Information processing system, information processing method, and information processing program

Abstract

To predict local population changes with high accuracy. To visualize future local population changes and make it easier to make decisions about appropriate downsizing and withdrawal of local businesses. Ultimately, to reduce business operating costs and enable sustainable business management without lowering the standard of living of residents. To apply this to a wide range of public policies and private business areas, such as local government infrastructure planning, urban development, measures against vacant homes, medical demand forecasting, and PPP projects, and to provide scientific forecast information that serves as the basis for implementing infrastructure downsizing measures as a financial measure for local governments in an era of population decline. [Solution] An information processing system equipped with a micro population forecasting unit that forecasts local future population using a mathematical model or machine learning model that calculates the probability of occurrence of each of multiple population change events, generated using resident information data.

Description

[Technical Field]

[0001] The present invention relates to an information processing system, an information processing method, and an information processing program. [Background technology]

[0002] In the above technical field, Patent Document 1 discloses a technology for predicting future population numbers for each mesh, which is created by dividing a target area, based on past statistical population numbers for each mesh. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2012-160143 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the technique described in the above document had low accuracy in predicting local population changes.

[0005] An object of the present invention is to provide a technique for solving the above-mentioned problems. [Means for solving the problem]

[0006] In order to achieve the above object, the system according to the present invention comprises: A macro mortality prediction unit that predicts future mortality over a wide area using a machine learning model for calculating the probability of death on a regional basis, which is generated using basic resident register data; and A micro mortality prediction unit that predicts local future mortality populations using a machine learning model for calculating the probability of death at the individual level, which is generated using resident registration data; and a mortality prediction adjustment unit that compares the wide-area future mortality predicted by the macro mortality prediction unit with the local future mortality predicted by the micro mortality prediction unit and adjusts prediction parameters of the micro mortality prediction unit; a macro birth population forecasting unit that forecasts future birth populations over a wide area using a machine learning model for calculating the probability of births occurring on a regional basis, which is generated using basic resident register data; and a micro birth population forecasting unit that forecasts local future birth populations using a machine learning model for calculating the probability of births on an individual basis, which is generated using basic resident register data; a birth population prediction adjustment unit that compares the wide-area future birth population predicted by the macro birth population prediction unit with the local future birth population predicted by the micro birth population prediction unit, and adjusts prediction parameters of the micro birth population prediction unit; a local future population calculation unit that calculates a local future population by integrating the local future death population and future birth population predicted by the micro death population prediction unit and the micro birth population prediction unit after the prediction parameters have been adjusted by the death population prediction adjustment unit and the birth population prediction adjustment unit; and It is an information processing system equipped with the above.

[0007] In order to achieve the above object, the method according to the present invention comprises: a macro mortality prediction step in which the macro mortality prediction unit predicts future mortality over a wide area using a machine learning model for calculating the probability of death on a regional basis, which model is generated using resident registration data; a micro mortality prediction step in which the micro mortality prediction unit predicts local future mortality using a machine learning model for calculating the probability of death at the residence level, which model is generated using resident registration data; a mortality prediction adjustment step in which a mortality prediction adjustment unit compares the wide-area future mortality predicted by the macro mortality prediction unit with the local future mortality predicted by the micro mortality prediction unit, and adjusts prediction parameters of the micro mortality prediction unit; a macro birth population prediction step in which the macro birth population prediction unit predicts future birth populations over a wide area using a machine learning model for calculating birth probabilities on a regional basis, the model being generated using basic resident register data; a micro birth population prediction step in which the micro birth population prediction unit predicts a local future birth population using a machine learning model for calculating the probability of births occurring in each residential unit, which model is generated using resident basic register data; a birth population forecast adjustment step in which a birth population forecast adjustment unit compares the wide-area future birth population predicted by the macro birth population forecast unit with the local future birth population predicted by the micro birth population forecast unit, and adjusts the forecast parameters of the micro birth population forecast unit; an integration step in which the local future population calculation unit integrates the local future death population and future birth population predicted by the micro death population prediction unit and the micro birth population prediction unit after the prediction parameters have been adjusted by the death population prediction adjustment unit and the birth population prediction adjustment unit, thereby calculating the local future population; The information processing method includes:

[0008] In order to achieve the above object, the program according to the present invention comprises: A macro mortality prediction step that predicts future mortality over a wide area using a machine learning model for calculating the probability of death on a regional basis, which is generated using basic resident register data; A micro-mortality prediction step predicts future local mortality using a machine learning model for calculating the probability of death at the residential level, which is generated using basic resident register data; a mortality prediction adjustment step of comparing the wide-area future mortality predicted in the macro mortality prediction step with the local future mortality predicted in the micro mortality prediction step and adjusting prediction parameters of the micro mortality prediction step; a macro-fertility prediction step that predicts future birth populations over a wide area using a machine learning model for calculating the probability of births occurring on a regional basis, which is generated using basic resident register data; a micro-birth population forecasting step that forecasts local future birth populations using a machine learning model for calculating the probability of births occurring at residential units, which is generated using resident registration data; The wide-area future birth population predicted in the macro birth population prediction step is compared with the local future birth population predicted in the micro birth population prediction step, and the micro birth population prediction is performed. Steps a birth population forecast adjustment step of adjusting the forecast parameters; an integration step of integrating the local future death population and future birth population predicted in the micro death population prediction step and the micro birth population prediction step after adjusting the prediction parameters in the death population prediction adjustment step and the birth population prediction adjustment step to calculate a local future population; It is an information processing program that causes a computer to execute the above. [Effects of the Invention]

[0009] According to the present invention, local population changes can be predicted with high accuracy. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a block diagram showing a configuration of an information processing system according to a first embodiment. [Figure 2] FIG. 10 is a diagram illustrating the significance of the existence of an information processing system according to a second embodiment. [Figure 3] FIG. 10 is a block diagram showing the configuration of an information processing system according to a second embodiment. [Figure 4] FIG. 10 is a block diagram showing the configuration of an information processing system according to a second embodiment. [Figure 5] FIG. 10 is a block diagram showing the configuration of an information processing system according to a second embodiment. [Figure 6] FIG. 10 is a block diagram showing the configuration of an information processing system according to a second embodiment. [Figure 7] FIG. 10 is a block diagram showing the configuration of an information processing system according to a second embodiment. [Figure 8]FIG. 10 is a block diagram showing the configuration of an information processing system according to a second embodiment. [Figure 9] FIG. 10 is a block diagram showing the configuration of an information processing system according to a second embodiment. [Figure 10] FIG. 10 is a diagram showing an example of a display screen of the information processing system according to the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. However, the components described in the following embodiments are merely examples and are not intended to limit the technical scope of the present invention.

[0012] [First embodiment] An information processing system 100 according to a first embodiment of the present invention will be described with reference to Fig. 1. The information processing system 100 is a system that predicts local population fluctuations with high accuracy.

[0013] As shown in FIG. 1, the information processing system 100 includes a micro population prediction unit 101.

[0014] The micro population prediction unit 101 predicts a local future population 115 using mathematical or machine learning models 111, 112 generated using resident information data 110 and calculating occurrence probabilities 113, 114 of multiple types of population change events.

[0015] The resident information data 110 is, for example, but not limited to, basic resident register data. The population change events include at least one of birth, death, migration from a local area, and migration into a local area.

[0016] According to this embodiment, it is possible to predict the local future population with higher accuracy.

[0017] [Second embodiment] Next, an information processing system according to a second embodiment of the present invention will be described with reference to Figures 2 and subsequent figures. Figure 2 is a diagram for explaining the necessity of the information processing system according to this embodiment. Currently, 201, the number of aging equipment (facilities) and facilities with declining users is steadily increasing year by year, but only ad hoc investigations and repairs are carried out in response to accidents or reports from citizens, and although the risks of accidents and declines in public services are recognized, they are not being adequately addressed. This puts pressure on the finances and personnel of local governments and businesses, resulting in a negative spiral.

[0018] This is because various facilities are not being updated or repaired based on objective priorities. No matter how much they are repaired, the work never seems to end. With limited income, expenses only increase. It is impossible to repair all facilities at once, and the local government also faces a heavy workload, dealing with complaints from residents.

