在時間ヒートマップ７０５は、第１推定条件および第２推定条件を充足した推定対象区画における面滞在時間１１７のヒートマップである。すなわち、色の濃淡により面滞在時間１１７の長さが表示される。面滞在時間ヒートマップ７０５は、人流方向データ１１２を用いて、クラウドサーバ１０１が面配置レイアウト６０５に面滞在時間１１７を示す色の濃淡を重畳することにより作成される。 The face staying time heat map 705 is a heat map of the face staying time 117 in the estimation target section that satisfies the first estimation condition and the second estimation condition. That is, the length of the surface staying time 117 is displayed by the shade of color. The face staying time heat map 705 is created by the cloud server 101 superimposing the color gradation indicating the face staying time 117 on the face arrangement layout 605 using the people flow direction data 112 .
　面滞在時間一覧７０６は、棚面Ｓ＃ごとの面滞在時間の推定値を示す推定結果である。具体的には、たとえば、面滞在時間一覧７０６は、棚面ＩＤ４０３と、面滞在時間１１７と、を有する。 A surface staying time list 706 is an estimation result showing an estimated value of the surface staying time for each shelf surface S#. Specifically, for example, the surface staying time list 706 has a shelf surface ID 403 and a surface staying time 117 .
　実行ボタン７０７は、押下により、第１推定条件選択プルダウン７０３および第２推定条件選択プルダウン７０４の第１推定条件および第２推定条件を充足した、推定対象区画ＩＤ入力欄７０１または推定対象区画名選択プルダウン７０２によって特定された推定対象区画について、面滞在時間１１７を棚面Ｓ＃ごとに算出し、面滞在時間ヒートマップ７０５および面滞在時間一覧７０６を表示するためのユーザインタフェースである。 When the execution button 707 is pressed, the estimation target section ID input field 701 or the estimation target section name selection that satisfies the first estimation condition and the second estimation condition of the first estimation condition selection pull-down menu 703 and the second estimation condition selection pull-down menu 704 is performed. This is a user interface for calculating surface stay time 117 for each shelf surface S# for the estimation target section specified by pull-down 702 and displaying surface stay time heat map 705 and surface stay time list 706 .
　モード変更ボタン７０８は、押下により、表示画面７００の表示内容を面滞在時間推定モードから推奨棚割推定モードに、および、推奨棚割推定モードから面滞在時間推定モードに、変更するためのユーザインタフェースである。 A mode change button 708 is a user interface for changing the display contents of the display screen 700 from the surface stay time estimation mode to the recommended planogram estimation mode, and from the recommended planogram estimation mode to the surface stay time estimation mode. is.
　図８は、推奨棚割推定モードを表示する表示画面７００の一例を示す説明図である。推奨棚割推定モードでは、表示画面７００は、推定対象区画ＩＤ入力欄７０１と、推定対象区画名選択プルダウン７０２と、棚種選択プルダウン８０１と、棚段数選択プルダウン８０２と、第１推定条件選択プルダウン７０３と、推定対象区画面配置データ８０４と、棚割レイアウト８０５と、実行ボタン８０７と、モード変更ボタン７０８と、を有する。 FIG. 8 is an explanatory diagram showing an example of a display screen 700 displaying the recommended planogram estimation mode. In the recommended shelf allocation estimation mode, the display screen 700 includes an estimation target section ID input field 701, an estimation target section name selection pull-down 702, a shelf type selection pull-down 801, a shelf number selection pull-down 802, and a first estimation condition selection pull-down. 703 , estimation target section plane layout data 804 , planogram layout 805 , execution button 807 , and mode change button 708 .
　棚種選択プルダウン８０１は、棚種の一覧を表示して棚種の選択を受け付けるユーザインタフェースである。棚種は、棚の種類である。棚段数選択プルダウン８０２は、棚段数の一覧を表示して棚段数の選択を受け付けるユーザインタフェースである。 The shelf type selection pull-down 801 is a user interface that displays a list of shelf types and accepts selection of a shelf type. The shelf type is the type of shelf. A shelf level selection pull-down 802 is a user interface for displaying a list of the number of shelves and accepting a selection of the number of shelves.
　推定対象区画面配置データ８０４は、推定対象区画の面配置レイアウト６０５である。棚割レイアウト８０５は、推定対象区画の各棚面Ｓ＃の棚割を示す図形データである。 The estimation target block surface layout data 804 is the surface layout layout 605 of the estimation target block. The planogram layout 805 is graphic data indicating the planogram of each shelf surface S# of the estimation target section.
　実行ボタン８０７は、押下により、棚種選択プルダウン８０１、棚段数選択プルダウン８０２および第１推定条件選択プルダウン７０３の棚種、棚段数、および第１推定条件を充足した、推定対象区画ＩＤ入力欄７０１または推定対象区画名選択プルダウン７０２によって特定された推定対象区画について、棚面Ｓ＃ごとの棚割を算出し、その算出結果として、棚割レイアウト８０５を表示するためのユーザインタフェースである。 When the execution button 807 is pressed, the estimation target section ID input field 701 that satisfies the shelf type, the number of shelves, and the first estimation condition of the shelf type selection pull-down 801, the number of shelves selection pull-down 802, and the first estimation condition selection pull-down 703 is displayed. Alternatively, it is a user interface for calculating the planogram for each shelf surface S# for the estimation target section specified by the estimation target section name selection pull-down 702 and displaying the planogram layout 805 as the calculation result.
　＜分析システムのシステム構成例２＞
