Program, estimate device and method which estimate statistical information from ratio of number of parking vehicles for each vehicle type at each area

Abstract

To provide a program, an estimate device and a method which estimate economical statistical information of residents about an arbitrary area.SOLUTION: An estimate device comprises: a statistical database which stores statistical information, for each area of a first area unit; a vehicle database which stores the number of registered vehicles for each vehicle type, for each area of the first area unit; number-of-vehicles ratio calculation means which calculates a ratio of the number of registered vehicles for each vehicle type with respect to the total number of registered vehicles, for each area of the first area unit, by using the vehicle database; and a machine learning engine which sets the ratio of the number of registered vehicles for each vehicle type as an explanation variable, and trains the statistical information of the statistical database as an objective variable, about the first area unit, and inputs the ratio of the number of parking vehicles for each vehicle type, and estimates the statistical information, about a second area unit different from the first area unit.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a technique for estimating statistical information on residents for each area (region). [Background technology]

[0002] Estimating the statistical information of residents for each area is necessary not only for commercial business activities but also for various economic activities. Conventionally, there is a technology that can appropriately predict economic indicators for each area regardless of publicly known information such as news (see, for example, Patent Document 1). According to this technology, a prediction model is prepared for each economic indicator to be predicted. The prediction model is trained using the number of terminals in each area collected over a specified period as an explanatory variable and the economic indicator as a target variable. Using this, economic indicators can be predicted from the number of terminals for each area collected from terminals. Here, it is also possible to count the number of devices staying in an area for each user attribute (age group, gender, etc.), which makes it possible to predict economic indicators for the area according to user attributes.

[0003] In addition, the Ministry of Internal Affairs and Communications publishes income by municipality (segregation) as statistical information (see, for example, Non-Patent Document 1). This makes it possible to know, for example, average income and average assets for each municipality area. In other words, a user's address can be said to have a high explanatory power for expressing the user's economic indicators. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2019-046376 A [Non-patent literature]

[0005] [Non-Patent Document 1] Ministry of Internal Affairs and Communications, Survey of Municipal Taxation Status, [online], [searched on July 20, 2023], Internet <URL:https: / / www.soumu.go.jp / main_sosiki / jichi_zeisei / czaisei / czaisei_seido / ichiran09_19.html> [Non-Patent Document 2] Kanto Regional Transport Bureau, Number of owned vehicles by city, ward, town, and village, [online], [Searched on July 20, 2023], Internet<URL:https: / / wwwtb.mlit.go.jp / kanto / jidou_gian / toukei / tiiki_betu.html> [Non-Patent Document 3] Supervised learning, [online], [searched July 20, 2023], Internet<https: / / ja.wikipedia.org / wiki / %E6%95%99%E5%B8%AB%E3%81%82%E3%82%8A%E5%AD%A6%E7%BF%92> Summary of the Invention [Problem to be solved by the invention]

[0006] For example, a residential area may be lined with luxury detached houses, while another may be lined with low-rent public housing. Residents in the same area tend to have relatively similar economic statistics, such as average income.

[0007] However, according to the above-mentioned Non-Patent Document 1, only the average income etc. in the city, ward, town, village unit is known. In other words, it is not possible to obtain statistical information such as the average income etc. for an area of ​​a regional unit with a smaller granularity than the regional unit of a city, ward, town, village. Furthermore, even if the technology described in the above-mentioned Patent Document 1 is used, it is not possible to estimate the average income of users in each area.

