Household estimation system and household estimation program

Abstract

To provide a household estimation system obtaining demographic that pays attention to individual households, and a household estimation program.SOLUTION: A household estimation system includes at least one processor accessible to a database storing building data and a database storing statistical data. The one processor at least identifies a plurality of residential blocks in an object area, determines as a household element a residential block corresponding to a household attribute on the basis of a household distribution showing the number of households every household attribute in the object area, sets a household-resident relationship between the household element and the resident attribute at random on the basis of a resident distribution showing the number of people every resident attribute in the object area, determines a household-resident relationship on the basis of a difference between a given estimated distribution calculated based on the setting of the household-resident relationship and a corresponding reference distribution by repeating processing to change a part of the household-resident relationship at random, and outputs household data showing a household element including the household attribute and the resident attribute.SELECTED DRAWING: Figure 9

Description

[Technical field]

[0001] One aspect of the present disclosure relates to a household estimation system and a household estimation program. [Background technology]

[0002] Conventionally, there has been known a technique for generating mesh data on a map that can be used as statistical data (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2013-228701 A Summary of the Invention [Problem to be solved by the invention]

[0004] The purpose is to obtain population statistics focusing on individual households. [Means for solving the problem]

[0005] A household estimation system according to an aspect of the present disclosure includes at least one processor that can access a building database that stores building data related to residential sections and a statistical database that stores statistical data related to demographics. The at least one processor identifies a plurality of residential sections in a target area from the building data, determines residential sections corresponding to household attributes as household elements based on a household distribution indicating the number of households for each household attribute in the target area obtained from the statistical data, randomly sets household-resident relationships between the household elements and resident attributes based on a resident distribution indicating the number of people for each resident attribute in the target area obtained from the statistical data, repeats a replacement process that randomly changes a part of the household-resident relationships based on an error between a given estimated distribution calculated based on the setting of the household-resident relationships and a corresponding reference distribution obtained from the statistical data, determines household-resident relationships, and outputs household data indicating the household elements including the household attributes and the resident attributes.

[0006] A household estimation program according to one aspect of the present disclosure causes a computer to execute the following steps: accessing a building database that stores building data related to residential sections, and identifying multiple residential sections within a target area from the building data; accessing a statistical database that stores statistical data related to demographics, and determining residential sections corresponding to the household attributes as household elements based on a household distribution indicating the number of households for each household attribute in the target area obtained from the statistical data; randomly setting household-resident relationships between household elements and resident attributes based on a resident distribution indicating the number of people for each resident attribute in the target area obtained from the statistical data; repeating a replacement process that randomly changes part of the household-resident relationships based on an error between a given estimated distribution calculated based on the setting of the household-resident relationships and a corresponding reference distribution obtained from the statistical data, to determine the household-resident relationships; and outputting household data indicating household elements including the household attributes and the resident attributes.

[0007] In this aspect, residential sections corresponding to household attributes are estimated as household elements based on the household distribution. Then, a correspondence between the household elements and resident attributes is randomly set based on the resident distribution, and the correspondence is determined while being repeatedly changed based on comparison with statistical data. Then, household data indicating the household elements including the household attributes and resident attributes is output. This household data can provide detailed demographic statistics focusing on individual households.

[0008] According to one aspect of the present disclosure, a household estimation system includes at least one processor having access to a building database that stores building data relating to residential areas and a statistical database that stores statistical data relating to demographics. At least one processor identifies multiple residential sections within the target area from the building data, randomly sets a section-personnel relationship between the residential section and household members based on a personnel-household distribution indicating the number of households per household member in the target area obtained from the statistical data, repeats the process of randomly changing part of the section-personnel relationship based on an error between a given estimated distribution calculated based on the section-personnel relationship and a corresponding reference distribution obtained from the statistical data, to determine the residential section corresponding to the household members as a household element, randomly sets household-family relationships between household elements and family types based on a family-household distribution indicating the number of households for each family type in the target area obtained from the statistical data, and repeats the process of randomly changing part of the household-family relationships based on an error between a given estimated distribution calculated based on the household-family relationship and a corresponding reference distribution obtained from the statistical data, to determine the household-family relationships, and outputs household data indicating household elements including household members and family type.

[0009] A household estimation program according to one aspect of the present disclosure includes the steps of: accessing a building database that stores building data related to residential sections, and identifying a plurality of residential sections within a target area from the building data; accessing a statistical database that stores statistical data related to demographics, and randomly setting a section-personnel relationship between residential sections and household members based on a person-household distribution indicating the number of households per household member in the target area obtained from the statistical data; and randomly changing a portion of the section-personnel relationship based on an error between a given estimated distribution calculated based on the section-personnel relationship and a corresponding reference distribution obtained from the statistical data. The method has a computer execute the steps of: repeating the process of determining residential areas corresponding to the household members as household elements; randomly setting household-family relationships between household elements and family types based on a family-household distribution indicating the number of households for each family type in the target area obtained from statistical data; repeating the process of randomly changing some of the household-family relationships based on the error between a given estimated distribution calculated based on the household-family relationships and a corresponding reference distribution obtained from statistical data to determine household-family relationships; and outputting household data indicating household elements including the household members and family type.

[0010] In this aspect, a block-person relationship between household members and residential blocks is randomly set based on the person-household distribution, and the block-person relationship is repeatedly changed based on a comparison with statistical data to determine a residential block corresponding to the household members as a household element. Next, a household-family relationship between the household element and family type is randomly set based on the family-household distribution, and the household-family relationship is repeatedly changed and determined based on a comparison with statistical data. Then, household data indicating household elements including household members and family type is output. This household data can provide detailed demographic statistics focusing on individual households. Effect of the Invention

[0011] According to one aspect of the present disclosure, demographics can be obtained that focus on individual households. [Brief description of the drawings]

[0012] [Figure 1] FIG. 2 is a diagram illustrating an example of a functional configuration of the household estimation system according to the embodiment. [Diagram 2] FIG. 2 is a diagram illustrating an example of a hardware configuration of a computer used in the household estimation system. [Diagram 3] FIG. 2 is a diagram illustrating an example of a data structure of building data. [Figure 4] FIG. 4 is a diagram illustrating an example of a data structure of statistical data. [Diagram 5] FIG. 4 is a diagram illustrating an example of a data structure of statistical data. [Figure 6] FIG. 4 is a diagram illustrating an example of a data structure of statistical data. [Figure 7] FIG. 4 is a diagram illustrating an example of a data structure of statistical data. [Figure 8] FIG. 4 is a diagram illustrating an example of a data structure of statistical data. [Figure 9] 1 is a flowchart showing a series of processes for household estimation by the household estimation system. [Figure 10] 13 is a flowchart showing a process for identifying a residential section. [Figure 11] 13 is a flowchart illustrating a process for determining household elements. [Figure 12] 13 is a flowchart showing a process for determining a family type of each household element. [Figure 13] 13 is a flowchart showing a process for determining resident attributes of each household element. [Figure 14] 13 is a flow chart illustrating a process for determining household-resident relationships. [Figure 15] FIG. 13 is a diagram showing an example of a display of household data. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same or equivalent elements are denoted by the same reference numerals, and duplicated descriptions will be omitted.