　図９は、実施例１にかかるクラウドシステム１００のシステム構成例２を示す説明図である。構成例２は、面滞在時間推定モデルを学習する場合の構成の一例である。図９では、図１にはない構成について説明する。クラウドサーバ１０１は、人流方向データ記憶部９０１と、面滞在時間記憶部９０２と、面滞在時間推定モデル学習部９０３と、を有する。 <System configuration example 2 of analysis system>
FIG. 9 is an explanatory diagram of a system configuration example 2 of the cloud system 100 according to the first embodiment. Configuration example 2 is an example of a configuration for learning a surface stay time estimation model. In FIG. 9, a configuration not shown in FIG. 1 will be described. The cloud server 101 has a people flow direction data storage unit 901 , a face stay time storage unit 902 , and a face stay time estimation model learning unit 903 .
　人流方向データ記憶部９０１は、人流方向データ１１２を記憶する。面滞在時間記憶部９０２は、面滞在時間を記憶する。面滞在時間は、たとえば、各区画１３１の棚面Ｓ＃ごとの人１３４の滞在時間の実測値である。 The people flow direction data storage unit 901 stores the people flow direction data 112 . The page stay time storage unit 902 stores the page stay time. The face stay time is, for example, the actually measured value of the stay time of the person 134 for each shelf face S# of each section 131 .
　面滞在時間推定モデル学習部９０３は、人流方向データ１１２、区間周辺環境データ、および面配置データ４００における棚面Ｓ＃の配置情報４０４（位置および法線方向）を説明変数とし、面滞在時間を目的変数として、機械学習により面滞在時間推定モデル１１６を学習する。 The surface stay time estimation model learning unit 903 uses the people flow direction data 112, the section surrounding environment data, and the arrangement information 404 (position and normal direction) of the shelf surface S# in the surface arrangement data 400 as explanatory variables, and calculates the surface stay time. As an objective variable, the surface stay time estimation model 116 is learned by machine learning.
　なお、図９に示した面滞在時間推定モデル学習部９０３は、具体的には、たとえば、記憶デバイス５０２に記憶されたプログラムをプロセッサ５０１に実行させることにより実現される機能である。また、図９に示した人流方向データ記憶部９０１および面滞在時間記憶部９０２は、記憶デバイス５０２に記憶される。 It should be noted that the surface stay time estimation model learning unit 903 shown in FIG. 9 is, specifically, a function realized by causing the processor 501 to execute a program stored in the storage device 502, for example. In addition, the people flow direction data storage unit 901 and the surface stay time storage unit 902 shown in FIG. 9 are stored in the storage device 502 .
　＜分析システムのシステム構成例２＞
　図１０は、実施例１にかかるクラウドシステム１００のシステム構成例３を示す説明図である。構成例３は、棚割モデルを学習する場合の構成の一例である。図１０では、図１および図９にはない構成について説明する。クラウドサーバ１０１は、学習用棚割データ１００１と、学習用棚割データ記憶部１００２と、客単価記憶部１００３と、棚割推定モデル１００４と、棚割推定モデル記憶部１００５と、棚割推定モデル学習部１０１０と、を有する。 <System configuration example 2 of analysis system>
FIG. 10 is an explanatory diagram of a system configuration example 3 of the cloud system 100 according to the first embodiment. Configuration example 3 is an example of a configuration for learning a planogram model. In FIG. 10, a configuration not shown in FIGS. 1 and 9 will be described. The cloud server 101 includes a learning planogram data 1001, a learning planogram data storage unit 1002, a customer unit price storage unit 1003, a planogram estimation model 1004, a planogram estimation model storage unit 1005, and a planogram estimation model. and a learning unit 1010 .
　学習用棚割データ１００１は、学習用の棚割データである。棚割データは、どの棚面Ｓ＃のどの棚段に何が配置されているかを示すデータである。学習用棚割データ記憶部１００２は、学習用棚割データ１００１を記憶する。客単価記憶部１００３は、区画１３１ごとの客単価を記憶する。棚割推定モデル１００４は、棚割を推定する学習モデルである。棚割推定モデル記憶部１００５は、棚割推定モデル１００４を記憶する。 The planogram data for learning 1001 is planogram data for learning. The shelving allocation data is data indicating what is arranged on which shelf of which shelf surface S#. The planogram data storage unit 1002 for learning stores the planogram data 1001 for learning. The customer unit price storage unit 1003 stores the customer unit price for each section 131 . The planogram estimation model 1004 is a learning model for estimating the planogram. The planogram estimation model storage unit 1005 stores the planogram estimation model 1004 .
　棚割推定モデル学習部１０１０は、人流方向データ１１２、区間周辺環境データ、面配置データ４００における棚面Ｓ＃の配置情報４０４（位置および法線方向）、および客単価を説明変数とし、学習用棚割データ１００１を目的変数として、機械学習により棚割推定モデル１００４を学習する。 The planogram estimation model learning unit 1010 uses the people flow direction data 112, the section surrounding environment data, the arrangement information 404 (position and normal direction) of the shelf surface S# in the surface arrangement data 400, and the customer unit price as explanatory variables, and uses them as explanatory variables for learning. A planogram estimation model 1004 is learned by machine learning using the planogram data 1001 as an objective variable.
