Income specification method and income specification program
By analyzing movement trajectory data to identify stay locations and adjust income estimates based on land prices and stay times, the method accurately determines an individual's income, improving SDT policy search accuracy.
Patent Information
- Application Number
- JP2023083401
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-19
- Publication Date
- 2024-11-29
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing methods fail to accurately estimate an individual's income based on movement trajectory data, relying on average customer unit prices of stores, which leads to inaccuracies in income estimation.
The method involves acquiring movement trajectory data, identifying stay locations where an individual stays for a predetermined time, determining the land price and stay time at these locations, and using an income estimation model to calculate income, adjusting the estimate based on the duration of stay and land prices.
This approach allows for accurate income identification without requiring external attribute information, enhancing the precision of income estimation and improving the accuracy of policy searches in Digital Twin (SDT) systems.
Smart Images

Figure 2024166956000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to an income determination method and an income determination program. [Background technology]
[0002] Social Digital Twin (SDT) is a technology that digitally reproduces human behavior in an attempt to solve problems and explore policies. In order to reproduce people's real-world behavior in a virtual space using SDT, people's movement trajectories and attribute information, such as age, gender, and income, are required. Attribute information is an important factor in determining people's reactions to policies.
[0003] Prior art includes, for example, a system that obtains the current income and expenditure of an individual who is active in a specified area from a bank account or the like to estimate the future income. Another system estimates the user's home location from early morning or nighttime location information, estimates the place of employment from daytime location information, estimates the standard of living from the land price of the home, estimates the annual income from the place of employment using company list information, and delivers information content such as advertisements based on the estimation results. Another system obtains the user's location information and information on the places of movement, and the information on the places of movement stores information such as the gender, age, and annual income based on the average customer price of the target user, and estimates the user's annual income based on the information on the places of movement to which the user has moved (for example, see Patent Documents 1 to 3 below). [Prior art documents] [Patent documents]
[0004] [Patent Document 1] JP 2018-101324 A [Patent Document 2] JP 2010-237305 A [Patent Document 3] JP 2018-180996 A Summary of the Invention [Problem to be solved by the invention]
[0005] However, the conventional technology could not accurately estimate the income of the subject. In Patent Document 3, the income of the subject is estimated only based on the average customer spending at the store where the subject moves, and the income of each subject cannot be accurately estimated.
[0006] In one aspect, the present invention aims to be able to accurately identify a subject's income. [Means for solving the problem]
[0007] According to one embodiment, an income determination method and an income determination program are proposed, which acquire movement trajectory data of a subject, and determine the income of the subject based on a place where the subject stayed for a predetermined time or more in the movement trajectory data and the subject's stay time at the place. Effect of the Invention
[0008] According to one embodiment, it becomes possible to precisely determine a subject's income. [Brief description of the drawings]
[0009] [Figure 1] FIG. 1 is an explanatory diagram illustrating an example of an income specification method according to an embodiment. [Diagram 2] FIG. 2 is an explanatory diagram of the use of attribute information in policy exploration using SDT. [Diagram 3] Figure 3 is an explanatory diagram of the policy search task. [Figure 4] FIG. 4 is a block diagram of an example of functions of the income identifying device according to the embodiment. [Diagram 5] FIG. 5 is a block diagram illustrating an example of a hardware configuration of the income identifying device. [Figure 6] FIG. 6 is a table showing an example of the stored contents of the movement trajectory data. [Figure 7] FIG. 7 is a table showing an example of the stored contents of place-purpose data. [Figure 8]FIG. 8 is a table showing an example of the stored contents of the stay location / time data. [Figure 9] FIG. 9 is a diagram showing an example of estimating the home position based on the stay location and time data. [Figure 10] FIG. 10 is a chart showing various data used to create the income estimation model. [Figure 11] FIG. 11 is an explanatory diagram of an example of creating an income estimation model. [Figure 12] FIG. 12 is a table showing examples of various data used in the income specification process. [Figure 13] FIG. 13 is a flowchart illustrating an example of a process performed by the income identification device according to the embodiment. [Figure 14] FIG. 14 is a flowchart illustrating an example of a process performed by the income identification device according to the embodiment. [Figure 15] FIG. 15 is a block diagram showing an example of service application based on the income determination result. [Figure 16] FIG. 16 is an explanatory diagram of an example of information content delivery corresponding to the income level of a target person. [Figure 17] FIG. 17 is an explanatory diagram showing an example of an income specifying method according to another embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0010] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an embodiment of an income determination method and an income determination program according to the present invention will be described in detail with reference to the drawings.
[0011] (One Example of an Income Identification Method According to an Embodiment) FIG. 1 is an explanatory diagram showing an example of an income specification method according to an embodiment. For example, the income specification method of the embodiment specifies a place of stay such as a commercial facility where the subject stayed for a predetermined time or more based on the movement trajectory data of the subject whose income is to be estimated, acquires the land price of the place of stay, and calculates the income based on the acquired land price. Then, a weight is set based on the stay time at the place of stay, and the calculated income is adjusted using the weight, thereby estimating the income of the subject more accurately. This allows the income to be specified even when attribute information on the income of the subject cannot be obtained. Furthermore, for example, the attribute information including the specified income is used to enable the distribution of information, such as information content, adapted to the subject according to income.