[0019] In contrast, Vision 202 uses locally predicted future population to prioritize facilities that need repair and renewal, formulates business plans, and reduces financial burdens and risks. By narrowing down the targets for renewal with an eye to the future, efficient and effective business plans can be realized. The future user population for each facility is predicted from locally predicted future population, and facilities with a high future user population are given priority for renewal over facilities with a low future user population. This effectively reduces complaints from residents and accelerates a positive spiral.

[0020] Conventionally, population forecasting has relied on intermittent, wide-area macro-statistical data, such as the national census, which is released every 5 to 10 years, but this data does not allow for highly accurate local population forecasting. Therefore, in this embodiment, the probability of occurrence of multiple population change events, such as deaths and births, is calculated using machine learning models constructed for each population change event, based on resident registration data held by local governments nationwide. The probability of occurrence of population change events for each individual or household is used to predict future population at the building level.

[0021] For example, in this embodiment, the future population is predicted for each population change event of "death," "migration," "emigration," "relocation," and "birth," and then these are integrated to predict the future population for each dwelling unit.Furthermore, the future population prediction can generate both "dwelling unit" and "small area unit" as outputs.

[0022] 3 is a diagram showing the configuration of an information processing system 300 according to this embodiment. The information processing system 300 is a system that predicts local population fluctuations with high accuracy, and includes a data cleansing unit 301 that cleans basic resident register data 310, a micro population forecasting unit 302, a regional aggregation unit 303, and a business plan generation unit 304.

[0023] The data cleansing unit 301 extracts and generates data necessary for population forecasting using personal identification numbers, individual addresses, dates of birth, gender, moving dates, reasons for moving, household numbers, relationships, and the like contained in the basic resident register data 310. Here, the basic resident register data 310 is used as an example, but any resident information data, such as customer data managed by credit card companies or insurance companies, can also be used. Specifically, the data cleansing unit 301 detects the occurrence of population change events over a certain period of time in the past from the basic resident register data 310, and tallies the number of population change events that occurred for any aggregation unit (year, by region, by building, etc.).

[0024] The micro population forecasting unit 302 uses a mathematical model or machine learning model that calculates the probability of occurrence of each of a plurality of population change events to forecast the number of occurrences of local population change events and further calculate the local future population. Here, the population change events include at least one of births, deaths, migration from a local area, and migration into a local area.

[0025] Specifically, the micro population prediction unit 302 uses the cleansed resident register data as training data to generate a mathematical model or machine learning model that calculates the probability of occurrence of each of multiple population change events.The generated mathematical model or machine learning model is then used to predict the number of future population change events and calculate local (for example, building-by-building) future populations.In this case, individual and household-by-household population attribute information is used as explanatory variables.

[0026] Of the population change events of "death," "migration," "emigration," "relocation," and "birth," death and birth are predicted on an individual basis. In other words, personal data from the Basic Resident Register data is extracted as training data, and the probability of death and birth (childbirth) for each individual is calculated.

[0027] Meanwhile, for "emigration," learning and predictions are performed by dividing the cases into two types, both macro and micro. One is when the entire household has moved out. By learning from past such cases, a machine learning model is generated that calculates the probability that the entire household will move out in the future, and the future number of households moving out is calculated, and then the out-migration population is calculated. The other is when only part of a household has moved out. In this case, by learning from past such cases, a machine learning model is generated that calculates the future probability that each individual in the household will move out, and then the future out-migration population is calculated.

[0028] Meanwhile, learning and predictions for "migration" and "relocation" are also performed by dividing them into two types of cases, macro and micro, but these are separated by whether the destination is originally a vacant residence or not. When someone moves into or moves into a residence that was originally vacant, the entire household is moving, so learning and predictions are performed on a household basis, and the future number of households moving in and moving out is determined, and then the in-migration population and moving out population are calculated. When someone moves into or moves into a residence that was not originally vacant, the entire household is not moving, so learning and predictions are performed on an individual basis, and the future in-migration population and moving out population are predicted. Future in-migration population and moving out population can also be predicted by incorporating the number of vacant homes prediction. The number of vacant homes can also be predicted using the results of death prediction, emigration prediction, and relocation prediction.