　なお、図９に示した棚割推定モデル学習部１０１０は、具体的には、たとえば、記憶デバイス５０２に記憶されたプログラムをプロセッサ５０１に実行させることにより実現される機能である。また、図９に示した学習用棚割データ記憶部１００２および客単価記憶部１００３は、記憶デバイス５０２に記憶される。 Note that the planogram estimation model learning unit 1010 shown in FIG. 9 is, specifically, a function realized by causing the processor 501 to execute a program stored in the storage device 502, for example. The learning planogram data storage unit 1002 and the customer unit price storage unit 1003 shown in FIG. 9 are stored in the storage device 502 .
　＜面滞在時間推定処理＞
　図１１は、クラウドサーバ１０１による面滞在時間推定処理手順例を示すフローチャートである。クラウドサーバ１０１は、情報処理装置１０３で入力された推定対象区画の区画ＩＤ４０２（以下、推定対象区画ＩＤ４０２）と、第１推定条件と、第２推定条件と、を、情報処理装置１０３から受信する（ステップＳ１１０１）。 <Surface Stay Time Estimation Processing>
FIG. 11 is a flow chart showing an example of a page stay time estimation processing procedure by the cloud server 101 . The cloud server 101 receives from the information processing device 103 the section ID 402 of the section to be estimated input by the information processing device 103 (hereinafter referred to as the section ID 402 to be estimated), the first estimation condition, and the second estimation condition. (Step S1101).
　クラウドサーバ１０１は、人流方向分析部１２１により、推定対象区画について、第１推定条件および第２推定条件に該当する人流方向データ１１２を取得する（ステップＳ１１０２）。 The cloud server 101 uses the people flow direction analysis unit 121 to acquire the people flow direction data 112 corresponding to the first estimation condition and the second estimation condition for the estimation target section (step S1102).
　クラウドサーバ１０１は、推定対象区画の面滞在時間推定モデル１１６に、ステップＳ１１０２で取得した人流方向データ１１２、区画周辺環境データおよび面配置データを入力して、推定値である面滞在時間１１７を算出する（ステップＳ１１０３）。 The cloud server 101 inputs the people flow direction data 112, the block surrounding environment data, and the surface arrangement data acquired in step S1102 to the surface stay time estimation model 116 of the estimation target block, and calculates the surface stay time 117, which is an estimated value. (step S1103).
　クラウドサーバ１０１は、面滞在時間１１７を面滞在時間ヒートマップ７０５と、面滞在時間一覧７０６とを生成し（ステップＳ１１０４）、情報処理装置１０３に送信する（ステップＳ１１０５）。これにより、図７に示したような表示画面７００が情報処理装置１０３に表示される。 The cloud server 101 generates a screen stay time heat map 705 and a screen stay time list 706 from the screen stay time 117 (step S1104), and transmits them to the information processing device 103 (step S1105). As a result, a display screen 700 as shown in FIG. 7 is displayed on the information processing apparatus 103 .
　＜客単価推定処理＞
　図１２は、客単価推定部１２３による客単価推定処理手順例を示すフローチャートである。クラウドサーバ１０１は、情報処理装置１０３から、客単価データ１１９の推定対象区画ＩＤ４０２および推定対象期間を取得する（ステップＳ１２０１）。 <Customer Unit Price Estimation Processing>
FIG. 12 is a flowchart showing an example of a procedure for estimating the unit price per customer by the unit price estimating unit 123 for customers. The cloud server 101 acquires the estimation target section ID 402 and the estimation target period of the customer unit price data 119 from the information processing device 103 (step S1201).
　クラウドサーバ１０１は、ステップＳ１２０１で選択された推定対象区画について、推定対象期間内において、人流方向データ１１２に基づいて、人１３４が棚面Ｓ＃に腕を伸ばした回数を取得する（ステップＳ１２０２）。具体的には、たとえば、クラウドサーバ１０１は、棚面Ｓ＃における商品の陳列位置ごとに、人１３４が棚面Ｓ＃に腕を伸ばした回数を取得する。 The cloud server 101 obtains the number of times that the person 134 stretched his or her arm toward the shelf surface S# within the estimation target period based on the people flow direction data 112 for the estimation target section selected in step S1201 (step S1202). . Specifically, for example, cloud server 101 acquires the number of times that person 134 has extended his/her arm to shelf surface S# for each product display position on shelf surface S#.
　クラウドサーバ１０１は、推定対象期間内に推定対象区画に陳列された商品の売上金額を情報処理装置１０３から取得する（ステップＳ１２０３）。 The cloud server 101 acquires from the information processing device 103 the sales amount of the products displayed in the estimation target section within the estimation target period (step S1203).
　クラウドサーバ１０１は、ステップＳ１２０３で取得した推定対象期間内における推定対象区画に陳列された商品ごとに、当該商品の売上金額を、当該商品の陳列位置に腕を伸ばした回数で除することにより、商品ごとの客単価を算出する（ステップＳ１２０４）。 The cloud server 101 divides the sales amount of each product displayed in the estimation target section within the estimation target period acquired in step S1203 by the number of times the arm is stretched to the display position of the product. A customer unit price for each product is calculated (step S1204).
　クラウドサーバ１０１は、ステップＳ１２０４で算出した客単価を情報処理装置１０３に送信する（ステップＳ１２０５）。これにより、情報処理装置１０３に客単価データ１１９が表示される。 The cloud server 101 transmits the customer unit price calculated in step S1204 to the information processing device 103 (step S1205). Thereby, the customer unit price data 119 is displayed on the information processing device 103 .
　＜棚割推定処理＞
　図１３は、棚割推定部１２２による棚割推定処理手順例を示すフローチャートである。クラウドサーバ１０１は、推定対象期間における推定対象区画での商品群の売上個数および客単価データ１１９を取得する（ステップＳ１３０１）。具体的には、たとえば、クラウドサーバ１０１は、情報処理装置１０３および商業施設１０２にＰＯＳ（Ｐｏｉｎｔ　Ｏｆ　Ｓａｌｅｓ）システムが導入されている場合は、情報処理装置１０３のＰＯＳデータから、推定対象期間における推定対象区画での商品群の売上個数および客単価を取得する。 <Planogram estimation process>