[0008] Therefore, an object of the present invention is to provide a program, an estimation device, and a method for estimating economic statistical information of residents for any area. [Means for solving the problem]

[0009] According to the present invention, there is provided a program for causing a computer installed in a device for estimating economic statistical information of residents in an area to function, the program comprising: a statistical database that accumulates statistical information for each area of ​​the first regional unit; a vehicle database that accumulates the number of registered vehicles of each vehicle type for each area of ​​a first regional unit; a vehicle number ratio calculation means for calculating a ratio of the number of registered vehicles of each vehicle type to the total number of registered vehicles for each area of ​​a first regional unit using the vehicle database; A machine learning engine that is trained for a first regional unit using the ratio of the number of registered vehicles of each vehicle type as an explanatory variable and the statistical information of the statistical database as a target variable, and that inputs the ratio of the number of parked vehicles of each vehicle type for a second regional unit different from the first regional unit and estimates the statistical information; The present invention is characterized in that it causes a computer to function.

[0010] According to another embodiment of the program of the present invention, The statistical information is the average income or average assets of residents, The first regional unit shall be the city, ward, town, or village unit. The second local unit is a smaller local unit than the first local unit. It is also preferable to function the computer in such a manner.

[0011] According to another embodiment of the program of the present invention, Vehicle type is based on model and / or year It is also preferable to function the computer in such a manner.

[0012] According to another embodiment of the program of the present invention, an address database in which a second area unit is registered for each user; a user weighting coefficient assigning means for assigning a weighting coefficient to a target user based on statistical information of a second area unit in which the target user resides, using an address database; It is also preferable to make the computer function as follows.

[0013] According to another embodiment of the program of the present invention, A user database in which economic features are registered for each user; economic feature updating means for weighting the economic feature of a target user with a weighting factor assigned by a user weighting factor assigning means, using a user database; It is also preferable to make the computer function as follows.

[0014] According to the present invention, there is provided an estimation device for estimating economic statistical information of residents in an area, comprising: a statistical database that accumulates statistical information for each area of ​​the first regional unit; a vehicle database that accumulates the number of registered vehicles of each vehicle type for each area of ​​a first regional unit; a vehicle number ratio calculation means for calculating a ratio of the number of registered vehicles of each vehicle type to the total number of registered vehicles for each area of ​​a first regional unit using the vehicle database; A machine learning engine that is trained for a first regional unit using the ratio of the number of registered vehicles of each vehicle type as an explanatory variable and the statistical information of the statistical database as a target variable, and that inputs the ratio of the number of parked vehicles of each vehicle type for a second regional unit different from the first regional unit and estimates the statistical information; The present invention is characterized by having the following.

[0015] According to the present invention, there is provided an estimation method for an apparatus for estimating economic statistical information of residents in an area, comprising the steps of: The device is a statistical database that accumulates statistical information for each area of ​​the first regional unit; a vehicle database that accumulates the number of registered vehicles of each vehicle type for each area of ​​a first regional unit; Machine learning engine and having A first step of calculating a ratio of the number of registered vehicles of each vehicle type to the number of all registered vehicles in each area of ​​a first regional unit using the vehicle database; A second step of training a machine learning engine using the ratio of the number of registered vehicles of each vehicle type as an explanatory variable for the first regional unit and the statistical information of the statistical database as a target variable, and inputting the ratio of the number of parked vehicles of each vehicle type for a second regional unit different from the first regional unit, and estimating statistical information; The present invention is characterized by carrying out the following steps. Effect of the Invention

[0016] According to the program, estimation device, and method of the present invention, economic statistical information of residents in any area can be estimated. [Brief description of the drawings]

[0017] [Figure 1] FIG. 2 is a functional configuration diagram of an estimation device according to the present invention. [Diagram 2] FIG. 2 is an explanatory diagram of a statistical database and a vehicle database in the present invention. [Diagram 3] FIG. 2 is an explanatory diagram showing a training stage of the estimation device according to the present invention. [Figure 4] FIG. 2 is an explanatory diagram illustrating an estimation stage of the estimation device according to the present invention. [Diagram 5] FIG. 13 is an explanatory diagram of assigning a weighting coefficient to each user in the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0018] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[0019] FIG. 1 is a functional configuration diagram of an estimation device according to the present invention.