[0014] [System Overview] The household estimation system 1 according to the embodiment is a computer system for obtaining demographics focusing on individual households. The household estimation system 1 estimates the residential section in which the household resides and details of each household based on building data on the residential section and statistical data on regional demographics. A household is a unit of a group of one or more people living together. Building data refers to data on individual buildings that physically exist in the real world. A residential section is a space in which one household can reside. Typically, a residential section is a unit of real estate that is the subject of an ownership or lease contract. For example, one detached house corresponds to one residential section, and an apartment building has multiple residential sections. The statistical data in the present disclosure is data obtained by conducting a survey on population (e.g., a census), and indicates, for example, the distribution of the population or number of households based on various attributes.

[0015] [System configuration] FIG. 1 is a diagram showing an example of a functional configuration of a household estimation system 1 according to an embodiment. The household estimation system 1 includes an attribute assignment unit 11, a residential section extraction unit 12, a household estimation unit 13, a resident estimation unit 14, and an output unit 15 as functional elements. The attribute assignment unit 11 is a functional element that assigns a regional attribute to a residential section candidate, which is a property that may be processed as a residential section. The regional attribute refers to the nature or characteristics of a region. The residential section extraction unit 12 is a functional element that extracts residential sections from residential section candidates. The household estimation unit 13 is a functional element that estimates residential sections in which households are expected to exist as household elements. The household estimation unit 13 associates household attributes with at least some residential sections and determines the association as a household element. The household attribute refers to the nature or characteristics of a household. The resident estimation unit 14 is a functional element that estimates the resident attribute of each household element. The resident attribute refers to the nature or characteristics of the residents who make up the household. The output unit 15 is a functional module that outputs household data indicating the household attribute and the household element including the resident attribute.

[0016] The household estimation system 1 can communicate with the database group 2 via a communication network to read and write data. The configuration and type of the communication network are not limited in any way and may be designed according to any policy. For example, the communication network may be constructed using the Internet, an intranet, or a combination of these. The communication network may also be constructed using a wired network, a wireless network, or a combination of these.

[0017] FIG. 2 is a diagram showing an example of a hardware configuration of a computer 110 used in the household estimation system 1. The computer 110 is an example of one or more computers constituting the household estimation system 1. The computer 110 has a control circuit 100. The control circuit 100 has one or more processors 101, a memory 102, a storage 103, a communication port 104, and an input / output port 105. The processor 101 executes an operating system and an application program. The storage 103 is composed of a storage medium such as a hard disk, a non-volatile semiconductor memory, or a removable medium (e.g., a magnetic disk, an optical disk, etc.), and stores the operating system and the application program. The memory 102 temporarily stores a program loaded from the storage 103, or a calculation result by the processor 101. In one example, the processor 101 functions as each of the above-mentioned functional modules by executing a program in cooperation with the memory 102. The communication port 104 performs data communication with other devices via a communication network NW according to a command from the processor 101. The input / output port 105 executes input / output of electrical signals to / from input / output devices (user interfaces) such as a keyboard, a mouse, and a monitor (not shown) in accordance with instructions from the processor 101 .

[0018] [program] The program for causing a computer to function as the household estimation system 1 includes program codes for causing the computer to function as an attribute assignment unit 11, a residential section extraction unit 12, a household estimation unit 13, a resident estimation unit 14, and an output unit 15. At least a part of the program may be provided after being fixedly recorded on a non-transitory recording medium such as a CD-ROM, a DVD-ROM, or a semiconductor memory. Alternatively, at least a part of the program may be provided via a communication network as a data signal superimposed on a carrier wave. The provided program is stored in the storage 103, and the processor 101 executes the program in cooperation with the memory 102 to realize functional elements corresponding to the program.

[0019] The household estimation system 1 may be configured with one or more computers. When multiple computers are used, the household estimation system 1 is configured by connecting these computers to each other via a communication network.

[0020] [Data Structure] The database group 2 is a device (storage unit) that stores various data used for processing by the household estimation system 1. The database group 2 may be provided in a computer or computer system different from the household estimation system 1, or may be a component of the household estimation system 1. In one example, the database group 2 stores data provided from various computer systems, devices, or apparatuses. As shown in FIG. 1, in this embodiment, the database group 2 includes a building database 21 that stores building data and a statistical database 22 that stores statistical data. The structures of these data will be described in detail with reference to FIGS. 3 to 8. FIG. 3 is a diagram showing an example of the data structure of the building data. FIGS. 4 to 8 are diagrams showing an example of the data structure of the statistical data.

[0021] In one example, the building data includes detached house data 211, apartment building data 212, and facility data 213. The detached house data 211 is data related to a detached house. In one example, each record of the detached house data 211 includes a building ID, a building type, a barycentric coordinate, a shape, and an area. The apartment building data 212 is data related to an apartment building. In one example, each record of the apartment building data 212 includes a building ID, a building type, a barycentric coordinate, a shape, an area, and room information. The facility data 213 is data related to a facility. A facility refers to a building where a specific person stays for a certain period of time. Examples of facilities include a care facility, a nursing home, and a hospital. In one example, each record of the facility data 213 includes a building ID, a building type, a barycentric coordinate, a shape, and an area.

[0022] The building ID is an identifier for uniquely identifying each building. The building type is the type or classification of the building, for example, "private house," "apartment," "office / residence," "hospital," "nursing home," "care facility," etc. The centroid coordinates are coordinates indicating the center of gravity of the land on which the building is located. The shape indicates the planar or three-dimensional shape of the building. The area is expressed, for example, by the total floor area of ​​the building. The room information of the apartment building data indicates each room (each dwelling unit) in the apartment building, and includes, for example, the room name, the floor on which it is located, etc.

[0023] In one example, the statistical data includes small area data 221, personnel-household distribution data 222, family-household distribution (7 categories) data 223, resident distribution data 224, housing-household distribution data 225, area-household distribution data 226, family-household distribution (17 categories) data 227, couple age distribution data 228, household member age distribution data 229, and family-age distribution data 230.

[0024] The small area data 221 is data related to a small area that is an area that constitutes a given administrative area. For example, a small area may be defined by a region system unique to Japan called "cho-cho-aza" that indicates divisions within an administrative area by a town name and a block code. In one example, each record of the small area data 221 includes a small area ID, an administrative area ID, a small area name, a centroid coordinate, an area, and a perimeter. The small area ID is an identifier for uniquely identifying each small area. The administrative area ID is an identifier for uniquely identifying each administrative area. The administrative area is, for example, a prefecture or a city, ward, town, or village. The centroid coordinate is a coordinate indicating the centroid of the small area. The perimeter is the length of the outer edge of the small area.

[0025] The personnel-household distribution data 222 is data indicating the number of households by household member in each small area. Household members are the number of people constituting a household, and are an example of a household attribute. In one example, each record of the personnel-household distribution data 222 includes a small area ID, data items for general households, and data items for institutional households. The data items for general households include the number of one-person households, the number of two-person households, the number of three-person households, the number of four-person households, the number of five-person households, the number of six-person households, and the number of households with seven or more people. The data items for institutional households include the number of households and the number of household members. An ordinary household is a household that lives in a detached house or an apartment building. An institutional household is a household that stays in a facility.

[0026] Family-household distribution data (7 categories) 223 is data showing the number of households and total number of people for each family type in each small area. Family type refers to the type of relationship between the residents who make up a household, and is an example of a household attribute. In one example, each record of family-household distribution data (7 categories) 223 includes a small area ID and the number of households and total number of people for each of the seven family types. In one example, the seven family types are "relatives only," "couple only," "couple and children," "households other than nuclear families," "households including non-relatives," "single," and "unknown."