[0012] The income identification method of the embodiment can be realized by an income identification device 100 such as a server executing a program. The income identification device 100 can also be used as one function (income identification) of the policy search performed by the SDT. In this case, the server has, for example, the function of the income identification device 100 and the function of the SDT. Furthermore, the function of the server can also be realized by a cloud on a network.
[0013] In the income specification method of the embodiment, a process for specifying the income of a subject is carried out, including the following processes (1) to (4). (1) Acquisition of subject movement trajectory data (2) Identifying the subject's place of stay (3) Obtain the land price of the place of stay and the duration of stay, and estimate the subject's income based on the land price. (4) Adjustment of the estimated income of the subjects based on weights corresponding to the time spent at the place of stay
[0014] An example of the process performed by the income specifying device 100 will be described with reference to FIG. (1) The movement trajectory data of the target person U can be, for example, location information output by a global positioning system (GPS) of a smartphone or the like carried by the target person U. The income specifying device 100 acquires movement trajectory data m, which is a change in the location information of the target person U over time, from, for example, a management server or the like that manages the location information and map information of the target person U.
[0015] (2) The target person's place of stay is specified, for example, by specifying the place of stay where the target person stayed for a predetermined time or more from the movement trajectory data m. For example, the income identifying device 100 identifies the place of stay such as the facility where the target person U stayed based on the movement trajectory data m on the map information E managed by the management server or the like. For example, when the target person U purchases goods or the like at the commercial facilities P01, P02, he / she stays for a predetermined time within the range of the commercial facilities P01, P02 on the corresponding map information E. The income identifying device 100 can specify the place where the target person stayed for the predetermined time as the place of stay (commercial facility) P01, P02 based on the movement trajectory data m.
[0016] (3) Acquire the land price and duration of stay of the place of stay, and estimate the income of the subject based on the land price, among which the income identifying device 100 can acquire the land price by referring to a database of land price data related to the place of stay (commercial facility) identified on the map information E. In addition, the income identifying device 100 can calculate the duration of stay based on the time that the subject U stayed within the range of the places of stay (commercial facility) P01, P02.
[0017] In addition, regarding the estimation of the income of the subject based on land prices, for example, the income identification device 100 estimates the income of the subject U using each data of the land price data, the identified stay places (commercial facilities) P01, P02 of the subject U, and the stay time.
[0018] The income identification device 100 may identify the income of the subject by using a previously created income estimation model. The income estimation model is a model to be referred to in order to estimate the income of the subject U, and for example, the income identification device 100 creates the model in advance as a relational expression combining various information related to the income of a plurality of individuals. The various information includes, for example, the annual income of each individual, the home position and the land price of the home, the place of stay based on the movement trajectory data m, the stay time, the land price of the place of stay (facility), and the like.
[0019] Then, the income identification device 100 estimates the income of the subject U by inputting various information about the subject U (place of stay, duration of stay, land price of the home, etc.) indicated by the above-mentioned movement trajectory data m to the income estimation model.
[0020] In this way, the income identification device 100 of the embodiment calculates the individual's income based on the movement trajectory data m of the subject U, using the "land price of the home" and the "land price and stay time of the place of stay (commercial facility)" which are considered to be positively correlated with income.
[0021] Here, we will explain how various information is used in income determination by the income determination device 100. Regarding the land price of a home, it is assumed that a high land price of a home = a high income. Regarding the land price and stay time of a commercial facility, there is a positive correlation between prices and land prices. Also, it is known from research reports that there is a positive correlation between the Consumer Price Regional Difference Index and land prices (see Technical Document 1 below). There is also a positive correlation between purchase price and stay time. Although there seems to be no correlation between stay time and sales price, it has been reported that there is actually a large causal relationship between each of them (see Technical Document 2 below). Also, it has been reported that what influences purchases is the length of stay time more than the subject U's desire to buy. (See Technical Document 3 below).
[0022] Technical Document 1: Takayuki Yamashita, Research on Mass Retailers and Retail Price Fluctuations, Shizuoka University Academic Repository, March 1998 Technical Document 2: What kind of restaurant do you want to stay at for a long time?, Nikkei TECH, 2007 / 5 / 28, https: / / xtech.nikkei.com / it / article / COLUMN / 20070523 / 271995 / Technical Document 3: Considering pedestrian movement behavior in station buildings, EKISUMER Vol. 37, Station Consumption Research Center, 2018 / 7 / 18, http: / / www.jeki.co.jp / ekishoken / upload / docs / EKISUMER_vol_37.pdf
[0023] From the above, the income identifying device 100 estimates the income of the subject U based on the premise that the subject U stays for a long time in a place of stay (commercial facility) with high land prices = high purchase prices = high income.
[0024] (4) Regarding the adjustment of the estimated income based on the weight corresponding to the stay time at the place of stay, for example, the income identification device 100 sets a weight w that has a larger influence on the estimated value of income as the stay time of the subject U at the places of stay (commercial facilities) P01 and P02 is longer. By using this weight w, income is identified more accurately. Since the probability of purchasing products, etc. increases the longer the stay time at the place of stay, it is identified that the income is higher as the stay time is longer. By adjusting the estimated income according to the length of the stay time, it becomes possible to identify the income accurately.
[0025] For example, the income identifying device 100 can set a weight w in advance in the income estimation model. In this case, the income identifying device 100 obtains a weight w for income for each of a plurality of individuals who are samples for creating the income estimation model. Then, the income identifying device 100 creates an income estimation model consisting of a predetermined relational expression including the calculated weight w.