FIG. 13 is a flow chart showing an example of the shelf allocation estimation processing procedure by the shelf allocation estimation unit 122 . The cloud server 101 acquires the number of products sold in the estimation target section in the estimation target period and the customer unit price data 119 (step S1301). Specifically, for example, when a POS (Point Of Sales) system is introduced in the information processing apparatus 103 and the commercial facility 102, the cloud server 101 performs estimation in the estimation target period from the POS data of the information processing apparatus 103. Acquire the number of products sold and the unit price per customer in the target section.
　また、クラウドサーバ１０１は、情報処理装置１０３および商業施設１０２にＰＯＳシステムが導入されていない場合は、情報処理装置１０３で管理者が入力した推定対象期間における推定対象区画での商品群の売上個数を情報処理装置１０３から取得する。また、クラウドサーバ１０１は、客単価推定部１２３から推定対象期間における客単価データ１１９を取得する。 If the information processing device 103 and the commercial facility 102 do not have a POS system installed, the cloud server 101 calculates the number of sales of the product group in the estimation target section during the estimation target period entered by the administrator in the information processing device 103. is acquired from the information processing apparatus 103 . The cloud server 101 also acquires the customer unit price data 119 in the estimation target period from the customer unit price estimation unit 123 .
　クラウドサーバ１０１は、区画１３１における商品ごとの売上個数および客単価でグループ分けする（ステップＳ１３０２）。 The cloud server 101 groups the products in the section 131 by the number of products sold and the unit price per customer (step S1302).
　図１４は、商品群のグルーピング例を示す説明図である。図１４において、横軸を商品の客単価とし、そのしきい値をＴｈ１とする。また、縦軸を商品の売上個数とし、そのしきい値をＴｈ２とする。 FIG. 14 is an explanatory diagram showing an example of grouping product groups. In FIG. 14, the horizontal axis is the customer unit price of the product, and the threshold is Th1. The vertical axis is the number of products sold, and the threshold is Th2.
　客単価がしきい値Ｔｈ１以上で、かつ、売上個数がしきい値Ｔｈ２以上となる商品の集合をグループＧ１とする。客単価がしきい値Ｔｈ１より小さく、かつ、売上個数がしきい値Ｔｈ２以上となる商品の集合をグループＧ２とする。客単価がしきい値Ｔｈ１以上で、かつ、売上個数がしきい値Ｔｈ２より小さい商品の集合をグループＧ３とする。客単価がしきい値Ｔｈ１より小さく、かつ、売上個数がしきい値Ｔｈ２より小さい商品の集合をグループＧ４とする。 Group G1 is a set of products whose unit price per customer is equal to or greater than threshold Th1 and whose number of sales is equal to or greater than threshold Th2. A group G2 is defined as a set of products whose unit price per customer is smaller than the threshold value Th1 and whose number of sales is equal to or greater than the threshold value Th2. A group G3 is a set of products whose unit price per customer is equal to or greater than the threshold value Th1 and whose number of sales is smaller than the threshold value Th2. A group G4 is a set of products whose unit price per customer is smaller than the threshold Th1 and whose number of sales is smaller than the threshold Th2.
　図１４では、クラウドサーバ１０１は、区画１３１内の陳列対象商品群をグループＧ１～Ｇ４の４つのグループに分類したが、２以上であれば、グループ数は４つに限定されない。 In FIG. 14, the cloud server 101 classified the display target product groups in the section 131 into four groups G1 to G4, but the number of groups is not limited to four as long as there are two or more.
　図１３に戻り、クラウドサーバ１０１は、グループの商品陳列の優先順位を設定する（ステップＳ１３０３）。図１４に示した例では、クラウドサーバ１０１は、グループＧ１→グループＧ２→グループＧ３→グループＧ４の順に設定する。すなわち、売上個数がＴｈ２以上のグループＧ１、Ｇ２が最優先され、そのうち、客単価がＴｈ１のグループＧ１が優先される。グループＧ１、Ｇ２の後は、客単価がＴｈ２以上のグループＧ３がグループＧ４よりも優先される。 Returning to FIG. 13, the cloud server 101 sets the product display priority of the group (step S1303). In the example shown in FIG. 14, the cloud server 101 sets in order of group G1→group G2→group G3→group G4. That is, the groups G1 and G2 whose number of sales is Th2 or more are given the highest priority, and among them, the group G1 whose unit price per customer is Th1 is given priority. After the groups G1 and G2, the group G3 having a customer unit price of Th2 or higher is given priority over the group G4.
　なお、売れない商品の売り上げを改善したい場合は、ユーザ設定により、クラウドサーバ１０１は、グループＧ１→グループＧ４→グループＧ３→グループＧ２の順に設定する。なお、クラウドサーバ１０１は、売上個数だけで商品陳列の優先順位を設定してもよく、客単価だけで商品陳列の優先順位を設定してもよい。 If it is desired to improve the sales of unsold products, the cloud server 101 sets in the order of group G1→group G4→group G3→group G2 according to user settings. Note that the cloud server 101 may set the priority order of product display based only on the number of sales, or may set the priority order of product display based only on the unit price per customer.