[0027] The resident distribution data 224 is data showing the population by age and sex in each small area. In one example, each record of the resident distribution data 224 includes a small area ID, a male population by age, and a female population by age. For example, the population by age is generally tallied in 5-year increments, such as the population aged 0-4, the population aged 5-9, ..., the population aged 95-99, and the population aged 100 or over.

[0028] The housing-household distribution data 225 is data indicating the number of households by building type in each small area. In one example, each record of the housing-household distribution data 225 includes a small area ID and the number of households for each building type. In one example, the building types include detached houses, row houses, apartment buildings, 1-2 story houses, 3-5 story houses, 6-10 story houses, 11 story or higher houses, and facilities. Row houses and apartment buildings are classifications of collective housing. Row houses are buildings in which each dwelling unit has an entrance that directly connects to the outside, while apartment buildings are buildings in which multiple dwelling units share an entrance.

[0029] The area-household distribution data 226 is data showing the number of households for each combination of building type and area in each administrative region. In one example, each record of the area-household distribution data 226 includes an administrative region ID and the number of households for each combination of building type and area. Examples of building types include detached houses, row houses, and apartment buildings. Examples of area classifications include 29 m2, 30 m2, 40 m2, 50 m2, 60 m2, 70 m2, 80 m2, 90 m2, 100 m2, 120 m2, 140 m2, 160 m2, 180 m2, 200 m2, 250 m2, 300 m2, 400 m2, 500 m2, 600 m2, 700 m2, 800 m2, 100 m2, 120 m2, 140 m2, 160 m2, 18 ... 2 Below 30m 2 ~49m 2 , 50m 2 ~69m 2 , 70m 2 ~99m 2 , 100m 2 ~149m 2 , and 150m 2 Including the above.

[0030] The family-household distribution (17 categories) data 227 is data showing the number of households and the total number of people for each family type in each administrative region. In one example, each record of the family-household distribution (17 categories) data 227 includes an administrative region ID and the number of households and the total number of people for each of the 17 family types. In one example, the 17 family types include "married couple only," "married couple and children," "male parent and children," "female parent and children," "married couple and parents," "married couple and single parent," "married couple, children, and parents," "married couple, children, and single parent," "married couple and other relatives," "married couple, children, and other relatives," "married couple, parents, and other relatives," "married couple, children, parents, and other relatives," "siblings only," "other households other than nuclear families," "households including non-relatives," "single," and "unknown."

[0031] The couple age distribution data 228 is data showing the number of households for each age of husband and wife in each administrative region. In one example, each record of the couple age distribution data 228 includes an administrative region ID and the number of households for each combination of husband and wife ages. In one example, ages are divided into one-year intervals from 15 to 84 years old, and further divided into 85 years old and over.

[0032] The household member age distribution data 229 is data showing the number of households for each age of the household head and household members in each administrative region. In one example, each record of the household member age distribution data 229 includes an administrative region ID and the number of households for each combination of the household head and household member ages. In one example, the household head's age is divided into one-year increments from 15 to 84 years old, and further divided into a category for ages 85 and over. The household members' ages are divided into one-year increments from 0 to 84 years old, and further divided into a category for ages 85 and over.

[0033] The family-age distribution data 230 is data showing the population for each combination of family type, age, and gender in each administrative region. In one example, each record of the family-age distribution data 230 includes an administrative region ID and the population for each combination of family type, age, and gender. In one example, 17 family types are used in the family-age distribution data 230, similar to the family-household distribution (17 categories) data 227. The population by age may be tabulated in 5-year increments in principle, similar to the resident distribution data 224.

[0034] The method of preparing each data is not limited. For example, the detached house data 211, the apartment house data 212, and the facility data 213 may be stored in the building database 21 by a map administrator or a building administrator. The small area data 221 may be generated from cadastral survey data. The personnel-household distribution data 222, the family-household distribution data (7 divisions) 223, the resident distribution data 224, the housing-household distribution data 225, the area-household distribution data 226, the family-household distribution (17 divisions) data 227, the couple age distribution data 228, the household member age distribution data 229, and the family-age distribution data 230 may be generated from census data.

[0035] The data structure of each piece of data is not limited to the examples shown in Figures 3 to 8, and may be designed according to any policy. For example, any piece of data may be normalized or denormalized according to any policy and stored in one or more data tables.

[0036] [System Operation] The operation of the household estimation system 1, that is, the household estimation method according to the present embodiment, will be described with reference to Fig. 9. Fig. 9 is a flowchart showing a series of processes for household estimation by the household estimation system 1 as a process flow S1.

[0037] In step S11, the attribute assigning unit 11 selects an administrative region. In one example, the attribute assigning unit 11 may select an administrative region based on a user operation, or may automatically select an administrative region based on a given rule. As described above, the selected administrative region may be a prefecture or a city, ward, town, or village.

[0038] The processing from step S12 onwards is executed for each target area within the selected administrative region. The target area is an area selected as a target for estimating demographic statistics focusing on individual households. For example, the target area is each of one or more small areas that make up the selected administrative region.

[0039] In step S12, the attribute assignment unit 11 identifies residential section candidates for each target area in the administrative region. In one example, the attribute assignment unit 11 identifies residential section candidates for each target area as follows. That is, the attribute assignment unit 11 defines the geographical range of the target area based on the centroid coordinates, shape, and area of ​​the target area indicated by the small area data 221. The attribute assignment unit 11 also refers to the detached house data 211, the apartment building data 212, and the facility data 213 to select one or more buildings whose centroid coordinates are located within the geographical range. Then, the attribute assignment unit 11 identifies one or more residential section candidates corresponding to the selected buildings.

[0040] In step S13, the attribute assigning unit 11 assigns area attributes to each of the identified residential section candidates. For example, the attribute assigning unit 11 assigns a small area ID, an administrative area ID, an administrative area name (e.g., a prefecture name and a city, town, or village name), and a small area name corresponding to the target area. In one example, the attribute assigning unit 11 acquires the administrative area name corresponding to the administrative area ID from master data (not shown) that manages a list of administrative areas.

[0041] In step S14, the residential section extraction unit 12 identifies residential sections for each target area. Details of step S14 will be described with reference to Fig. 10. Fig. 10 is a flowchart showing the process of identifying residential sections in the target area.

[0042] In step S1401, the residential area extraction unit 12 selects one target area.

[0043] In step S1402, the residential section extraction unit 12 determines whether or not at least one residential section candidate exists in the target area.

[0044] In step S1402, if a residential section candidate exists (YES in step S1402), the process proceeds to step S1403. In step S1403, the residential section extraction unit 12 identifies each residential section in the target area based on the building type of the residential section candidate.

[0045] In one example, the residential area extraction unit 12 refers to the personnel-household distribution data 222 to calculate the total number N of private households in the target area. t Then, the residential area extraction unit 12 calculates N t The residential section extraction unit 12 first identifies residential section candidates whose building type is "private house", "apartment house", "office / residence", "office building", "individual house", "building ... t If the number of residential section candidates of "individual house" and "apartment house" is smaller than the total number N, the residential section extraction unit 12 identifies each residential section candidate whose building type is "apartment house" as a residential section. t If the number of residential blocks is smaller than N, the residential block extraction unit 12 identifies each residential block candidate whose building type is "office / residence" as a residential block. tThe residential section extraction unit 12 identifies residential section candidates in order starting from the first priority until residential sections are identified. The residential section extraction unit 12 may identify only the residential section candidates with the first priority as residential sections, or may identify residential section candidates with the first priority to the jth priority as residential sections. In one example, the residential section extraction unit 12 sets a flag "1" to the residential section corresponding to the building indicated by the building data.