[0026] An example of income identification using weights will be described with reference to FIG. 1. Based on the movement trajectory data m of the target person U, the income identification device 100 identifies two commercial facilities P01 and P02 as the places of stay of the target person U. In this case, the income identification device 100 identifies the land price "A yen" and the stay time "T A", commercial facility P02's land price "B million yen" and stay time "T B Based on this, the income of subject U is determined.
[0027] The income identification device 100 determines the dwell time: T A >T B Based on the weight: w A >w B The commercial facility P01 has weight w A × A million yen, commercial facility P02 is weight w B After adjusting the weights by ×B million, A × A million yen + weight w B × B million yen is identified as the income of subject U and output. The detailed process of identifying income using the income estimation model and weighting of accommodation facilities including home land prices will be described later.
[0028] In this way, according to the income specification device of the embodiment, when there are multiple places of stay (commercial facilities), the income of the subject is estimated based on the land prices of each of the multiple places of stay. Furthermore, the income specification device can specify the income of the subject more accurately by adjusting the weighting so that the longer the stay time is, the higher the weighting is.
[0029] (Issues with existing technology) Figure 2 is an explanatory diagram of the use of attribute information in SDT policy exploration. SDT requires the movement trajectory and attribute information (age, gender, income, etc.) of subjects. Attribute information is an important factor that determines the reaction of people (subjects) to policies when exploring proposed policies.
[0030] For example, as shown in Figure 2, an example of searching for a policy to charge fees on congested roads in order to alleviate traffic congestion will be described. SDT searches for the reaction of subjects A and B to the policy (charging) based on the value of the attribute information "income." When a charging area X is set up in a specific congested area on the map shown in Figure 2 to alleviate congestion, SDT estimates the reaction of how the subjects will behave (movement trajectory). Charging area X is an area in which a fee is charged to subjects who pass through this charging area X.
[0031] Here, suppose that the attribute information "income" of subject A is 200,000 yen, and the attribute information "income" of subject B is 800,000 yen. The dotted line in the figure shows the route taken by the subject before the implementation of the policy, and the solid line in the figure shows the route taken by the subject after the implementation of the policy.
[0032] Subject A, who has a low income, will respond by taking a route that avoids paid area X after the policy is implemented. On the other hand, subject B, who has a high income, will not avoid paid area X even after the policy is implemented. In this way, SDT uses the subjects' movement trajectories and "income" from among the attribute information to explore the subjects' reactions before and after the policy.
[0033] However, compared to movement trajectory data, which can be obtained relatively easily using GPS and other devices, attribute information on subjects is difficult to obtain. In particular, few companies provide "income" among the various attribute information, due to the need to protect personal information. In addition, movement trajectory data with attributes attached is expensive. In this way, if the "income" of a subject cannot be obtained, it becomes difficult to explore measures as shown in Figure 2.
[0034] Next, we will explain the relationship between attribute information and movement trajectory data in the existing technology SDT. We will explain using an example of SDT to search for policies for a synthetic population in an area S' on a movement trajectory based on the actual population of that area. SDT assigns the "income" attribute information obtained from the synthetic population data to the movement trajectory data of the target person. SDT obtains the number of residents and attribute information in an area S from the resident statistical information of that area. The synthetic population data is population data of a virtual population group in area S' on a movement trajectory created from the resident statistical information.
[0035] Then, SDT creates synthetic population data so that the distribution of attribute information in region S' is identical to the distribution of the actual population in region S. For example, SDT randomly assigns the distribution of the attribute information "income" in the synthetic population data in region S' to the movement trajectories so that it is identical to the distribution of the actual population in region S.
[0036] Figure 3 is an explanatory diagram of the problem of policy search. As mentioned above, when attribute information is randomly assigned to movement trajectory data, a value of attribute information "income" that does not match the subject's actual income may be assigned. When attribute information "income" that does not match the movement trajectory data is assigned, the subject responds to the policy differently from reality, which reduces the accuracy of policy search.
[0037] For example, suppose that attribute information is randomly assigned to movement trajectory data in the policy search for charging tolls on congested roads described with reference to Figure 2. In this case, as shown in Figure 3, subject B does not avoid the charging area X shown by the solid line in reality, but in the policy search using SDT, he or she will react by avoiding the charging area. In this case, the effect of the policy will not be measured correctly, and the accuracy of the policy search will be reduced.
[0038] In contrast, as described above, the income identifying device 100 of the embodiment calculates the individual's income based on the movement trajectory data of the subject, using the "land price and duration of stay at the place of stay (commercial facility)" which are considered to have a positive correlation with income. Furthermore, by adding the "land price of the home", the accuracy of the calculated income value can be further improved. Thus, according to the income identifying device 100 of the embodiment, it is not necessary to obtain the attribute information "income" from outside, and the income value can be calculated with high accuracy.
[0039] In addition, in SDT, it is possible to assign attribute information "income" that matches the subject's own income. By assigning matching attribute information "income", subjects will react more realistically to policies, improving the accuracy of policy exploration in SDT. For example, in the policy exploration mentioned above (charging fees for congested roads), it becomes possible to obtain more accurate responses from each subject according to their income, as shown in Figure 2.