　図１３に戻り、クラウドサーバ１０１は、推定対象区画について未選択棚面Ｓ＃があるか否かを判断する（ステップＳ１３０４）。未選択棚面Ｓ＃がある場合（ステップＳ１３０４：Ｙｅｓ）、クラウドサーバ１０１は、未選択棚面Ｓ＃のうち面滞在時間１１７が最長の棚面Ｓ＃を選択する（ステップＳ１３０５）。 Returning to FIG. 13, the cloud server 101 determines whether or not there is an unselected shelf surface S# for the estimation target section (step S1304). If there is an unselected shelf surface S# (step S1304: Yes), the cloud server 101 selects the shelf surface S# with the longest surface stay time 117 among the unselected shelf surface S#s (step S1305).
　クラウドサーバ１０１は、未選択グループがあるか否かを判断する（ステップＳ１３０６）。未選択グループがある場合（ステップＳ１３０６：Ｙｅｓ）、クラウドサーバ１０１は、ステップＳ１３０３で設定された優先順位が最も高い未選択グループを１つ選択する（ステップＳ１３０７）。 The cloud server 101 determines whether or not there is an unselected group (step S1306). If there are unselected groups (step S1306: Yes), the cloud server 101 selects one unselected group with the highest priority set in step S1303 (step S1307).
　クラウドサーバ１０１は、選択グループに商品が残存しているか否かを判断する（ステップＳ１３０８）。商品が残存している場合（ステップＳ１３０８：Ｙｅｓ）、クラウドサーバ１０１は、選択グループから商品を選択する（ステップＳ１３０９）。選択された商品は選択グループから削除される。選択グループに商品が残存していない場合（ステップＳ１３０８：Ｎｏ）、ステップＳ１３０６に戻る。 The cloud server 101 determines whether or not the product remains in the selected group (step S1308). If the product remains (step S1308: Yes), the cloud server 101 selects the product from the selection group (step S1309). The selected items are removed from the selection group. If no product remains in the selected group (step S1308: No), the process returns to step S1306.
　図１５は、グループに所属する商品群の一例を示す説明図である。たとえば、グループＧ１には、「おもちゃＡ」、「知育玩具Ａ」、「いたずら玩具Ａ」、「知恵の輪Ａ」が含まれる。グループ内の商品の選択順は限定されない。具体的には、たとえば、選択順は、ランダムでもよく、名前順でもよく、売上金額順でもよく、客単価順でもよい。 FIG. 15 is an explanatory diagram showing an example of product groups belonging to a group. For example, group G1 includes "toy A", "educational toy A", "mischievous toy A", and "wisdom ring A". The order in which the products in the group are selected is not limited. Specifically, for example, the order of selection may be random, name order, sales amount order, or customer unit price order.
　クラウドサーバ１０１は、選択商品が選択棚面Ｓ＃に追加可能か否かを判断する（ステップＳ１３１０）。具体的には、たとえば、商品は形状データおよび体積データを有しており、棚面Ｓ＃も棚段ごとに残存収納形状および残存容積データを有する。クラウドサーバ１０１は、選択商品の形状データおよび体積データと、選択棚面Ｓ＃の各棚段の残存収納形状および残存容積データに基づいて、いずれかの棚段に選択商品が収納可能か否かを判断する。 The cloud server 101 determines whether the selected product can be added to the selected shelf surface S# (step S1310). Specifically, for example, the product has shape data and volume data, and the shelf surface S# also has residual storage shape and residual volume data for each shelf. The cloud server 101 determines whether or not the selected product can be stored on any shelf based on the shape data and volume data of the selected product and the remaining storage shape and remaining volume data of each shelf of the selected shelf surface S#. to judge.
　選択棚面Ｓ＃のいずれの棚段にも選択商品が追加可能でない場合（ステップＳ１３１０１３１０：Ｎｏ）、ステップＳ１３０４に戻る。一方、選択棚面Ｓ＃のいずれかの棚段に選択商品が追加可能である場合、クラウドサーバ１０１は、選択商品を選択棚面Ｓ＃の棚段に選択商品を追加する。 If the selected product cannot be added to any shelf of the selected shelf surface S# (step S13101310: No), return to step S1304. On the other hand, if the selected product can be added to any shelf level of the selected shelf face S#, the cloud server 101 adds the selected product to the shelf level of the selected shelf face S#.
　図１６は、商品追加例１を示す説明図である。図１６では、棚１３２は、高さ方向に棚段Ｔ１～Ｔ７を有する。棚段Ｔ１～Ｔ７を区別しない場合は、棚段Ｔと称す。人１３４の目線や立った姿勢で棚段Ｔに手が届く範囲を、ゴールデンゾーン１６００と称す。本例では、ゴールデンゾーン１６００は、棚段Ｔ３～Ｔ５である。ゴールデンゾーン１６００内の棚段Ｔ３～Ｔ５が、選択商品の陳列先として優先的に選択され、ゴールデンゾーン１６００内の棚段Ｔ３～Ｔ５に選択商品が陳列可能な残存スペースがなくなると、ゴールデンゾーン１６００外の棚段Ｔ１、Ｔ２、Ｔ６、Ｔ７については、ゴールデンゾーン１６００から距離が近い順に選択される。 FIG. 16 is an explanatory diagram showing example 1 of product addition. In FIG. 16, shelf 132 has shelves T1-T7 in the height direction. The racks T1 to T7 are referred to as the racks T when not distinguished. A range within which the shelf T can be reached by the person 134's line of sight or in a standing posture is referred to as a golden zone 1600 . In this example, the golden zone 1600 is trays T3-T5. The shelves T3 to T5 in the golden zone 1600 are preferentially selected as the display destination of the selected product. Outer trays T1, T2, T6, and T7 are selected in descending order of distance from the golden zone 1600 .
　図１７は、商品追加例２を示す説明図である。図１７に示すように、棚段Ｔ７もゴールデンゾーン１６００に含めてもよい。どの棚段Ｔがゴールデンゾーン１６００に含まれるか、棚段Ｔの選択順をどのようにするかは、事前に管理者が情報処理装置１０３からクラウドサーバ１０１に設定可能である。 FIG. 17 is an explanatory diagram showing example 2 of product addition. Tray T7 may also be included in golden zone 1600, as shown in FIG. Which shelf T is included in the golden zone 1600 and how the shelf T is selected can be set in the cloud server 101 from the information processing device 103 in advance by the administrator.