[0046] In one example, the residential area extraction unit 12 determines whether there is a facility (e.g., a nursing home, a hospital, etc.) to which the household can be associated. If such a facility is present, the residential area extraction unit 12 refers to the personnel-household distribution data 222 and calculates the number of facility households N in the target area. h Then, the residential area extraction unit 12 identifies N h The residential section extraction unit 12 identifies residential section candidates as residential sections. The method of identification is the same as that of the residential section of a general household described above. On the other hand, if there is no facility with which the household can be associated, the residential section extraction unit 12 identifies the centroid coordinates of the target area as the residential section. This process is prepared assuming an inconsistency between data, such as a facility household not recorded in the building database 21 but recorded in the statistical database 22. In one example, the residential section extraction unit 12 sets a flag "1" to the residential section corresponding to the facility indicated in the building data, and sets a flag "2" to the residential section indicated by the centroid coordinates of the target area.

[0047] In step S1404, the residential section extraction unit 12 estimates the floor area of ​​each residential section. The residential section extraction unit 12 may calculate the floor area of ​​a detached house as total floor area / total number of rooms. The residential section extraction unit 12 may calculate the floor area of ​​an apartment building as (total floor area x 0.8) / total number of rooms, assuming that 20% of the building is shared space.

[0048] In step S1402, if there are no residential section candidates (NO in step S1402), the process proceeds to step S1405. In step S1405, the residential section extraction unit 12 identifies the centroid coordinates of the target area as a residential section. This process is prepared assuming an inconsistency between data, where a building (residential section) is not recorded in the building database 21 but a household is recorded in the statistical database 22. In one example, the residential section extraction unit 12 sets a flag "2" to the residential section indicated by the centroid coordinates.

[0049] The flag set for the residential section is information indicating whether the residential section corresponds to an actual building or whether it is set in accordance with the barycentric coordinates of the target area.

[0050] In step S1406, the residential area extraction unit 12 determines whether all target areas have been processed. If there are any target areas that have not been processed (NO in step S1406), the process returns to step S1401. The residential area extraction unit 12 selects the next target area in step S1401, and executes the processes from step S1402 onwards for that target area. If all target areas have been processed (YES in step S1406), the residential area extraction unit 12 ends step S14.

[0051] Returning to Fig. 9, in step S15, the household estimation unit 13 determines household elements for each target area based on the personnel-household distribution. Details of step S15 will be described with reference to Fig. 11. Fig. 11 is a flowchart showing the process of determining household elements in a target area.

[0052] In step S1501, the household estimation unit 13 selects one target area.

[0053] In step S1502, the household estimation unit 13 randomly sets a section-personnel relationship, which is a correspondence relationship between the residential section and the household members. In one example, the household estimation unit 13 refers to the member-household distribution data 222 to read a record corresponding to the target area, and identifies the number of households for each household member indicated in the record. Then, the household estimation unit 13 randomly associates household members with the residential section in the target area based on the number of households. For example, assume that the number of one-person households is 10 and the number of two-person households is 8. In this case, the household estimation unit 13 randomly selects 10 residential sections that have not yet been associated with household members, and associates information that a one-person household is associated with these residential sections. Furthermore, the household estimation unit 13 randomly selects 8 residential sections in the target area that have not yet been associated with household members, and associates information that a two-person household is associated with these residential sections. The household estimation unit 13 performs similar processing for other household members to set the section-personnel relationship.

[0054] In step S1503, the household estimation unit 13 calculates the error between the estimated distribution based on the block-personnel relationship and the corresponding reference distribution. In one example, the household estimation unit 13 refers to the housing-household distribution data 225 to read records corresponding to the target area, and acquires the number of households for each building type as a first reference distribution. Furthermore, the household estimation unit 13 refers to the area-household distribution data 226 to read records corresponding to the administrative area to which the target area belongs, and acquires the number of households for each combination of building type and area. Then, the household estimation unit 13 converts each number of households into a ratio (percentage) to acquire a second reference distribution. The household estimation unit 13 generates a first estimated distribution corresponding to the first reference distribution and a second estimated distribution corresponding to the second reference distribution based on the block-personnel relationship. Since each residential block has information on the building type and estimated floor area, the household estimation unit 13 can generate the first and second estimated distributions based on this information. The block-personnel relationship is set for the target area, whereas the area-household distribution data 226 is aggregated for each administrative area. Therefore, the household estimation unit 13 assumes that the distribution indicated by the area-household distribution data 226 is uniform across each small area (target area) constituting the administrative area, and acquires the second reference distribution and the second estimated distribution by the ratio. The household estimation unit 13 calculates a first error of the first estimated distribution and the first reference distribution, and a second error of the second estimated distribution and the second reference distribution, and sets a current error based on these errors according to a given rule. The current error may be represented by a set of the first error and the second error, or may be represented by a set of one or more values ​​obtained by integrating the first error and the second error by a given calculation formula.

[0055] In step S1504, the household estimation unit 13 randomly changes a part of the section-personnel relationships. The household estimation unit 13 randomly changes a part of the section-personnel relationships by randomly selecting a part of the household members associated with the residential section and randomly moving the selected household members to another residential section. The household estimation unit 13 moves one or more pieces of information on the household members. The random movement may be moving the household members from another residential section to a residential section with which the household members are not associated. Alternatively, the random movement may be swapping the household members between two residential sections.

[0056] In step S1505, the household estimation unit 13 recalculates the error between the estimated distribution based on the randomly changed block-person relationships and the corresponding reference distribution. This recalculated error is called a post-change error. In one example, the household estimation unit 13 recalculates the first error and the second error, and calculates the post-change error based on these errors.

[0057] In step S1506, the household estimation unit 13 determines whether the error has become smaller as a result of the random change. The household estimation unit 13 performs this determination by comparing the post-change error with the current error.

[0058] If the post-change error is smaller than the current error (YES in step S1506), the process proceeds to step S1507. In step S1507, the household estimation unit 13 stores the randomly changed block-person relationship. Also, the household estimation unit 13 updates the current error with the post-change error.

[0059] If the post-change error is equal to or greater than the current error (NO in step S1506), the process proceeds to step S1508. In step S1508, the household estimation unit 13 discards the randomly changed block-personnel relationships and the post-change error. The household estimation unit 13 maintains the block-personnel relationships and the current error before the random change.

[0060] In step S1509, the household estimation unit 13 judges whether or not to further randomly change the block-personnel relationship. The condition for ending the random change may be, for example, that the random change has been repeated a given number of times. Alternatively, the end condition may be that the amount of decrease in the current error in a given number of past executions has become equal to or less than a given threshold, that is, that the current error has converged. If the random change is to be repeated (NO in step S1509), the household estimation unit 13 repeatedly executes the processing from step S1504 onward. On the other hand, if the random change is to be ended (YES in step S1509), the processing proceeds to step S1510.