[0040] (Example of the configuration of the income identification device) 4 is a block diagram showing an example of functions of the income identification device according to the embodiment. The income identification device 100 includes a storage unit 110 such as a plurality of databases (DBs), an income estimation model creation unit 121, a stay location and time data creation unit 122, a home position estimation unit 123, an income identification unit 124, and a communication unit 125.
[0041] The storage unit 110 stores a plurality of data for identifying the income of a subject. The storage unit 110 includes and stores, for example, true value data 111 of income for each subject, land price data 112 for each location such as a home or place of stay, and movement trajectory data 113 which is the movement trajectory of the subject. The storage unit 110 also includes and stores location-use data 114 which is the use for each location such as a home or place of stay, and stay location-time data 115 which is the stay time for each stay location of the subject. The storage unit 110 also stores an income estimation model 116 created by the income estimation model creation unit 121.
[0042] Based on the movement trajectory data 113 and the location-purpose data 114, the stay location / time data creation unit 122 creates stay location / time data 115 which indicates the locations where the subject stayed and the stay time at each location.
[0043] The home position estimation unit 123 estimates the home position based on the stay location / time data 115. The income estimation model creation unit 121 creates an income estimation model 116 based on the true income data 111, the land price data 112, and the stay location / time data 115.
[0044] The income determination unit 124 determines the income of the subject based on the created income estimation model 116, the subject's stay location and time data, and the land price data 112 of the stay location.
[0045] The income identification results of the subject to be estimated identified by the income identification unit 124 can be output via the communication unit 125 to a device that implements various applications that use the income identification results, such as an SDT or an information content distribution device.
[0046] The income identifying device 100 is communicatively connected to the subject's terminal 150 via the communication unit 125. The terminal 150 includes a location information acquiring unit 151 that acquires the current location (latitude, longitude, etc.) of the terminal 150 by GPS or the like, and a communication unit 152 that communicatively connects to the income identifying device 100. The income identifying device 100 outputs the current location of the terminal 150 acquired via the communication unit 125 to the stay location / time data creating unit 122.
[0047] (Example of hardware configuration of income identification device) Fig. 5 is a block diagram showing a hardware configuration example of the income identification device. In Fig. 5, the income identification device 100 can use a general-purpose server. The income identification device 100 includes a CPU (Central Processing Unit) 501 such as a processor, a memory 502, a network IF 503, a recording medium IF 504, and a recording medium 505. In addition, each component is connected to each other by a bus 500.
[0048] Here, the CPU 501 functions as a control unit that controls the entire revenue specifying device 100. The CPU 501 may have multiple cores. The memory 502 has, for example, a ROM (Read Only Memory), a RAM (Random Access Memory), and a flash ROM. Specifically, for example, the flash ROM stores an OS program, the ROM stores an application program, and the RAM is used as a work area for the CPU 501. The program stored in the memory 502 is loaded into the CPU 501 to cause the CPU 501 to execute the coded process.
[0049] The network IF 503 is connected to the network NW through a communication line, and is connected to an external computer (for example, the terminal 150 shown in FIG. 4) through the network NW. The network IF 503 manages the interface between the network NW and the inside of the device, and controls input and output of data from the external computer. For example, a modem or a LAN adapter can be used as the network IF 503.
[0050] The recording medium IF 504 controls reading / writing of data from / to the recording medium 505 under the control of the CPU 501. The recording medium 505 stores the written data. The recording medium 505 may be a magnetic disk such as an HDD, an optical disk, an SSD (Solid State Drive), a USB (Universal Serial Bus) memory, or the like.
[0051] In addition to the above-mentioned components, the income specifying device 100 may have, for example, an input device, a display, and the like.
[0052] The functions of the income estimation model creation unit 121, the stay location / time data creation unit 122, the home position estimation unit 123, and the income identification unit 124 shown in Fig. 4 can be realized by executing a program on the CPU 501 shown in Fig. 5. The data in the storage unit 110 shown in Fig. 4 can be stored using the memory 502 and the recording medium 504 shown in Fig. 5. The function of the communication unit 125 shown in Fig. 4 can be realized using the network IF 503 shown in Fig. 5.
[0053] The terminal 150 shown in Fig. 4 can also be functionally realized using hardware similar to that shown in Fig. 5. The location information acquisition unit 151 shown in Fig. 4 can be functionally realized by a GPS sensor. The terminal 150 can be, for example, a smartphone carried by the subject, and in this case, the terminal 150 has a GPS function (location information acquisition unit 151) as well as functions such as a touch display and a speaker.
[0054] (Contents of data stored in the income identification device) Next, the contents of various data stored in the income specifying device will be described along with examples of their creation.
[0055] (Storage contents of movement trajectory data) 6 is a table showing an example of the stored contents of movement trajectory data. The movement trajectory data 113 is created based on the location information of the subject of the terminal 150 acquired by the stay location / time data creation unit 122. The movement trajectory data 113 includes items such as the personal ID, time, latitude, and longitude of each subject.
[0056] In the example of Figure 6, the record of the personal ID "I01" has the time "t I01 ", latitude "lat I01 ", longitude "lon I01 " is stored.
[0057] (Location-use data storage contents) 7 is a table showing an example of the stored contents of the location-purpose data. The location-purpose data 114 is information showing facilities (corresponding to places of stay) such as commercial facilities on a map and the purpose of the facility, and can use, for example, map information created in advance. The location-purpose data 114 includes items such as location ID, shape of the facility, and purpose.