[0020] According to FIG. 1, an estimation device 1 estimates economic statistical information of residents in a given area. The estimation device 1 includes a statistical database 101, a vehicle database 102, a vehicle count ratio calculation unit 11, and a machine learning engine 12. The estimation device 1 may also include an address database 103, a user database 104, a user weighting coefficient assignment unit 13, and an economic feature update unit 14. These functional components are realized as programs that cause a computer installed in the device to function. The process flow of these functional components can also be understood as an estimation method.

[0021] In general, the type and year of vehicles owned by residents (users) tend to correlate strongly with the average income and average assets of those residents. In other words, for example, residents in the same income range tend to own vehicles in a relatively similar price range (grade). On the other hand, residents in different income ranges tend to own vehicles in relatively different price ranges. In other words, this is because owning a vehicle is constrained by the income available to purchase it. Residents who own expensive vehicles tend to have higher incomes and assets. <Income range> <Price range of vehicle owned> 0~4 million yen Inexpensive 4 million yen to 6 million yen ↓ 6 million yen to 9 million yen ↓ 9 million yen to 12 million yen ↓ 12 million yen to 20 million yen ↓ 20 million yen and up Expensive

[0022] <Training stage> FIG. 2 is an explanatory diagram of the statistical database and the vehicle database in the present invention.

[0023] [Statistics Database 101] The statistical database 101 accumulates statistical information for each area of ​​the first regional unit. The first area unit is the area unit of the training data, for example, a city, ward, town, or village. The statistical information may be, for example, the average income or average asset value of residents in each area of ​​the first regional unit. Such statistical information may be based on, for example, the "Survey of Municipal Taxation Status, etc." (see, for example, Non-Patent Document 1). This is information that publicizes the average income of residents in each area of ​​a city, town, or village.

[0024] According to FIG. 2, the statistical database 101 stores the following statistical information. <First regional unit (city, ward, town, village)> <Average income (average assets)> Nerima Ward: 4.5 million yen Koto Ward 4.8 million yen

[0025] [Vehicle Database 102] The vehicle database 102 accumulates the number of registered vehicles of each vehicle type for each area of ​​a first regional unit. Vehicle type may be based on model and / or year, for example. The vehicle database 102 may be, for example, the "number of owned automobiles by city, ward, town, and village" registered with a local transport bureau, and may be the number of registered automobiles of each vehicle type in each area of ​​the first regional unit (see, for example, Non-Patent Document 2).

[0026] According to FIG. 2, it is assumed that vehicles of each vehicle type in each area of ​​the first regional unit are registered in the vehicle database 102 as follows. <First local unit (city / ward / town / village)> <Vehicle ID> <Model and year> Nerima-ku 1234 X·2020 model 5678 Y・2021 model year 9012 Z・2021 model year 3456 X·2020 model year 7890 Z・2021 model year ... Koto Ward 2345 Z・2021 model year 6789 Y・2021 model year 0123 X·2020 model year 4567 Z·2021 model year 8901 X・2020 model year ...

[0027] Also, according to FIG. 2, the vehicle database 102 indicates the number of registered vehicles by vehicle type in each area of ​​the first regional unit. <First local unit (city / ward / town / village)> <Vehicle type> <Number of registered vehicles> Nerima-ku X·2020 model 132 units Y・2021 model year 509 units Z・2021 model year 301 units ... Total number of vehicles: 94,285 Koto Ward X 2020 model 90 units Y・2021 model year 205 units Z・2021 model year 106 units ... Total number of units: 41,000

[0028] FIG. 3 is an explanatory diagram showing the training stage of the estimation device in the present invention.

[0029] [Unit Quantity Ratio Calculation Section 11] The vehicle number ratio calculation unit 11 uses the vehicle database 102 to calculate the ratio of the number of registered vehicles of each vehicle type to the total number of registered vehicles for each area of ​​the first regional unit.