[0061] In step S1510, the household estimation unit 13 determines the block-personnel relationship at the time when the random change is completed as the block-personnel relationship of the target area. The household estimation unit 13 determines the residential block corresponding to the household members as the household elements. The household estimation unit 13 sets a household ID for each household element to uniquely identify each individual household.

[0062] In step S1511, the household estimation unit 13 determines whether all target areas have been processed. If there are target areas that have not been processed (NO in step S1511), the process returns to step S1501. The household estimation unit 13 selects the next target area in step S1501, and executes the processes from step S1502 onwards for that target area. If all target areas have been processed (YES in step S1511), the household estimation unit 13 ends step S15.

[0063] Returning to Fig. 9, in step S16, the household estimation unit 13 determines the family type of each household element for each target area based on the family-household distribution. Details of step S16 will be described with reference to Fig. 12. Fig. 12 is a flowchart showing the process of determining the family type of each household element in the target area.

[0064] In step S1601, the household estimation unit 13 selects one target area.

[0065] In step S1602, the household estimation unit 13 randomly sets household-family relationships, which are correspondence relationships between household elements and family types. In one example, the household estimation unit 13 refers to the family-household distribution data (7 categories) 223 to read a record corresponding to the target area, and identifies the number of households for each family type indicated in the record. Then, based on the number of households, the household estimation unit 13 randomly associates the family type with the household elements in the target area while matching the household members with the family type. For example, if the number of households with the family type "couple and children" is 9, the household members of that family type are at least 3, so the household estimation unit 13 randomly associates the family type with 9 household elements among the household elements with 3 or more members. For the family type "single", the household estimation unit 13 randomly associates the family type with a household element with 1 member. The household estimation unit 13 performs similar processing for other family types to set household-family relationships.

[0066] In step S1603, the household estimation unit 13 calculates the error between the estimated distribution based on the household-family relationship and the corresponding reference distribution. In one example, the household estimation unit 13 refers to the family-household distribution (17 categories) data 227, reads out a record corresponding to the administrative region to which the target region belongs, and acquires the number of households for each of the 17 family types. Then, the household estimation unit 13 converts the distribution of the number of households into the ratio of the number of households for the 7 family types to acquire the reference distribution. The household estimation unit 13 generates an estimated distribution corresponding to the reference distribution based on the household-family relationship. The household-family relationship is set for the target region, whereas the family-household distribution (17 categories) data 227 is aggregated for each administrative region. Therefore, the household estimation unit 13 assumes that the distribution indicated by the family-household distribution (17 categories) data 227 is uniform across each small region (target region) that constitutes the administrative region, and acquires the reference distribution and the estimated distribution by the ratio. The household estimation unit 13 calculates the error between the estimated distribution and the reference distribution as the current error.

[0067] In step S1604, the household estimation unit 13 randomly changes some of the household-family relationships. The household estimation unit 13 randomly swaps some of the family types associated with the household elements between at least two household elements while maintaining consistency between the household members and the family types.

[0068] In step S1605, the household estimation unit 13 recalculates the error between the estimated distribution based on the randomly changed household-family relationships and the corresponding reference distribution. That is, the household estimation unit 13 calculates the post-change error.

[0069] In step S1606, the household estimation unit 13 determines whether the error has become smaller as a result of the random change. The household estimation unit 13 performs this determination by comparing the post-change error with the current error.

[0070] If the post-change error is smaller than the current error (YES in step S1606), the process proceeds to step S1607. In step S1607, the household estimation unit 13 stores the randomly changed household-family relationships. Also, the household estimation unit 13 updates the current error with the post-change error.

[0071] If the post-change error is equal to or greater than the current error (NO in step S1606), the process proceeds to step S1608. In step S1608, the household estimation unit 13 discards the randomly changed household-family relationship. The household estimation unit 13 maintains the household-family relationship before the random change and the current error.

[0072] In step S1609, the household estimation unit 13 determines whether or not to further randomly change the household-family relationships. The condition for ending the random changes may be, for example, that the random changes have been repeated a given number of times. Alternatively, the ending condition may be that the current error has converged. If the random changes are to be repeated (NO in step S1609), the household estimation unit 13 repeatedly executes the processes in and after step S1604. On the other hand, if the random changes are to be ended (YES in step S1609), the process proceeds to step S1610.

[0073] In step S1610, the household estimation unit 13 determines the household-family relationships at the time when the random change is completed as the household-family relationships in the target area.

[0074] In step S1611, the household estimation unit 13 determines whether all target areas have been processed. If there are target areas that have not been processed (NO in step S1611), the process returns to step S1601. The household estimation unit 13 selects the next target area in step S1601, and executes the processes from step S1602 onwards for that target area. If all target areas have been processed (YES in step S1611), the household estimation unit 13 ends step S16.

[0075] Returning to Fig. 9, in step S17, the resident estimation unit 14 determines the resident attributes of each household element for each target area based on the resident distribution. Details of step S17 will be described with reference to Fig. 13. Fig. 13 is a flowchart showing the process of determining the resident attributes of each household element in the target area.

[0076] In step S171, the resident estimation unit 14 selects one target area.

[0077] In step S172, the resident estimation unit 14 randomly sets a household-resident relationship, which is a correspondence relationship between a household element and a resident attribute. In one example, the resident estimation unit 14 refers to the resident distribution data 224 to read a record corresponding to the target area, and identifies the population for each resident attribute (age and sex) indicated in the record. Then, based on the population, the resident estimation unit 14 randomly associates the attributes (age and sex) of each resident with the household elements in the target area while maintaining consistency between the household members, family type, and residents. For example, the resident estimation unit 14 associates the resident attribute with each household element such that the number of residents matches the household members. Alternatively, the resident estimation unit 14 associates the resident attribute with the household element such that a household element including children includes residents aged 19 or younger.

[0078] In step S173, the resident estimation unit 14 calculates the error between the estimated distribution based on the household-resident relationship and the corresponding reference distribution.

[0079] In one example, the resident estimation unit 14 refers to the couple age distribution data 228, reads records corresponding to the administrative region to which the target area belongs, and obtains the number of households for each age of the couple. Then, the resident estimation unit 14 converts the number of households into a ratio to obtain a first reference distribution. Also, the resident estimation unit 14 refers to the household member age distribution data 229, reads records corresponding to the administrative region, and obtains the number of households for each combination of the household head and the household member. Then, the resident estimation unit 14 converts the number of households into a ratio to obtain a second reference distribution. Furthermore, the resident estimation unit 14 refers to the family-age distribution data 230, reads records corresponding to the administrative region, and obtains the population for each combination of family type, age, and sex in 17 categories. The resident estimation unit 14 converts the family type from 17 categories to 7 categories, and then converts the population into a ratio to obtain a third reference distribution. Furthermore, the resident estimation unit 14 refers to the couple age distribution data 228, reads records corresponding to the administrative region to which the target area belongs, and calculates the ratio of the number of men and women for each age. Next, the resident estimation unit 14 refers to the household member age distribution data 229, reads out the record corresponding to the administrative district to which the target district belongs, and obtains the number of household heads by age. Based on the number of household heads and the ratio of males to females, the resident estimation unit 14 obtains the number of household heads by age and gender as a fourth reference distribution.