[0058] In the example in Figure 7, the record for location ID "P01" contains the latitude and longitude of the vertices of the facility's polygon, "(lat P011 ,lon P011 ),(lat P012 ,lon P012 ),(lat P013 ,lon P013 ), ..." and the use "commercial facility" is recorded. Also, the use of the record with location ID "P03" is "residential."
[0059] (Memory contents of location and time data) 8 is a chart showing an example of the stored contents of the stay place / time data. The stay place / time data creation unit 122 creates the stay place / time data 115 based on the movement trajectory data 113 and the place-purpose data 114. The stay place / time data 115 includes each item of an individual ID for each subject, a stay place ID (place ID in FIG. 7), and a stay time at the stay place.
[0060] The stay place / time data creation unit 122 determines a stay place where the stay time is equal to or longer than a predetermined threshold among facilities in the place-purpose data 114 corresponding to a position indicated by the movement trajectory of the subject indicated by the movement trajectory data 113, and creates the stay place / time data 115. At this time, the stay place / time data creation unit 122 determines that the subject is staying in the same place when the movement trajectory of the subject is moving within the shape of a certain place ID (the shape of a polygon shown in FIG. 7).
[0061] In the example of FIG. 8, the record of the individual ID "I01" contains the location of stay "P01 (commercial facility)" and the duration of stay (start to end) "(t I01P011 ~t I01P012 )" is recorded. In addition, the record for personal ID "I03" includes the location of stay "P03 (residence)" and the duration of stay (start to end) "(t I03P031 ~t I03P032 )" has been recorded.
[0062] (Estimating home location using location and time data) Fig. 9 is a chart showing an example of estimating a home position based on stay place / time data. The stay place / time data 115 shown in Fig. 9 has the same content as Fig. 8, and only the stay time item is recorded in actual time. The home position estimation unit 123 estimates the home position of the subject based on the stay place / time data 115.
[0063] The home position estimation unit 123 estimates, for example, from the stay location / time data 115, the location where the subject stays the longest during a time period during which the subject is assumed to be staying at home, for example, from late night to early morning (e.g., 11:00 p.m. to 6:00 a.m.), as the subject's home.
[0064] In the example of Fig. 9, the stay time of the individual ID "I01" at the stay location "P03" is "21:00 to 07:30". The home position estimation unit 123 estimates the home of the subject with the individual ID "I01" as the stay location "P03" and stores and holds this stay location as the home position data of the corresponding subject. In addition, if the true value of the home location is obtained by a questionnaire or the like given to the subject, the home position estimation unit 123 uses this true value as the stay location.
[0065] (Examples of various data for creating income estimation models) 10 is a chart showing various data used to create an income estimation model. The following describes various data samples collected by the income estimation model creation unit 121 for creating an income estimation model. The income estimation model is created by collecting data on multiple individuals as sample subjects whose true income values are known.
[0066] In the embodiment, the subject from which samples are collected to create an income estimation model is referred to as an "individual," and the subject whose income is estimated using the created income estimation model is referred to as a "subject."
[0067] Fig. 10(a) shows home position data 1001 estimated by the home position estimation unit 123. The home position data 1001 includes a location ID indicating a home location for each personal ID. Fig. 10(b) shows true income data 111. The true income data 111 includes income for each personal ID. The income estimation model creation unit 121 collects income values (data) of multiple individuals as samples.
[0068] FIG. 10(c) shows the land price data 112. The land price data 112 includes land prices by location ID. The land price data 112 may use information such as published land price official prices. FIG. 10(d) shows the stay location and time data 115 described above.
[0069] (Example of creating an income estimation model) FIG. 11 is an explanatory diagram of an example of creating an income estimation model. The income estimation model creation unit 121 substitutes the home position data 1001, the true income data 111, the land price data 112, and the stay place and time data 115 shown in FIG. 10(a) to (d) into the relational expression in FIG. 11(a). As a result, the weights w c ,The weight of my home w l The value of is calculated to create an income estimation model.
[0070] The relational expression in FIG. 11(a) specifically describes the data substitution state for the individual ID "I01" shown in FIG. 10. The home location of this individual "I01" is the location ID "P07" indicated by the home location data 1001. The land price of this home location is "1 million" indicated by the land price data 112. The income of the individual "I01" is "500,000" indicated by the true value data 111. The places of stay are two locations indicated by the stay location / time data 115 (stay location ID "P01", stay time "12:00-17:00" and stay location ID "P02", stay time "17:30-18:00"). The land prices of the two places of stay are "1.5 million" and "800,000" indicated by the land price data 112.
[0071] In addition, in the relational equation, (300 minutes / 330 minutes) and (30 minutes / 330 minutes) are terms that indicate the ratio of the time spent at the relevant location to the total time spent outside the home by the individual with personal ID "I01", which is 330 minutes (300 minutes spent in P01 + 30 minutes spent in P02).
[0072] Then, as shown in FIG. 11(b), the income estimation model creation unit 121 substitutes each of the sample individuals (personal IDs: I01, I02, I03) into the relational equation shown in FIG. 11(a) for each of the sample individuals (for example, more than 100 individuals). Then, the income estimation model creation unit 121 uses a general-purpose least squares method to find the relational equations for the individuals, w c ,w l The least squares method calculates the slope a and the tangent point b of the linear function y=ax+b that minimizes the error with the known multiple x, y values. In the data example shown in Figure 11, c =0.1305, w l =0.3169.