　図１３に戻り、ステップＳ１３１１のあと、ステップＳ１３０８に戻る。ステップＳ１３０４において未選択棚面Ｓ＃がない場合（ステップＳ１３０４：Ｎｏ）、または、ステップＳ１３０６において未選択グループがない場合（ステップＳ１３０６：Ｎｏ）、クラウドサーバ１０１は、棚割結果を情報処理装置１０３に出力する。これにより、情報処理装置１０３では、図８に示した表示画面７００が表示可能になる。 Returning to FIG. 13, after step S1311, return to step S1308. If there is no unselected shelf surface S# in step S1304 (step S1304: No), or if there is no unselected group in step S1306 (step S1306: No), the cloud server 101 sends the planogram result to the information processing apparatus 103. output to As a result, the display screen 700 shown in FIG. 8 can be displayed on the information processing apparatus 103 .
　なお、図１３では、ルールベースで棚割推定処理を実行したが、機械学習により棚割推定処理を実行してもよい。具体的には、たとえば、クラウドサーバ１０１は、商品の売上個数や商品の客単価、棚面Ｓ＃ごとの面滞在時間１１７を説明変数とし、棚１３２における商品の陳列位置を目的変数として、棚割学習モデルを区画１３１ごとに生成し、記憶デバイス５０２に格納しておく。そして、クラウドサーバ１０１は、予測対象商品の売上個数、客単価、推定対象区画の各棚面Ｓ＃ごとの面滞在時間１１７入力することにより、予測対象商品の棚１３２における商品の陳列位置を出力する。このように、クラウドサーバ１０１は、商品ごとの陳列位置を棚割結果として情報処理装置１０３に出力することができる。 In FIG. 13, the planogram estimation process is performed based on rules, but the planogram estimation process may be performed by machine learning. Specifically, for example, the cloud server 101 uses the number of products sold, the unit price per customer of products, and the surface stay time 117 for each shelf surface S# as explanatory variables, and the product display position on the shelf 132 as an objective variable. A split learning model is generated for each section 131 and stored in the storage device 502 . Then, the cloud server 101 outputs the product display position on the shelf 132 of the prediction target product by inputting the number of sales of the prediction target product, the customer unit price, and the surface stay time 117 for each shelf surface S# of the estimation target section. do. In this way, the cloud server 101 can output the display position of each product to the information processing device 103 as a planogram result.
　このように、本実施例によれば、面滞在時間１１７の推定容易化を図ることができ、面滞在時間１１７を考慮した効率的な棚割を設定することができ、管理者の負担軽減を図ることができる。また、面滞在時間１１７を考慮することで、区画１３１の賃料の適正化を図ることができる。 As described above, according to the present embodiment, it is possible to facilitate estimation of the page stay time 117, and to set an efficient planogram in consideration of the page stay time 117, thereby reducing the burden on the administrator. can be planned. In addition, by considering the face stay time 117, the rent of the section 131 can be optimized.
　なお、本発明は前述した実施例に限定されるものではなく、添付した特許請求の範囲の趣旨内における様々な変形例及び同等の構成が含まれる。たとえば、前述した実施例は本発明を分かりやすく説明するために詳細に説明したものであり、必ずしも説明した全ての構成を備えるものに本発明は限定されない。また、ある実施例の構成の一部を他の実施例の構成に置き換えてもよい。また、ある実施例の構成に他の実施例の構成を加えてもよい。また、各実施例の構成の一部について、他の構成の追加、削除、または置換をしてもよい。 It should be noted that the present invention is not limited to the above-described embodiments, and includes various modifications and equivalent configurations within the scope of the attached claims. For example, the above-described embodiments have been described in detail in order to explain the present invention in an easy-to-understand manner, and the present invention is not necessarily limited to those having all the described configurations. Also, part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Moreover, the configuration of another embodiment may be added to the configuration of one embodiment. Moreover, other configurations may be added, deleted, or replaced with respect to a part of the configuration of each embodiment.
　また、前述した各構成、機能、処理部、処理手段等は、それらの一部又は全部を、たとえば集積回路で設計する等により、ハードウェアで実現してもよく、プロセッサがそれぞれの機能を実現するプログラムを解釈し実行することにより、ソフトウェアで実現してもよい。 In addition, each configuration, function, processing unit, processing means, etc. described above may be implemented in hardware, for example, by designing a part or all of them with an integrated circuit, and the processor implements each function. It may be realized by software by interpreting and executing a program to execute.
　各機能を実現するプログラム、テーブル、ファイル等の情報は、メモリ、ハードディスク、ＳＳＤ（Ｓｏｌｉｄ　Ｓｔａｔｅ　Ｄｒｉｖｅ）等の記憶装置、又は、ＩＣ（Ｉｎｔｅｇｒａｔｅｄ　Ｃｉｒｃｕｉｔ）カード、ＳＤカード、ＤＶＤ（Ｄｉｇｉｔａｌ　Ｖｅｒｓａｔｉｌｅ　Ｄｉｓｃ）の記録媒体に格納することができる。 Information such as programs, tables, files, etc. that realize each function is recorded on storage devices such as memory, hard disk, SSD (Solid State Drive), or IC (Integrated Circuit) card, SD card, DVD (Digital Versatile Disc) Can be stored on media.
　また、制御線や情報線は説明上必要と考えられるものを示しており、実装上必要な全ての制御線や情報線を示しているとは限らない。実際には、ほとんど全ての構成が相互に接続されていると考えてよい。 In addition, the control lines and information lines indicate those that are considered necessary for explanation, and do not necessarily indicate all the control lines and information lines necessary for implementation. In practice, it can be considered that almost all configurations are interconnected.