[0080] The resident estimation unit 14 generates a first estimated distribution corresponding to the first reference distribution, a second estimated distribution corresponding to the second reference distribution, a third estimated distribution corresponding to the third reference distribution, and a fourth estimated distribution corresponding to the fourth reference distribution based on the household-resident relationship. The household-resident relationship is set for the target area, whereas the couple age distribution data 228, the household member age distribution data 229, and the family-age distribution data 230 are collected for each administrative area. Therefore, the resident estimation unit 14 acquires the reference distribution and the estimated distribution by ratio, assuming that the distributions indicated by the respective data are uniform across the individual small areas (target areas) that make up the administrative area.

[0081] The occupant estimation unit 14 calculates a first error of the first estimated distribution and the first reference distribution, a second error of the second estimated distribution and the second reference distribution, a third error of the third estimated distribution and the third reference distribution, and a fourth error of the fourth estimated distribution and the fourth reference distribution, and sets a current error based on these errors according to a given rule. The current error may be represented by a set of the first error, the second error, the third error, and the fourth error, or may be represented by a set of one or more values ​​obtained by integrating the first error, the second error, the third error, and the fourth error using a given calculation formula.

[0082] In step S174, the resident estimation unit 14 executes a process for determining a household-resident relationship. Details of step S174 will be described with reference to Fig. 14. Fig. 14 is a flowchart showing the process for determining a household-resident relationship.

[0083] In step S1741, the resident estimation unit 14 executes a replacement process that randomly changes some of the household-resident relationships. The resident estimation unit 14 randomly replaces some of the residents associated with household elements between at least two household elements while maintaining consistency between household members, family type, and residents. In this replacement process, the resident estimation unit 14 moves m residents. The resident estimation unit 14 may replace only some of the residents or all of the residents for a certain household element.

[0084] In step S1742, the resident estimation unit 14 recalculates the error between the estimated distribution based on the randomly changed household-resident relationships and the corresponding reference distribution. In one example, the resident estimation unit 14 recalculates the first error, the second error, the third error, and the fourth error, and calculates the changed error based on these errors.

[0085] In step S1743, the occupant estimation unit 14 determines whether the error has been reduced by the random change. The occupant estimation unit 14 performs this determination by comparing the post-change error with the current error.

[0086] If the post-change error is smaller than the current error (YES in step S1743), the process proceeds to step S1744. In step S1744, the resident estimation unit 14 stores the randomly changed household-resident relationship. The resident estimation unit 14 also updates the current error with the post-change error.

[0087] If the post-change error is equal to or greater than the current error (NO in step S1743), the process proceeds to step S1745. In step S1745, the resident estimation unit 14 discards the randomly changed household-resident relationship and the post-change error. The resident estimation unit 14 maintains the household-resident relationship before the random change and the current error.

[0088] In step S1746, the resident estimation unit 14 determines whether or not to execute a mutation process. In the present disclosure, a mutation process refers to a process of randomly replacing more residents than the number in the replacement process (step S1741). In one example, if the number of times that the replacement process (step S1741) has been executed is (k-1) times or less, the resident estimation unit 14 does not execute the mutation process, and the process proceeds to step S1749. If the number of times that the replacement process has been executed is k times or more, the process proceeds to step S1747.

[0089] In step S1747, the resident estimation unit 14 determines whether the current error is smaller than the error k times ago (the current error at the point in time k times ago). If the current error is smaller than the error k times ago (YES in step S1747), the resident estimation unit 14 does not execute the mutation process and the process proceeds to step S1749. If the current error is equal to or larger than the error k times ago (NO in step S1747), the process proceeds to step S1748.

[0090] In step S1748, the resident estimation unit 14 executes a mutation process that randomly changes some of the household-resident relationships. The resident estimation unit 14 randomly swaps some of the residents associated with the household elements between at least two household elements while maintaining consistency between household members, family types, and residents. In this mutation process, the resident estimation unit 14 swaps n (>m) residents. That is, the resident estimation unit 14 randomly changes a larger proportion of the household-resident relationships than in the swap process. In this step, the resident estimation unit 14 may swap only some residents or all residents for a certain household element. After step S1748, the process returns to step S1741.

[0091] As shown in steps S1747 and S1748, the occupant estimation unit 14 executes a mutation process during the repetition of the replacement process, based on the history of errors obtained by the repetition of the replacement process.

[0092] In step S1749, the resident estimation unit 14 determines whether or not to further randomly change the household-resident relationship. The condition for ending the random change may be, for example, that the replacement process has been repeated a given number of times. Alternatively, the ending condition may be that the current error has converged. If the random change is to be repeated (NO in step S1749), the resident estimation unit 14 repeatedly executes the processes from step S1741 onward. On the other hand, if the random change is to be ended (YES in step S1749), step S174 ends.

[0093] Returning to FIG. 13, in step S175, the resident estimation unit 14 determines the household-resident relationship at the time when the random change is completed as the household-resident relationship of the target area.

[0094] In step S176, the resident estimation unit 14 determines whether all target areas have been processed. If there are any target areas that have not been processed (NO in step S176), the process returns to step S171. The resident estimation unit 14 selects the next target area in step S171, and executes the processes from step S172 onwards for that target area. If all target areas have been processed (YES in step S176), the resident estimation unit 14 ends step S17.

[0095] Returning to FIG. 9, in step S18, the output unit 15 outputs household data related to the household elements. For example, the output unit 15 displays the household data on a monitor. FIG. 15 is a diagram showing an example of the display of household data. In this example, the output unit 15 displays a screen 300 showing household data. The screen 300 includes a map 301 and household details 302. A plurality of points 310 showing individual household elements are plotted on the map 301. In response to a point 310 being selected by a user operation, the output unit 15 displays household details 302 corresponding to the point 310. In the example of FIG. 15, the household details 302 show information on the household element corresponding to the selected point 311.

[0096] The method of outputting the household data is not limited. For example, the household data may be stored in a given database, transmitted to another computer, or printed. The method of expressing the household data is also not limited. For example, the household data may be expressed using at least one of a map, a table, and a graph. In any case, the user can understand from the household data where (in which building) what kind of households exist in the target area.

[0097] As shown in steps S15 to S17, the household estimation system 1 repeats a process of randomly changing a part of the information related to the household elements so as to reduce the error between the estimated distribution and the reference distribution, i.e., so that the estimated distribution approaches the reference distribution. This can be said to be a process that introduces the Monte Carlo method, which is a method of finding an approximate solution by repeating trials using random numbers. Step S17 can be said to be a process that incorporates mutation, which is an element of a genetic algorithm, into the Monte Carlo method.

[0098] [effect] As described above, the household estimation system according to one aspect of the present disclosure includes at least one processor that can access a building database that stores building data related to residential sections and a statistical database that stores statistical data related to demographics. The at least one processor identifies a plurality of residential sections in a target area from the building data, determines residential sections corresponding to household attributes as household elements based on a household distribution indicating the number of households for each household attribute in the target area obtained from the statistical data, randomly sets household-resident relationships between the household elements and resident attributes based on a resident distribution indicating the number of people for each resident attribute in the target area obtained from the statistical data, repeats a replacement process that randomly changes a part of the household-resident relationships based on an error between a given estimated distribution calculated based on the setting of the household-resident relationships and a corresponding reference distribution obtained from the statistical data, determines household-resident relationships, and outputs household data indicating the household elements including the household attributes and the resident attributes.