[0029] It is also possible to integrate water usage, real estate registration information, building information such as plateaus, and regional attribute data, convert it into a format suitable for machine learning models, and use it as training data or explanatory variables. This will enable spatial distribution predictions (local population predictions) that take into account various geographic information (building information, administrative district boundary information, etc.), water usage, real estate registration information, distance from stations, population density, etc.

[0030] The mathematical model or machine learning model learns and predicts the number of occurrences (probability of occurrence) of population change events from the next fiscal year onwards on an individual or household basis. Gradient boosting decision tree algorithms such as LightGBM and XGBoost may be used as the machine learning model. The training data is basically the past record of changes in the Basic Resident Register. The micro population forecasting unit 302 probabilistically simulates the occurrence of future events for individuals and households and spatially maps them on a map.

[0031] The micro population prediction unit 302 includes an adjustment unit 321 that adjusts prediction parameters of a mathematical model or machine learning model by comparing a wide-area future population predicted based on past wide-area population fluctuations with a local future population predicted from the predicted number of future population change events. In other words, the adjustment unit adjusts the prediction parameters by comparing the predicted number of wide-area population change events with the predicted number of local and future population change events.

[0032] The regional aggregation unit 303 aggregates the local future population predicted by the micro population prediction unit 302 after the adjustment of the prediction parameters by the adjustment unit, and outputs the regional future population for each specified region or area (e.g., basic unit block, etc.).

[0033] The business plan generation unit 304 generates a local business plan based on the regional future population predicted by the micro population forecast unit 302 and compiled by the regional compilation unit 303. Specifically, for existing facilities, the local future population is used to calculate the future number of users. The appropriate cost for that number of users is calculated and the facility size (e.g., water pipe diameter) is determined. For public services, the system determines whether there are few users (poor cost performance) and issues a warning if the cost performance is poor. In response to the warning, local governments and businesses can withdraw, downsize, or consolidate their businesses and facilities. Specifically, the system proposes the appropriate number and location of city hall branch offices, the number and location of schools, libraries, and hospitals, the appropriate diameter of gas and water pipes, and the number and location of substations, all based on the regional future population. Accurate predictions of future infrastructure users enable effective reductions in infrastructure service and facility repair costs as local government financial measures, such as water pipe downsizing and public facility consolidation and closure. It is possible to shift from renewal plans based on the statutory useful life to a planning system that anticipates future demand and deterioration.

[0034] 4 is a diagram showing the detailed configuration of the micro population prediction unit 302. The micro population prediction unit 302 includes a death population prediction unit 401, an out-migration population prediction unit 402, an in-migration population prediction unit 403, a migration population prediction unit 404, a birth population prediction unit 405, and a local future population calculation unit 406.

[0035] The mortality population prediction unit 401 predicts the local future mortality population. The out-migration population prediction unit 402 predicts the local future out-migration population. The in-migration population prediction unit 403 predicts the local future in-migration population. The moving-in population prediction unit 404 predicts the future moving-in population. Moving-in and moving-out refer to changes of address across municipalities, while moving refers to changes of address within the same municipality. The birth population prediction unit 405 predicts the local future birth population. The local future population calculation unit 406 calculates the local future population by subtracting the future death population and future out-migration population from the current local population and adding the future in-migration population and future birth population. A decision is made each time as to whether the future moving-in population is a local subtraction or addition factor. The future moving-in population does not contribute to wide-area population changes.

[0036] Specifically, the local future population calculation unit 406 generates the "future population structure (by age and household structure)" for each residence.

[0037] For example, using the resident register data of a certain local government, the following predictions can be made. - Estimated number of deaths in XX town / district in 2025: X cases ·Predicted number of births in △△ district: Y cases Estimated number of people moving out of XX chome in one year: Z people Future distribution of household size by building Simulation of appropriate design of water pipe diameter based on future water demand estimates It is also possible to calculate local fluctuations in the number of vacant houses using the prediction results from the death population prediction unit 401, the out-migration population prediction unit 402, and the relocation population prediction unit 404.

[0038] Figure 5 is a diagram showing the detailed configuration of the mortality prediction unit 401. The mortality prediction unit 401 predicts the probability of death in each residence for each individual and estimates the mortality. The macro mortality prediction unit 401 includes a city / ward / town / village-level mortality prediction model construction unit 501 and a city / ward / town / village-level mortality prediction unit 502, and the micro mortality prediction unit includes an individual-level mortality prediction model construction unit 503, a local mortality prediction unit 504, and an adjustment unit 505.