[0030] According to FIG. 3, the vehicle number ratio calculation unit 11 calculates the ratio (%) of the number of registered vehicles of each vehicle type for each area of ​​the first regional unit as follows. <First regional unit (city / ward / town / village)> <Vehicle type> <Number ratio (%)> Nerima-ku X·2020 model 0.14(=132 / 94,285) Y 2021 model 0.54 (=509 / 94,285) Z 2021 model 0.32 (=301 / 94,285) ... Koto Ward X 2020 model year 0.22 (= 90 / 41,000) Y 2021 model 0.50 (=205 / 41,000) Z 2021 model 0.26 (=106 / 41,000)

[0031] [Machine Learning Engine 12] (Training stage) In the training stage, the machine learning engine 12 trains for the first area unit using the "proportion (%) of the number of registered vehicles of each vehicle type" as an explanatory variable and the "statistical information" of the statistical database 101 as a target variable. Here, the statistical information is assumed to be average income and average assets. In addition, the explanatory variable is a vector with the vehicle type as n-dimensional and the ratio as an element.

[0032] The machine learning engine 12 may be a regression analysis engine, such as a linear multiple regression analysis or a support vector machine (see, for example, Non-Patent Document 3). The learning model of the machine learning engine 12 may be one that is trained at a predetermined interval.

[0033] According to FIG. 3, the machine learning engine 12 performs supervised learning and is trained using the following training data. <Objective variable (average income)> <Explanatory variable (unit ratio (%))> 4.5 million yen (0.14, 0.54, 0.32, ...) 4.8 million yen (0.22, 0.50, 0.26, ...) Of course, even within the same city, ward, town, or village, there is a certain degree of variation in the income of residents, etc. However, when viewed on a city, ward, town, or village basis, it is believed that there is a strong correlation between the average income, etc. as statistical information, and the type of vehicle owned by the residents.

[0034] <Estimation stage> FIG. 4 is an explanatory diagram showing the estimation stage of the estimation device according to the present invention.

[0035] As shown in FIG. 4, the vehicle database 102 accumulates the number of registered vehicles of each vehicle type for each area of ​​the second regional unit. The second area unit may be an area smaller than the first area unit, for example, divided into a 100 m mesh.

[0036] The number of parked vehicles in each area of ​​the second regional unit may be collected manually, for example. Of course, the positioning information may be acquired by a GPS (Global Positioning System) installed in the vehicle and collected by a mobile phone communication facility. For example, the mobile phone communication facility collects the vehicle ID and positioning information from the vehicle during the nighttime hours (for example, from 10 p.m. to 4 a.m.). The vehicle's location during the nighttime hours is highly likely to be the owner's address. In addition, the vehicle's positioning information can be corrected by map matching using a map, so spatial errors are also reduced. Although GPS positioning information is generally expressed in terms of latitude and longitude, it may be expressed as map coordinates or converted into map mesh numbers.

[0037] Also, according to FIG. 4, it is assumed that vehicles of each vehicle type are parked in each area of ​​the second area unit in the vehicle database 102 as follows. <Secondary regional unit (mesh number)> <Vehicle ID> <Vehicle model and year> 5339550137 1234 X·2020 model year 5678 Y・2021 model year 9012 Z・2021 model year 3456 X·2020 model year 7890 Z・2021 model year ... 5339367358 2345 Z·2021 model year 6789 Y・2021 model year 0123 X·2020 model year 4567 Z·2021 model year 8901 X・2020 model year ...

[0038] Also, according to FIG. 4, the vehicle database 102 also indicates the number of parked vehicles by vehicle type in each area of ​​the second area unit, as follows: <Secondary local unit> <Vehicle type> <Number of parking spaces> 5339550137 X 2020 model 5 units Y・2021 model year 19 units Z 2021 model 11 units ... Total number of vehicles: 3,571 5339367358 X 2020 model 3 units Y・2021 model year 7 units Z 2021 model 3 units ... Total number of units: 1,400

[0039] The vehicle number ratio calculation unit 11 uses the statistical database 101 and the vehicle database 102 to calculate the ratio of the number of parked vehicles of each vehicle type to the total number of parked vehicles for each area of ​​the second regional unit.