[0073] (Example of income determination using income estimation model) The income determination unit 124 determines the income (income value I) of the subject using the income estimation model 116 shown in the following formula (1).
[0074]
number
[0075] Fig. 12 is a chart showing examples of various data used in the income identification process. The income identification unit 124 assigns the stay place and time data 115 shown in Fig. 12(a) of the subject (individual ID "I04") and the data of the land price data 112 by stay place shown in Fig. 12(b) to the income estimation model 116 to identify the income. It is assumed that the home position estimation unit 123 estimates the home position of the subject (individual ID "I04") shown in Fig. 12(a) as the stay place ID "P13".
[0076] For example, the income determination unit 124 determines the income value of the subject with the personal ID "I04" as follows: c =0.1305, w l = 0.3169, 0.1305 × {(120 minutes / 540 minutes) × 1.5 million) + (420 minutes / 540 minutes) × 8 million)} + 0.3169 × 4 million ≒ 252,000
[0077] (Example of processing of income specific device) 13 and 14 are flowcharts showing an example of processing of the income identifying device according to the embodiment. Fig. 13 shows an example of processing for creating an income estimation model performed in advance by the income identifying device 100, and the control unit (CPU 501) of the income identifying device 100 executes the following processing by a program.
[0078] First, the income identification device 100 (staying place / time data generating unit 122) generates the staying place / time data 115 based on the movement trajectory data of a sample individual whose true income value is known (step S1301).
[0079] Next, the income specification device 100 (home position estimation unit 123) estimates the position of the individual's home based on the movement trajectory data 113 (step S1302). The process of step S1302 is executed when the position of the home is unknown, and when the position of the individual's home can be acquired, the process can be omitted.
[0080] Next, the income identifying device 100 (income estimation model creating unit 121) acquires the individual's staying place and home land price data 112 based on the staying place / time data 115 (step S1303).
[0081] Then, the income identification device 100 (income estimation model creation unit 121) creates the income estimation model 116 based on the true income data 111, the stay place and time data 115, and the land price data 112 (step S1304), and ends the above process.
[0082] 14 is an example of income specification processing of a subject to be estimated using an income estimation model performed by the income specification device 100. First, the income specification device 100 (stay place / time data creation unit 122) creates the stay place / time data 115 based on the movement trajectory data of the subject whose income is to be estimated (step S1401).
[0083] Next, the income specification device 100 (home position estimation unit 123) estimates the position of the target person's home based on the movement trajectory data 113 (step S1402). The process of step S1402 is executed when the home position is unknown, and can be omitted when the home position of the individual can be acquired.
[0084] Next, the income identifying device 100 (income estimation model creating unit 121) acquires the staying place and home land price data 112 of the subject based on the staying place and time data 115 (step S1403).
[0085] Then, the income identifying device 100 (income identifying unit 124) identifies the income of the subject by inputting the staying place / time data 115 of the subject to the income estimation model 116 (step S1404), and ends the above processing.
[0086] (Examples of service application based on income determination results) FIG. 15 is a block diagram showing an example of a service application of the income identification result. The income identification result of the subject by the above-mentioned income identification device 100 can be applied to various services. The example of FIG. 15 is a configuration example in which the income identification result is output to an information content delivery unit 1500 to deliver information content according to the income of the subject. The function of the information content delivery unit 1500 may be a part of the function of the income identification device 100. In addition, the SDT may have the functions of the income identification device 100 and the information content delivery unit 1500.
[0087] Each function of the income identification device 100 shown in Fig. 15 is the same as that of the above-mentioned embodiment (see Fig. 4). In addition, the subject's terminal 150 includes a display unit 1510 that displays information content. The terminal 150 is a computer used by the subject, such as a PC (Personal Computer), a tablet PC, or a smartphone.
[0088] The income identifying device 100 outputs the income identification result of the subject to the information content delivery unit 1500 via the communication unit 125. The information content delivery unit 1500 includes information content data 1501, an information content extraction unit 1502, and a communication unit 1503.
[0089] The communication unit 1503 of the information content distribution unit 1500 outputs the income identification result of the subject output by the income identification device 100 to the information content extraction unit 1502. The information content extraction unit 1502 extracts information content corresponding to the income identification result (income value of the subject) from the information content data 1501, and outputs it to the terminal 150 via the communication unit 1503 as information content to be distributed. The terminal 150 displays and outputs the information content input via the communication unit 152 on the display unit 1510.
[0090] 16 is an explanatory diagram of an example of information content distribution corresponding to the income level of a target person. Information content data 1501 of the information content distribution unit 1500 includes a plurality of information contents according to high / low income level, for example, different income value stages.
[0091] The information content extraction unit 1502 extracts information content 1501a adapted to the income level corresponding to the income value indicated by the income determination result from the information content data 1501, and outputs (distributes) it to the terminal 150. This makes it possible to display information content adapted to the income of the subject on the display unit 1510 of the terminal 150.
[0092] In the above embodiment, the subject's place of stay is described as two places (commercial facilities P01 and P02), but the subject may stay in three or more places, and the weight for each stay time at each place of stay may be calculated, and the income value may be adjusted by the weight. The more places of stay there are, the more accurately the subject's income value can be estimated.