[0039] The city / ward / town / village-level mortality prediction model construction unit 501 constructs a city / ward / town / village-level mortality prediction model 521, which is a machine learning model for predicting the number of deaths at the city / ward / town / village level. Here, the city / ward / town / village-level mortality prediction model 521 is a machine learning model, but it may also be a mathematical model. The city / ward / town / village-level mortality prediction unit 502, which serves as a macro population prediction unit that predicts future populations over a wide area, predicts the number of deaths over a wide area using the city / ward / town / village-level mortality prediction model 521. Note that in this embodiment, the wide-area macro population is also predicted using a machine learning model, but the present invention is not limited to this, and values ​​calculated by the National Institute of Population Problems or the like may be used as the macro population.

[0040] The individual-based mortality prediction model construction unit 503 constructs the individual-based mortality prediction model 541, which is a machine learning model for predicting the number of deaths on an individual basis. Here, the individual-based mortality prediction model 541 is a machine learning model, but it may also be a mathematical model. The local mortality prediction unit 504, which serves as a micro population prediction unit that predicts local future populations, predicts the local number of deaths using the individual-based mortality prediction model 541.

[0041] The adjustment unit 505 compares the wide-area future population predicted based on past wide-area population fluctuations with the local future population predicted by the mathematical model or machine learning model, and adjusts the parameters of the mathematical model or machine learning model.

[0042] Specifically, for example, if the predicted future death population for a city, ward, town, or village is 3,000 and the cumulative micro-projected future death population is 3,400, the adjustment unit 505 adjusts the calculation method for the local future death population so that the cumulative micro-projected future death population is 3,000. The adjustment unit 505 adjusts the threshold for the probability of death occurrence so that a death flag is assigned (counted in the predicted number of deaths) to each individual, starting with the local area (individual) with the highest probability of death calculated by the individual-level death prediction model 541, until the number of population change events (the number of deaths at the city, ward, town, or village level) is reached, which is the macro-projection. For example, if the macro-projection number of deaths at the city, ward, town, or village level is 100, the classification probability, i.e., the number of deaths with a probability of death between 100% and 50% is 99, and the number of deaths with a probability of death between 100% and 49% is 100, the threshold is set to 49% (0 for deaths below 49%). This enables consistency with the number of population change events at the administrative district level under the total volume constraint.

[0043] The local future mortality output from the local mortality prediction unit 504 may be used to predict the local mortality for even further into the future. For example, the mortality for next year may be predicted, and the predicted mortality for next year may be used to predict the mortality for the year after next. By repeatedly predicting the mortality for a certain period in the future in this way, it is possible to accurately predict the mortality for a long time in the future, such as 20 years from now.

[0044] In constructing the individual mortality prediction model and local mortality prediction, water usage data may be used as an explanatory variable. For example, if water usage drops significantly, the probability of death may be estimated to be higher.

[0045] 6 is a block diagram showing the detailed configuration of the out-migration population prediction unit 402. The out-migration population prediction unit 402 constructs models separately for cases where all household members move out and cases where an individual moves out of a household, and predicts the macro-out-migration population and the micro-out-migration population.

[0046] The out-migration population prediction unit 402 includes a macro out-migration population prediction unit, which includes a city / ward / town / village-level out-migration household number prediction model construction unit 601, a city / ward / town / village-level out-migration household number prediction unit 602, a city / ward / town / village-level out-migration population prediction model construction unit 603, and a city / ward / town / village-level out-migration population prediction unit 604.

[0047] 5. The out-migration population prediction unit 402 also includes a micro out-migration population prediction unit, which includes a household-based out-migration household number prediction model construction unit 605, a household-based out-migration household number prediction unit 606, an individual-based out-migration population prediction model construction unit 607, and an individual-based out-migration population prediction unit 608. The out-migration population prediction unit 402 also includes an adjustment unit 609 having the same function as the adjustment unit 505 in FIG.