Claims (13)

  1. 　プログラムを実行するプロセッサと、前記プログラムを記憶する記憶デバイスと、を有する推定装置であって、
    　前記プロセッサは、
    　人が対面する１以上の面を有する区画が配置されている領域において、時系列な前記人の移動軌跡を示す人流方向データを取得する取得処理と、
    　前記取得処理によって取得された人流方向データと、前記区画内における前記面の位置と、に基づいて、前記面における前記人の滞在時間を算出する算出処理と、
    　前記算出処理によって算出された滞在時間を出力する出力処理と、
    　を実行することを特徴とする推定装置。     An estimating device having a processor that executes a program and a storage device that stores the program,
    The processor
    Acquisition processing for acquiring people flow direction data indicating a time-series movement trajectory of a person in an area in which a section having one or more surfaces facing a person is arranged;
    a calculation process of calculating the length of time the person stays on the surface based on the people flow direction data acquired by the acquisition process and the position of the surface in the section;
    an output process for outputting the stay time calculated by the calculation process;
    An estimation device characterized by executing
  2. 　請求項１に記載の推定装置であって、
    　前記取得処理では、前記プロセッサは、前記領域に配置された、物体の距離および移動方向を検出するセンサからの出力に基づいて、前記人流方向データを取得する、
    　ことを特徴とする推定装置。     The estimating device according to claim 1,
    In the acquisition process, the processor acquires the people flow direction data based on the output from a sensor that detects the distance and movement direction of an object and is placed in the area.
    An estimation device characterized by:
  3. 　請求項１に記載の推定装置であって、
    　前記算出処理では、前記プロセッサは、前記区画の周辺環境を示すデータに基づいて、前記滞在時間を算出する、
    　ことを特徴とする推定装置。     The estimating device according to claim 1,
    In the calculation process, the processor calculates the stay time based on data indicating the surrounding environment of the partition.
    An estimation device characterized by:
  4. 　請求項１に記載の推定装置であって、
    　前記記憶デバイスは、過去の前記人流方向データと前記面の位置とを説明変数とし、前記面における過去の滞在時間を目的変数として学習された面滞在時間推定モデルを記憶しており、
    　前記算出処理では、前記プロセッサは、前記人流方向データと前記面の位置とを前記面滞在時間推定モデルに入力することにより、前記滞在時間を算出する、
    　ことを特徴とする推定装置。     The estimating device according to claim 1,
    The storage device stores a face stay time estimation model learned using the past people flow direction data and the position of the face as explanatory variables, and using the past stay time at the face as an objective variable,
    In the calculation process, the processor calculates the staying time by inputting the people flow direction data and the position of the surface into the surface staying time estimation model.
    An estimation device characterized by:
  5. 　請求項４に記載の推定装置であって、
    　前記記憶デバイスは、過去の前記人流方向データと前記面の位置と前記区画の周辺環境を示すデータとを説明変数とし、前記面における過去の滞在時間を目的変数として学習された面滞在時間推定モデルを記憶しており、
    　前記算出処理では、前記プロセッサは、前記人流方向データと前記面の位置と前記区画の周辺環境を示すデータとを前記面滞在時間推定モデルに入力することにより、前記滞在時間を算出する、
    　ことを特徴とする推定装置。     The estimating device according to claim 4,
    The storage device is a surface staying time estimation model trained using the past people flow direction data, the position of the surface, and the data indicating the surrounding environment of the section as explanatory variables, and using the past stay time at the surface as an objective variable. is remembered and
    In the calculation process, the processor calculates the stay time by inputting the people flow direction data, the position of the surface, and the data indicating the surrounding environment of the section into the surface stay time estimation model.
    An estimation device characterized by:
  6. 　請求項１に記載の推定装置であって、
    　前記プロセッサは、
    　前記区画において前記面を構成する棚に配置される商品の売上個数と、前記面の前記滞在時間と、に基づいて、前記棚における前記商品の配置位置を推定する棚割推定処理を実行し、
    　前記出力処理では、前記プロセッサは、前記棚割推定処理によって推定された前記商品の配置位置を出力する、
    　ことを特徴とする推定装置。     The estimating device according to claim 1,
    The processor
    executing planogram estimation processing for estimating the arrangement position of the product on the shelf based on the number of sales of the product arranged on the shelf that constitutes the surface in the section and the stay time on the surface;
    In the output process, the processor outputs the arrangement position of the product estimated by the planogram estimation process.
    An estimation device characterized by:
  7. 　請求項６に記載の推定装置であって、
    　前記棚割推定処理では、前記プロセッサは、前記区画に配置される商品群について前記売上個数に基づいて優先順位を設定し、前記優先順位にしたがって、前記棚における前記商品の配置位置を推定する、
    　ことを特徴とする推定装置。     The estimating device according to claim 6,
    In the planogram estimation process, the processor sets a priority order based on the number of sales for a group of products arranged in the section, and estimates the arrangement position of the product on the shelf according to the priority order.
    An estimation device characterized by:
  8. 　請求項１に記載の推定装置であって、
    　前記プロセッサは、
    　前記区画において前記面を有する棚に配置される商品の客単価と、前記面の前記滞在時間と、に基づいて、前記棚における前記商品の配置位置を推定する棚割推定処理を実行し、
    　前記出力処理では、前記プロセッサは、前記棚割推定処理によって推定された前記商品の配置位置を出力する、
    　ことを特徴とする推定装置。     The estimating device according to claim 1,
    The processor
    executing planogram estimation processing for estimating the arrangement position of the product on the shelf based on the unit price per customer of the product arranged on the shelf having the face in the section and the stay time on the face,
    In the output process, the processor outputs the arrangement position of the product estimated by the planogram estimation process.
    An estimation device characterized by:
  9. 　請求項８に記載の推定装置であって、
    　前記棚割推定処理では、前記プロセッサは、前記区画に配置される商品群について前記客単価に基づいて優先順位を設定し、前記優先順位にしたがって、前記棚における前記商品の配置位置を推定する、
    　ことを特徴とする推定装置。     The estimating device according to claim 8,
    In the planogram estimation process, the processor sets a priority order based on the customer unit price for a group of products arranged in the section, and estimates the arrangement position of the product on the shelf according to the priority order.
    An estimation device characterized by:
  10. 　請求項８に記載の推定装置であって、
    　前記取得処理では、前記プロセッサは、前記人が前記商品の配置位置に腕を伸ばした回数を取得し、
    　前記プロセッサは、
    　前記腕を伸ばした回数と、前記商品の売上個数と、に基づいて、前記商品の前記客単価を推定する客単価推定処理を実行し、
    　前記棚割推定処理では、前記商品の売上個数と、前記客単価推定処理によって推定された前記商品の客単価と、に基づいて、前記棚における前記商品の配置位置を推定する、
    　ことを特徴とする推定装置。     The estimating device according to claim 8,
    In the obtaining process, the processor obtains the number of times the person has extended his/her arm to the product placement position,
    The processor
    executing a customer unit price estimation process for estimating the customer unit price of the product based on the number of times the arm is stretched and the number of sales of the product;
    In the planogram estimation process, the arrangement position of the product on the shelf is estimated based on the number of sales of the product and the customer unit price of the product estimated by the customer unit price estimation process.
    An estimation device characterized by:
  11. 　請求項１に記載の推定装置であって、
    　前記プロセッサは、
    　前記区画において前記面を有する棚に配置される商品の売上個数および客単価と、前記面の前記滞在時間と、に基づいて、前記棚における前記商品の配置位置を推定する棚割推定処理を実行し、
    　前記出力処理では、前記プロセッサは、前記棚割推定処理によって推定された前記商品の配置位置を出力する、
    　ことを特徴とする推定装置。     The estimating device according to claim 1,
    The processor
    Execution of planogram estimation processing for estimating the arrangement position of the product on the shelf based on the number of products sold and the unit price per customer placed on the shelf having the face in the section and the stay time on the face death,
    In the output process, the processor outputs the arrangement position of the product estimated by the planogram estimation process.
    An estimation device characterized by:
  12. 　請求項６に記載の推定装置であって、
    　前記棚割推定処理では、前記プロセッサは、前記棚の高さ方向の位置に基づいて、前記棚における前記商品の配置位置を推定する、
    　ことを特徴とする推定装置。     The estimating device according to claim 6,
    In the planogram estimation process, the processor estimates the arrangement position of the product on the shelf based on the position of the shelf in the height direction.
    An estimation device characterized by:
  13. 　プログラムを実行するプロセッサと、前記プログラムを記憶する記憶デバイスと、を有する推定装置による推定方法であって、
    　前記プロセッサは、
    　１以上の面を有する区画が配置されている領域において、時系列な人の移動軌跡を示す人流方向データを取得する取得処理と、
    　前記取得処理によって取得された人流方向データと、前記区画内における前記面の位置と、に基づいて、前記面における前記人の滞在時間を算出する算出処理と、
    　前記算出処理によって算出された滞在時間を出力する出力処理と、
    　を実行することを特徴とする推定方法。     An estimation method by an estimation device having a processor that executes a program and a storage device that stores the program,
    The processor
    Acquisition processing for acquiring people flow direction data indicating a time-series movement trajectory of people in an area in which a section having one or more surfaces is arranged;
    a calculation process of calculating the length of time the person stays on the surface based on the people flow direction data acquired by the acquisition process and the position of the surface in the section;
    an output process for outputting the stay time calculated by the calculation process;
    An estimation method characterized by performing
PCT/JP2022/033035 2022-02-03 2022-09-01 Estimation apparatus and estimation method WO2023149005A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2022015609A JP2023113330A (en) 2022-02-03 2022-02-03 Estimation device and estimation method
JP2022-015609 2022-02-03