[0093] (An example of an income specification method according to another embodiment) FIG. 17 is an explanatory diagram showing an example of an income specification method according to another embodiment. The income specification method according to another embodiment acquires movement trajectory data of a subject whose income is to be specified, and specifies the income of the subject based on a place where the subject stayed for a predetermined time or more in the movement trajectory data, the subject's stay time at the place, and the land price of the place. This allows the income to be specified even if attribute information on the subject's income cannot be obtained. Furthermore, for example, attribute information including the specified income can be used to enable the delivery of information, such as information content, adapted to subjects by income.
[0094] In another embodiment of the income specification method, the following processes (1) to (4) are included to specify the income of a subject. (1) Acquisition of subject movement trajectory data (2) Identifying the subject's place of stay (3) Time spent at the place of stay and land price acquisition (4) Identifying income based on land price and length of stay
[0095] In the income identification example shown in FIG. 17, the income identification device 100 identifies one place in the commercial facility P01 as the place of stay of the target person U1 based on the movement trajectory data m1. In this case, the income identification device 100 determines the stay time "T A " Based on the land price of "A million yen," determine the income of subject U1.
[0096] In addition, the income identification device 100 identifies one location in the commercial facility P02 as the place of stay of the target person U2 based on the movement trajectory data m2. In this case, the income identification device 100 identifies the stay time "T B " and based on the land price of "B million yen," determine the income of subject U2.
[0097] For example, suppose that subject U1 stays for "1 hour" at commercial facility P01, and subject U2 stays for "3 hours" at commercial facility P02. If the land prices of commercial facilities P01 and P02 are the same, the income identification device 100 will weight the income higher the longer the stay time, and calculate the income of subject U2 higher than that of subject U1. The land prices for each place of stay can be used as weights (coefficients) for income, similar to the stay times described above.
[0098] In yet another embodiment, the income of the subject may be determined only from the place of stay and the duration of stay without using the land price at the place of stay to calculate the income of the subject. Furthermore, the income may be weighted for each store name at the place of stay. For example, the luxury level of each store may be used as a weight based on the difference in store name. For example, a setting is stored in which a store with the name of a luxury department store has a higher luxury level (higher weight) than a store with the name of a 100-yen store. The income determining device 100 can determine the income of the subject by acquiring the name of the store where the subject stayed and the duration of stay at this store.
[0099] According to the embodiment described above, it is possible to improve the accuracy of various measures and problem solving in SDT. As for "income" among the attribute information of subjects handled by SDT, it is not necessary to purchase it from an external institution, etc., and income values can be specified with high accuracy, so that the reaction in SDT to the proposed measures and problems can be made closer to reality, and the accuracy of measures and problem solving can be improved.
[0100] In addition, in the embodiment, an example of a distribution service of information content adapted to the income of a target person has been described, but the present invention is not limited to this, and the income value identified by the income identification device can also be used to monitor the status of targets based on income levels and to predict future behavior, among various other services.
[0101] As described above, the income identification device acquires the movement trajectory data of the subject, and performs processing to identify the income of the subject based on the places where the subject stayed for a predetermined time or more in the movement trajectory data and the duration of the subject's stay at the places. This makes it possible to estimate the income information of the subject based on the movement state of the subject, and to identify the income of the subject without the trouble of obtaining information from an external source.
[0102] The income identification device may also include a process of acquiring the land price of the place of stay and identifying the income of the subject, including the acquired land price. In this way, by including the land price of the place of stay, the income of the subject can be identified more accurately.
[0103] The income determination device may also adjust the income of the subject based on a weight that increases the impact on the estimated income the longer the subject's stay time at the stay location. Based on the positive correlation between stay time and purchase price, and the longer the stay time, the more it affects the purchase, and therefore the higher the subject's income, weighting the stay time allows for a more accurate estimation of the subject's income.
[0104] The income identification device may also identify a place of stay where the subject stayed for a predetermined time or more from a plurality of samples of individual movement trajectory data whose true income is known in advance, obtain the land price of the place of stay, create an income estimation model for estimating the individual's income based on the true income of the individual, the place of stay, the stay time at the place of stay, and the land price, and input information on the stay time of the subject at the place of stay into the created income estimation model to create an income estimation model for estimating the income of the subject. Then, information on the stay time of the subject at the place of stay can be input into the created income estimation model to estimate the income of the subject. In this way, by using the income estimation model, it becomes possible to accurately identify the income of the subject based on the behavioral status of the subject that is similar to the behavioral status of the individual indicated by the model.
[0105] The income identification device may estimate that a place where the subject stays for a predetermined time or more in a predetermined time period is the subject's home, and adjust the subject's income including the estimated land price and weight of the home. The land price of the home is also effective in identifying income, and the accuracy of identification can be further improved.
[0106] The income identification device may also represent the identified income of the subject in a virtual space based on the subject's real-world behavior, and output the income as attribute information handled by the digital twin that performs processing for a predetermined policy search. This makes it possible to provide part of the attribute information required for various policy searches performed by the digital twin.
[0107] In addition, the income specification device may perform the process of searching for a predetermined measure by delivering information content adapted to the income of the target person to the target person. This makes it possible to deliver information content that is adapted to the income of the target person from among a large number of information contents and that is more suitable for each target person.
[0108] The income identification method described in this embodiment can be realized by executing a prepared program on a computer such as a PC or a workstation. The income identification program described in this embodiment is recorded on a computer-readable recording medium and is executed by being read from the recording medium by the computer. The income identification program described in this embodiment may also be distributed via a network such as the Internet.