[0048] The city / ward / town / village-level household outmigration prediction model construction unit 601 extracts the number of city / ward / town / village-level household outmigration in the event that all household members move out from the cleansed resident register data, and uses this as training data to construct a city / ward / town / village-level household outmigration prediction model 621 for predicting the number of city / ward / town / village-level household outmigration in the event that all household members move out in the future.

[0049] The city / ward / town / village out-migration population prediction model construction unit 603 extracts the city / ward / town / village out-migration population when some household members move out from the cleansed resident register data, and uses this as training data to construct a city / ward / town / village out-migration population prediction model 641 for predicting the city / ward / town / village out-migration population when some household members move out in the future.

[0050] The household-based outmigration household number prediction model construction unit 605 extracts the number of households that will outmigrate on a household basis when all household members move out from the cleansed resident register data, and uses this as training data to construct a household-based outmigration household number prediction model 661 for predicting the number of households that will outmigrate on a household basis when all household members move out in the future.

[0051] The individual-level out-migration population prediction model construction unit 607 extracts the individual-level out-migration population when some household members move out from the cleansed resident register data, and uses this as training data to construct an individual-level out-migration population prediction model 681 for predicting the individual-level out-migration population when some household members move out in the future.

[0052] The city / ward / town / village-level out-migration household number prediction unit 602 estimates that, for example, for households consisting of only elderly couples and family households that do not include residents aged 18-25, all household members will move out, and predicts the number of out-migration households over a wide area using the city / ward / town / village-level out-migration household number prediction model 621.

[0053] The city / ward / town / village-level out-migration population prediction unit 604 estimates that some household members will move out, for example, for households that do not consist of only elderly couples and family households that include residents aged 18-25, and predicts the out-migration population over a wide area using the city / ward / town / village-level out-migration population prediction model 641.

[0054] The household-based out-migration household number prediction unit 606 estimates that, for example, for households consisting of only elderly couples and family households that do not include residents aged 18-25, all household members will move out, and predicts the local number of out-migration households using the household-based out-migration household number prediction model 661.

[0055] The individual-based outmigration population prediction unit 608 estimates that some household members will move out, for example, for households that do not consist of only elderly couples and family households that include residents aged 18-25, and predicts the local outmigration population using the individual-based outmigration population prediction model 681.

[0056] The local future out-migration population 610 can be calculated by multiplying the local number of out-migration households predicted by the household-level out-migration household number prediction unit 606 by the number of households, and then adding the local out-migration population predicted by the individual-level out-migration population prediction unit 608.

[0057] The adjustment unit 609 adjusts the prediction parameters of the household-based out-migration household number prediction unit 606 using the predictions of the city / ward / town / village-based out-migration household number prediction unit 602. The adjustment unit 609 also adjusts the prediction parameters of the individual-based out-migration population prediction unit 608 using the predictions of the city / ward / town / village-based out-migration population prediction unit 604.

[0058] In other words, the threshold value of the probability of out-migration is adjusted so that an out-migration flag is assigned to each household in descending order of the probability of out-migration calculated by the household-level out-migration household number prediction unit 606 until the total number of out-migration households reaches a certain number of out-migration households at the municipality level predicted by the municipality-level out-migration household number prediction unit 602. This makes it possible to match the number of population change events at the administrative district level under the total volume constraint.

[0059] The predicted future out-migration population 610 may be used to predict the future out-migration population for even further into the future. For example, the out-migration population for next year may be predicted, and the predicted out-migration population for next year may be used to predict the out-migration population for the year after next. By repeatedly predicting the out-migration population for a certain period of time in this way, it is possible to accurately predict the out-migration population for a long time in the future, such as 20 years from now.

[0060] 7 is a block diagram showing the detailed configuration of the in-migration population prediction unit 403. The in-migration population prediction unit 403 constructs models separately for cases where all household members move in and cases where an individual moves into an existing household, and predicts the macro-in-migration population and the micro-in-migration population.

[0061] The in-migration population prediction unit 403 includes a macro in-migration population prediction unit, which includes a city / ward / town / village-level in-migration household number prediction model construction unit 701, a city / ward / town / village-level in-migration household number prediction unit 702, a city / ward / town / village-level in-migration population prediction model construction unit 703, and a city / ward / town / village-level in-migration population prediction unit 704.