[0099] A household estimation program according to one aspect of the present disclosure causes a computer to execute the following steps: accessing a building database that stores building data related to residential sections, and identifying multiple residential sections within a target area from the building data; accessing a statistical database that stores statistical data related to demographics, and determining residential sections corresponding to the household attributes as household elements based on a household distribution indicating the number of households for each household attribute in the target area obtained from the statistical data; randomly setting household-resident relationships between household elements and resident attributes based on a resident distribution indicating the number of people for each resident attribute in the target area obtained from the statistical data; repeating a replacement process that randomly changes part of the household-resident relationships based on an error between a given estimated distribution calculated based on the setting of the household-resident relationships and a corresponding reference distribution obtained from the statistical data, to determine the household-resident relationships; and outputting household data indicating household elements including the household attributes and the resident attributes.

[0100] In this aspect, residential sections corresponding to household attributes are estimated as household elements based on the household distribution. Then, a correspondence between the household elements and resident attributes is randomly set based on the resident distribution, and the correspondence is determined while being repeatedly changed based on comparison with statistical data. Then, household data indicating the household elements including the household attributes and resident attributes is output. This household data can provide detailed demographic statistics focusing on individual households.

[0101] In another aspect of the household estimation system, at least one processor may perform a mutation process that randomly changes a larger proportion of the household-resident relationships during the repetition of the replacement process based on a history of errors obtained by the repetition of the replacement process. If the replacement process is performed alone, the household-resident relationships may fall into a local solution or the errors may not improve. By introducing the mutation process, the household-resident relationships are changed relatively more significantly, making it possible to increase the probability of correctly estimating the household-resident relationships.

[0102] In another aspect of the household estimation system, at least one processor may execute a mutation process when an error after a predetermined number of repetitions of the replacement process is equal to or greater than an error before the repetitions. By executing the mutation process when the error does not decrease, the household-resident relationship can be more accurately estimated without executing the mutation process unnecessarily many times.

[0103] In another aspect of the household estimation system, the household attributes include household members and family type, and at least one processor may execute the following steps to determine the residential section corresponding to the household members as a household element: a step of randomly setting a section-personnel relationship between the residential section and the household members based on a person-household distribution indicating the number of households for each household member in the target area obtained from statistical data; a step of repeating a process of randomly changing a part of the section-personnel relationship based on an error between a given estimated distribution calculated based on the section-personnel relationship and a corresponding reference distribution obtained from statistical data, to determine the residential section corresponding to the household members; a step of randomly setting a household-family relationship between the household element and the family type based on a family-household distribution indicating the number of households for each family type in the target area obtained from statistical data; and a step of repeating the process of randomly changing a part of the household-family relationship based on an error between a given estimated distribution calculated based on the household-family relationship and a corresponding reference distribution obtained from statistical data, to determine the household-family relationship, to determine the residential section corresponding to the household attributes as a household element. In this case, the block-personnel relationship between the household members and the residential block is set randomly based on the person-household distribution, and the block-personnel relationship is repeatedly changed based on comparison with statistical data to determine the residential block corresponding to the household members as the household element. Next, the household-family relationship between the household element and the family type is set randomly based on the family-household distribution, and the household-family relationship is repeatedly changed and determined based on comparison with statistical data. This series of processes allows for accurate estimation of household attributes including the household members and family type.

[0104] According to one aspect of the present disclosure, a household estimation system includes at least one processor having access to a building database that stores building data relating to residential areas and a statistical database that stores statistical data relating to demographics. At least one processor identifies multiple residential sections within the target area from the building data, randomly sets a section-personnel relationship between the residential section and household members based on a personnel-household distribution indicating the number of households per household member in the target area obtained from the statistical data, repeats the process of randomly changing part of the section-personnel relationship based on an error between a given estimated distribution calculated based on the section-personnel relationship and a corresponding reference distribution obtained from the statistical data, to determine the residential section corresponding to the household members as a household element, randomly sets household-family relationships between household elements and family types based on a family-household distribution indicating the number of households for each family type in the target area obtained from the statistical data, and repeats the process of randomly changing part of the household-family relationships based on an error between a given estimated distribution calculated based on the household-family relationship and a corresponding reference distribution obtained from the statistical data, to determine the household-family relationships, and outputs household data indicating household elements including household members and family type.

[0105] A household estimation program according to one aspect of the present disclosure includes a step of accessing a building database that stores building data related to residential sections and identifying a plurality of residential sections within a target area from the building data; a step of accessing a statistical database that stores statistical data related to demographics and randomly setting a section-personnel relationship between residential sections and household members based on a person-household distribution indicating the number of households per household member in the target area obtained from the statistical data; and a step of randomly changing a part of the section-personnel relationship based on an error between a given estimated distribution calculated based on the section-personnel relationship and a corresponding reference distribution obtained from the statistical data. The method has a computer execute the steps of: repeating the process of determining residential areas corresponding to the household members as household elements; randomly setting household-family relationships between household elements and family types based on a family-household distribution indicating the number of households for each family type in the target area obtained from statistical data; repeating the process of randomly changing some of the household-family relationships based on the error between a given estimated distribution calculated based on the household-family relationships and a corresponding reference distribution obtained from statistical data to determine household-family relationships; and outputting household data indicating household elements including the household members and family type.

[0106] In this aspect, a block-person relationship between household members and residential blocks is randomly set based on the person-household distribution, and the block-person relationship is repeatedly changed based on a comparison with statistical data to determine a residential block corresponding to the household members as a household element. Next, a household-family relationship between the household element and family type is randomly set based on the family-household distribution, and the household-family relationship is repeatedly changed and determined based on a comparison with statistical data. Then, household data indicating household elements including household members and family type is output. This household data makes it possible to obtain detailed demographic statistics focusing on individual households.

[0107] [Variations] The present disclosure has been described in detail above based on the embodiments. However, the present disclosure is not limited to the above embodiments. The present disclosure can be modified in various ways without departing from the spirit and scope of the present disclosure.

[0108] The household estimation unit 13 may execute step S15 after executing step S16. That is, the household estimation unit 13 may determine the correspondence between the residential section and the family type as a household element based on the family-household distribution, and then determine the household members of the household element based on the member-household distribution. Either step S15 or S16 for determining the household element (i.e., the residential section corresponding to the household attribute) may be omitted.

[0109] The household estimation system 1 does not need to include the resident estimation unit 14, and therefore step S17 may be omitted.

[0110] The household estimation system 1 may not include the attribute assignment unit 11 and the residential section extraction unit 12, and therefore steps S12 to S14 may be omitted. In other words, the household estimation system may generate household data on the premise that the residential sections of the target area have already been identified.

[0111] The resident estimation unit 14 may determine the household-resident relationship without using the mutation process, that is, steps S1746 to S1748 may be omitted.

[0112] In step S174, the resident estimation unit 14 may further take into consideration the age difference between generations to determine whether to save or discard the random changes. The resident estimation unit 14 may use, as the age difference, at least one of the following: the age difference between a parent and a child, and the age difference between a household head and a parent.

[0113] In steps S173 and S1742, the resident estimation unit 14 calculates the number of children whose age difference with their parents is less than a given threshold based on the household-resident relationship. For example, the resident estimation unit 14 calculates the age difference between the child and the female parent, and for children with only a male parent, calculates the age difference between the two. The resident estimation unit 14 may calculate the difference in age categories listed in the family-age distribution data 230 as the age difference, and calculate the number of children whose category difference is less than 3. In step S1743, the resident estimation unit 14 further determines whether the number of children has decreased from before the random change. If the number of children has decreased and the error has become smaller, the resident estimation unit 14 saves the random change in step S1744, and if the number of children has not decreased, discards the random change in step S1745 regardless of whether the error has become smaller.