[0040] According to FIG. 4, it is assumed that the vehicle number ratio calculation unit 11 calculates the ratio (%) of the number of registered vehicles of each vehicle type for each area of ​​the second regional unit as follows. <Secondary regional unit> <Vehicle type> <Volume ratio (%)> 5339550137 X 2020 model 0.14(=5 / 3,571) Y 2021 model 0.53(=19 / 3,571) Z 2021 model 0.31 (=11 / 3,571) ...

[0041] [Machine Learning Engine 12] (Estimation stage) In the estimation step, the machine learning engine 12 inputs the ratio of the number of registered vehicles of each vehicle type for a second regional unit different from the first regional unit, and estimates statistical information. The second regional unit is a regional unit smaller than the first regional unit. When the first regional unit is a city, town, or village unit, the second regional unit can estimate statistical information in an area of ​​a regional unit (such as a block or mesh) smaller than the first regional unit.

[0042] According to FIG. 4, the machine learning engine 12 inputs the following explanatory variables and estimates the objective variable: <Explanatory variable (unit ratio (%))> <Objective variable (average income)> (0.14, 0.53, 0.31, ...) 4.16 million yen That is, for area 5339550137 in the second regional unit, the average income of its residents can be estimated to be 4.16 million yen.

[0043] FIG. 5 is an explanatory diagram of how a weighting coefficient is assigned to each user in the present invention.

[0044] [Address Database 103] The address database 103 registers areas in second regional units for each user. The address database 103 may be a database in which a second area unit is linked to an address registered in advance for each user, for example, from a resident's card. Alternatively, the positioning information may be acquired by a GPS installed in a smartphone owned by a user and collected by a mobile phone communication facility. For example, the mobile phone communication facility collects positioning information from a user's smartphone during the nighttime (e.g., from 10 p.m. to 4 a.m.). The location of the smartphone during the nighttime is highly likely to be the user's address.

[0045] [User weighting coefficient assignment unit 13] The user weighting coefficient assigning unit 13 uses the address database 103 to assign to the target user a "weighting coefficient" based on statistical information of the second area unit where the target user resides. Here, a weighting factor is assigned based on the average income estimated for the second area unit that the user resides in. Of course, the average income value may be used as the weighting factor as is, or the average income value may be input to an arbitrary function and the output may be used as the weighting factor. Weighting factor = f (average income of the second geographical area where the target user lives)

[0046] [User Database 104] The user database 104 registers economic features for each user. The economic feature amount is an arbitrary feature amount based on the behavioral tendency of the user, and may be, for example, the following: That is, it is a numerical value based on the behavior of each user that is likely to change depending on the economic power. Average time spent outside on holidays ·Travel frequency Frequency of eating out ...

[0047] [Economic Features Update Part 14] The economic feature update unit 14 uses the user database 104 to weight the economic feature of the target user by the weighting factor assigned by the user weighting factor assignment unit 13 . In this way, by assigning the weighting coefficient estimated by the present invention to the economic feature quantity for each user, it is possible to improve the accuracy of the explanatory power of the economic feature quantity.

[0048] <Grouping multiple users> In another embodiment, for example, assume that a plurality of groups to which a plurality of users belong are formed. In this case, an economic feature between the groups can be obtained by calculating an average weighting coefficient (average value of weighting coefficients of a plurality of users) for each group. Of course, the economic strength between the groups can also be compared using the average weighting coefficient. This can also be used as any economic indicator (for example, retail sales amount, economic activity index, etc.).