Publications (1)

Publication Number Publication Date
WO2023149005A1 true WO2023149005A1 (en) 2023-08-10

Family

ID=87553277

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2022/033035 WO2023149005A1 (en) 2022-02-03 2022-09-01 Estimation apparatus and estimation method

Country Status (2)

Country Link
JP (1) JP2023113330A (en)
WO (1) WO2023149005A1 (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160335586A1 (en) * 2015-05-12 2016-11-17 Oracle International Corporation Display space optimization
JP2016218821A (en) * 2015-05-22 2016-12-22 由紀貞 深谷 Marketing information use device, marketing information use method and program

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160335586A1 (en) * 2015-05-12 2016-11-17 Oracle International Corporation Display space optimization
JP2016218821A (en) * 2015-05-22 2016-12-22 由紀貞 深谷 Marketing information use device, marketing information use method and program

Also Published As

Publication number Publication date
JP2023113330A (en) 2023-08-16

Similar Documents

Publication Publication Date Title
US10163115B2 (en) Control method for displaying merchandising information on information terminal
JP5635709B2 (en) Display of data about physical retail environment on virtual illustration of physical retail environment
US10347021B1 (en) Visualization of multiple objects on a dashboard
JP6825628B2 (en) Flow line output device, flow line output method and program
US20160005052A1 (en) Information processing system and information processing method
WO2016174878A1 (en) Action analysis system and action analysis method
CN113222705A (en) Commodity display optimization method, device and equipment
US20150118663A1 (en) System and Method for Predicting and Communicating Data Relating to a Partner&#39;s Mood and Sensitivity
US11899833B2 (en) Systems and methods for interacting with augmented reality content using a dual-interface
WO2023149005A1 (en) Estimation apparatus and estimation method
WO2002091266A1 (en) Image providing device
JP2018185683A (en) Display control program, display control method, and information processing apparatus
JP2010097398A (en) Device and method for displaying merchandise information
JP2002230127A (en) Planogram create method and its device
JP2019164635A (en) Information processing apparatus, information processing method, and program
US20230142288A1 (en) Information processing device, information processing method, and recording medium
WO2019187968A1 (en) Information processing system, information processing method, and storage medium
US20150347946A1 (en) Store opening cockpit for real-time preparation, simulation and execution of merchandise supply for grand opening events
JP5717122B2 (en) Interior material purchase cost estimation system and interior material purchase cost estimation program
WO2020054203A1 (en) Display device, method, and program
CN107015733B (en) Option information presentation system and method
JP6182171B2 (en) Demand prediction apparatus and program
JP2014074996A (en) Information processing system, information processing method, and program
JP6449578B2 (en) Purchase forecast analysis system and program thereof
JP7396736B1 (en) Cafeteria management system, cafeteria management method, and program

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22924910

Country of ref document: EP

Kind code of ref document: A1