[0114] In steps S173 and S1742, the resident estimation unit 14 calculates the number of parents whose age difference from the household head is less than a given threshold based on the household-resident relationship. The resident estimation unit 14 may calculate the difference in age categories described in the family-age distribution data 230 as the age difference, and calculate the number of parents whose category difference is less than 3. In step S1743, the resident estimation unit 14 further determines whether the number of parents has decreased from before the random change. If the number of parents has decreased and the error has become smaller, the resident estimation unit 14 saves the random change in step S1744, and if the number of parents has not decreased, the resident estimation unit 14 discards the random change in step S1745 regardless of whether the error has become smaller.

[0115] In this disclosure, the expression "at least one processor executes a first process, executes a second process, ... executes an nth process" or a corresponding expression is a concept including a case where the executing entity (i.e., the processor) of n processes from the first process to the nth process changes midway. In other words, this expression is a concept including both a case where all n processes are executed by the same processor and a case where the processor changes among the n processes according to an arbitrary policy.

[0116] The processing procedure of the method executed by at least one processor is not limited to the example in the above embodiment. For example, some of the steps (processing) described above may be omitted, or each step may be executed in a different order. In addition, any two or more of the steps described above may be combined, or some of the steps may be modified or deleted. Alternatively, other steps may be executed in addition to each of the steps described above.

[0117] In the present disclosure, when comparing the magnitude relationship of two errors, either of the two criteria "greater than or equal to" and "greater than" may be used, or either of the two criteria "less than or equal to" and "less than" may be used. The selection of such criteria does not change the technical meaning of the process of comparing the magnitude relationship of two numerical values. [Explanation of symbols]

[0118] 1...household estimation system, 2...database group, 11...attribute assignment section, 12...residential area extraction section, 13...household estimation section, 14...resident estimation section, 15...output section, 21...building database, 22...statistical database, 211...detached house data, 212...apartment data, 213...facility data, 221...small area data, 222...person-household distribution data, 223...family-household distribution data (7 categories), 224...resident distribution data, 225...housing-household distribution data, 226...area-household distribution data, 227...family-household distribution data (17 categories), 228...couple age distribution data, 229...household member age distribution data, 230...family-age distribution data.

Claims

1. at least one processor having access to a building database storing building data relating to residential units and a statistical database storing statistical data relating to demographics; the at least one processor: Identifying a plurality of said residential parcels within a target area from said building data; Based on a household distribution indicating the number of households for each household attribute in the target area obtained from the statistical data, the residential section corresponding to the household attribute is determined as a household element; Randomly setting a household-resident relationship between the household element and the resident attribute based on a resident distribution indicating the number of people for each resident attribute in the target area obtained from the statistical data; A replacement process is repeated to randomly change a part of the household-resident relationships based on an error between a given estimated distribution calculated based on the setting of the household-resident relationships and a corresponding reference distribution obtained from the statistical data, thereby determining the household-resident relationships; outputting household data indicative of the household elements including the household attributes and the resident attributes; Household estimation system.

2. the at least one processor executes a mutation process that randomly changes a greater proportion of the household-resident relationships during the repetition of the replacement process based on a history of the errors obtained by the repetition of the replacement process than during the replacement process; The household estimation system according to claim 1 .

3. the at least one processor executes the mutation process when the error after repeating the replacement process a predetermined number of times is equal to or greater than the error before the repetition; The household estimation system according to claim 2 .

4. The household attributes include household members and family type; the at least one processor: A step of randomly setting a block-personnel relationship between the residential block and the household members based on a person-household distribution indicating the number of households for each of the household members in the target area obtained from the statistical data; A step of repeating a process of randomly changing a part of the section-personnel relationship based on an error between a given estimated distribution calculated based on the section-personnel relationship and a corresponding reference distribution obtained from the statistical data, and determining the residential section corresponding to the household member as the household element; A step of randomly setting a household-family relationship between the household elements and the family types based on a family-household distribution indicating the number of households for each family type in the target area obtained from the statistical data; A step of determining the household-family relationships by repeating a process of randomly changing a part of the household-family relationships based on an error between a given estimated distribution calculated based on the household-family relationships and a corresponding reference distribution obtained from the statistical data; to determine the residential section corresponding to the household attribute as the household element; The household estimation system according to any one of claims 1 to 3.

5. accessing a building database storing building data relating to residential zones and identifying a plurality of said residential zones within a region of interest from said building data; A step of accessing a statistical database that stores statistical data on demographics, and determining the residential section corresponding to the household attribute as a household element based on a household distribution indicating the number of households for each household attribute in the target area obtained from the statistical data; A step of randomly setting a household-resident relationship between the household element and the resident attribute based on a resident distribution indicating the number of people for each resident attribute in the target area obtained from the statistical data; A step of determining the household-resident relationships by repeating a replacement process that randomly changes a part of the household-resident relationships based on an error between a given estimated distribution calculated based on the setting of the household-resident relationships and a corresponding reference distribution obtained from the statistical data; outputting household data indicative of the household elements including the household attributes and the resident attributes; A household estimation program that causes a computer to execute the above.

6. at least one processor having access to a building database storing building data relating to residential units and a statistical database storing statistical data relating to demographics; the at least one processor: Identifying a plurality of said residential parcels within a target area from said building data; Randomly set a block-personnel relationship between the residential block and the household members based on a person-household distribution indicating the number of households for each household member in the target area obtained from the statistical data; Based on the error between a given estimated distribution calculated based on the block-personnel relationship and a corresponding reference distribution obtained from the statistical data, a process of randomly changing a part of the block-personnel relationship is repeated to determine the residential block corresponding to the household members as a household element; Randomly setting household-family relationships between the household elements and the family types based on a family-household distribution indicating the number of households for each family type in the target area obtained from the statistical data; Based on an error between a given estimated distribution calculated based on the household-family relationships and a corresponding reference distribution obtained from the statistical data, a process of randomly changing a part of the household-family relationships is repeated to determine the household-family relationships; outputting household data indicating the household elements including the household members and the family type; Household estimation system.

7. accessing a building database storing building data relating to residential zones and identifying a plurality of said residential zones within a region of interest from said building data; A step of accessing a statistical database that stores statistical data on demographics, and randomly setting a block-personnel relationship between the residential block and the household members based on a person-household distribution indicating the number of households per household member in the target area obtained from the statistical data; A step of repeating a process of randomly changing a part of the section-personnel relationship based on an error between a given estimated distribution calculated based on the section-personnel relationship and a corresponding reference distribution obtained from the statistical data, and determining the residential section corresponding to the household members as a household element; A step of randomly setting a household-family relationship between the household elements and the family types based on a family-household distribution indicating the number of households for each family type in the target area obtained from the statistical data; A step of determining the household-family relationships by repeating a process of randomly changing a part of the household-family relationships based on an error between a given estimated distribution calculated based on the household-family relationships and a corresponding reference distribution obtained from the statistical data; outputting household data indicating the household elements including the number of household members and the family type; A household estimation program that causes a computer to execute the above.

Images