[0109] The following supplementary notes are further disclosed regarding the above-described embodiment.
[0110] (Appendix 1) Obtain the subject's movement trajectory data, Identifying the income of the subject based on a place where the subject stayed for a predetermined time or more in the movement trajectory data and a stay time of the subject at the place of stay. A method for determining income, the method comprising the steps of:
[0111] (Appendix 2) Obtain the land price of the place of stay, Identifying the income of said subject, including the value of said land acquired; 2. The method of claim 1, further comprising the steps of:
[0112] (Appendix 3) The income of the subject is adjusted based on a weight that has a larger impact on the estimated income as the subject's stay time at the place of stay is longer; 2. The income determination method according to claim 1,
[0113] (Appendix 4) From the movement trajectory data of multiple samples of individuals whose true income values are known in advance, we identify places where people stayed for a certain period of time or more, Obtain the land price of the place of stay; creating an income estimation model for estimating the individual's income based on the true value of the individual's income, the place of stay, the stay time at the place of stay, and the land price; Inputting information on the stay time of the subject at the stay location into the created income estimation model, and estimating the income of the subject. 2. The income determination method according to claim 1,
[0114] (Appendix 5) The place where the subject stayed for a predetermined time or more during a predetermined time period is presumed to be the subject's home, adjusting for said subject's income by including the estimated land value and weight of said home; 2. The income determination method according to claim 1,
[0115] (Appendix 6) The income of the identified subject is represented in a virtual space based on the subject's real-world behavior, and output as attribute information handled by the digital twin that performs a predetermined policy search process. 2. The income determination method according to claim 1,
[0116] (Appendix 7) The process of searching for a predetermined policy delivers information content adapted to the income of the subject to the subject; 7. The income determination method according to claim 6,
[0117] (Appendix 8) Obtain the subject's movement trajectory data, Identifying the income of the subject based on a place where the subject stayed for a predetermined time or more in the movement trajectory data and a stay time of the subject at the place of stay. A revenue determination program characterized by causing a computer to execute a process.
[0118] (Appendix 9) Obtain the subject's movement trajectory data, Identifying the income of the subject based on a place where the subject stayed for a predetermined time or more in the movement trajectory data and a stay time of the subject at the place of stay. A revenue specifying device having a control unit that performs processing.
[0119] (Appendix 10) The income identification device described in Appendix 9, characterized in that the control unit is provided as one function of a digital twin that represents the subject's real-world behavior in a virtual space and performs processing to search for specified policies. [Explanation of symbols]
[0120] 100 Revenue identification device 110 Storage section 111 True income data 112 Land Price Data 113 Movement trajectory data 114 Location-Use Data 115 Location and time data 116 Revenue Estimation Model 121 Revenue Estimation Model Creation Department 122 Stay location and time data creation section 123 Home location estimation unit 124 Revenue Specific Department 125,152 Communications Department 150 devices 151 Location information acquisition unit 501 CPU 502 Memory 503 Network Interface 505 Recording media 1001 Home location data 1500 Information Content Distribution Department 1501 Information Content Data 1501a Information Content 1502 Information Content Extraction Unit 1503 Communications Department 1510 Display section E Map information I Income Value NW Network P01, P02 Facilities (Commercial Facilities) m Movement trajectory data U(U1,U2) Target
Claims
1. Acquire the subject's movement trajectory data, Identifying the income of the subject based on a place where the subject stayed for a predetermined time or more in the movement trajectory data and a stay time of the subject at the place of stay. A method for determining income, the method comprising the steps of:
2. Obtain the land price of the place of stay; Identifying the income of said subject, including the value of said land acquired; 2. The method of claim 1, further comprising the steps of:
3. adjusting the income of the subject based on a weight that has a larger impact on the estimated income as the subject's stay time in the place of stay increases; 2. The method of claim 1, wherein the income is determined based on the amount of the income.
4. Identifying locations where individuals stayed for a given period of time or longer from the movement trajectory data of a plurality of samples of individuals whose true income values are known in advance; Obtain the land price of the place of stay; creating an income estimation model for estimating the individual's income based on the true value of the individual's income, the place of stay, the stay time at the place of stay, and the land price; Inputting information on the stay time of the subject at the stay location into the created income estimation model, and estimating the income of the subject.
2. The method of claim 1, wherein the income is determined based on the amount of the income.
5. The place where the subject stayed for a predetermined time or more during a predetermined time period is estimated to be the subject's home; adjusting for said subject's income by including the estimated land value and weight of said home; 2. The method of claim 1, wherein the income is determined based on the amount of the income.
6. The income of the identified subject is output as attribute information handled by a digital twin that represents the subject's real-world behavior in a virtual space and performs a predetermined policy search process.
2. The method of claim 1, wherein the income is determined based on the amount of the income.
7. The process of searching for a predetermined policy delivers information content adapted to the income of the subject to the subject.
7. The method of claim 6, wherein the income is determined by:
8. Acquire the subject's movement trajectory data, Identifying the income of the subject based on a place where the subject stayed for a predetermined time or more in the movement trajectory data and a stay time of the subject at the place of stay. A revenue determination program characterized by causing a computer to execute a process.
Citation Information
Patent Citations
Advertising system and method
JP2010237305A
Extraction device, extraction method and extraction program
JP2018101324A
Prediction device, prediction method, and prediction program
JP2018180996A