[0049] As described above in detail, according to the program, estimation device, and method of the present invention, economic statistical information of residents for any area can be estimated. Specifically, for example, it is possible to estimate the average income of residents in a regional area that is finer than the regional area of ​​a city, ward, town, or village. In addition, the weighting coefficient of the user that is assigned according to the average income of residents can be used in various economic activities.

[0050] Regarding the various embodiments of the present invention described above, various changes, modifications, and omissions within the scope of the technical idea and perspective of the present invention can be easily made by those skilled in the art. The above description is merely an example and is not intended to be restrictive in any way. The present invention is limited only by the scope of the claims and their equivalents. [Explanation of symbols]

[0051] 1 Estimation device 101 Statistical Database 102 Vehicle Database 103 Address Database 104 User Database 11 Unit Quantity Ratio Calculation Section 12 Machine Learning Engine 13 User weighting factor assignment unit 14 Economic feature update section

Claims

1. A program for causing a computer installed in a device for estimating economic statistical information of residents in an area to function, a statistical database that accumulates statistical information for each area of ​​the first regional unit; a vehicle database that accumulates the number of registered vehicles of each vehicle type for each area of ​​a first regional unit; a vehicle number ratio calculation means for calculating a ratio of the number of registered vehicles of each vehicle type to the total number of registered vehicles for each area of ​​the first regional unit using the vehicle database; a machine learning engine that is trained for a first regional unit using the ratio of the number of registered vehicles of each vehicle type as an explanatory variable and the statistical information of the statistical database as a target variable, and that inputs the ratio of the number of parked vehicles of each vehicle type for a second regional unit different from the first regional unit and estimates the statistical information; A program for causing a computer to function.

2. The statistical information is the average income or average assets of residents, The first regional unit shall be the city, ward, town, or village unit; The second local unit is a smaller local unit than the first local unit.

2. The program according to claim 1, which causes a computer to function as described above.

3. Vehicle type is based on model and / or year 2. The program according to claim 1, which causes a computer to function as described above.

4. an address database in which a second area unit is registered for each user; a user weighting coefficient assigning means for assigning a weighting coefficient to a target user based on statistical information of a second area unit in which the target user resides, using an address database; 2. The program according to claim 1, further comprising:

5. A user database in which economic features are registered for each user; economic feature updating means for weighting the economic feature of a target user with a weighting factor assigned by a user weighting factor assigning means, using a user database; 5. The program according to claim 4, which causes a computer to function as described above.

6. An estimation device for estimating economic statistical information of residents in an area, comprising: a statistical database that accumulates statistical information for each area of ​​the first regional unit; a vehicle database that accumulates the number of registered vehicles of each vehicle type for each area of ​​a first regional unit; a vehicle number ratio calculation means for calculating a ratio of the number of registered vehicles of each vehicle type to the total number of registered vehicles for each area of ​​the first regional unit using the vehicle database; a machine learning engine that is trained for a first regional unit using the ratio of the number of registered vehicles of each vehicle type as an explanatory variable and the statistical information of the statistical database as a target variable, and that inputs the ratio of the number of parked vehicles of each vehicle type for a second regional unit different from the first regional unit and estimates the statistical information; An estimation device comprising:

7. An estimation method for an apparatus for estimating economic statistical information of residents in an area, comprising: The device is a statistical database that accumulates statistical information for each area of ​​the first regional unit; a vehicle database that accumulates the number of registered vehicles of each vehicle type for each area of ​​a first regional unit; Machine learning engine and having A first step of calculating a ratio of the number of registered vehicles of each vehicle type to the total number of registered vehicles for each area of ​​a first regional unit using a vehicle database; a second step of training a machine learning engine using the ratio of the number of registered vehicles of each vehicle type as an explanatory variable for the first regional unit and the statistical information of the statistical database as a target variable, and inputting the ratio of the number of parked vehicles of each vehicle type for a second regional unit different from the first regional unit, and estimating statistical information; A method for estimating an apparatus, comprising:

Images