[0062] 5. The in-migration population prediction unit 403 also includes a micro-in-migration population prediction unit, which includes a household-level in-migration household number prediction model construction unit 705, a household-level in-migration household number prediction unit 706, an individual-level in-migration population prediction model construction unit 707, and an individual-level in-migration population prediction unit 708. The in-migration population prediction unit 403 also includes an adjustment unit 709 having the same function as the adjustment unit 505 in FIG.

[0063] The city / ward / town / village-level household in-migration prediction model construction unit 701 learns the number of households in-migrating to a city / ward / town / village when all household members move in from the cleansed resident register data as training data, and then constructs a city / ward / town / village-level household in-migration prediction model 721 for predicting the number of households in-migrating to a city / ward / town / village when all household members move in in the future.

[0064] The city / ward / town / village-level in-migration population prediction model construction unit 703 extracts and learns the in-migration population at city / ward / town / village level when some household members move in from the cleansed resident register data, and then constructs a city / ward / town / village-level in-migration population prediction model 741 for predicting the in-migration population at city / ward / town / village level when some household members move in in the future.

[0065] The household-level in-migration household number prediction model construction unit 705 extracts and learns the number of households moving in on a household-level basis when all household members move in from the cleansed resident register data. Then, it constructs a household-level in-migration household number prediction model 761 for predicting the number of households moving in on a household-level basis when all household members move in in the future.

[0066] The individual-level in-migration population prediction model construction unit 707 extracts and learns the individual-level in-migration population when some household members move in from the cleansed resident register data, and then constructs an individual-level in-migration population prediction model 781 for predicting the individual-level in-migration population when some household members move in in the future.

[0067] For example, when the new residence is a vacant house, the city / ward / town / village unit household in-migration number prediction unit 702 estimates that all household members will move in (a new household will be created), and predicts the number of households in-migration over a wide area using city / ward / town / village unit household in-migration number prediction model 721. The number of households in-migration over a wide area may also be predicted by time series analysis of the Basic Resident Register.

[0068] For example, if the new residence is not vacant, the city / ward / town / village-level in-migration population prediction unit 704 estimates that some household members will move in (no new household will be created, but the number of household members will increase), and predicts the in-migration population over a wide area using the city / ward / town / village-level in-migration population prediction model 741.

[0069] For example, when the new residence is a vacant house, the household-based moving-in household number prediction unit 706 estimates that all household members will move in, and predicts the local number of moving-in households using the household-based moving-in household number prediction model 761.

[0070] For example, if the new residence is not vacant, the individual-level in-migration population prediction unit 708 estimates that some household members will move in, and predicts the local in-migration population using the individual-level in-migration population prediction model 781.

[0071] The local future in-migration population 710 can be calculated by multiplying the local number of in-migration households predicted by the household-level in-migration household number prediction unit 706 by the number of household members, and then adding the local in-migration population predicted by the individual-level in-migration population prediction unit 708.

[0072] The adjustment unit 709 adjusts the prediction parameters of the household-based in-migration number prediction unit 706 using the predictions made by the city / ward / town / village-based in-migration number prediction unit 702. The adjustment unit 709 also adjusts the prediction parameters of the individual-based in-migration population prediction unit 708 using the predictions made by the city / ward / town / village-based in-migration population prediction unit 704.

[0073] In other words, the threshold for the probability of in-migration is adjusted so that an in-migration occurrence flag is assigned to each household in order of the probability of in-migration calculated by the household-level in-migration household number prediction unit 706, until the total number of in-migration households reaches a certain number of in-migration households at the municipality level predicted by the municipality-level in-migration household number prediction unit 702. This makes it possible to match the number of population change events at the administrative ward level under the total amount constraint.

[0074] The predicted future in-migration population 710 may be used to predict future in-migration populations in the future. For example, the in-migration population for next year may be predicted, and the predicted in-migration population for next year may be used to predict the in-migration population for the year after next. By repeatedly predicting the in-migration population for a certain period of time in this way, it is possible to accurately predict the in-migration population for a long time in the future, such as 20 years from now.

[0075] The in-migration population prediction unit 403 may predict the probability and number of people moving in for each building or area, according to patterns such as singles, marriages, nuclear families, etc. Specifically, it may predict the future number of wide-area marriages using the past number of wide-area marriages, and estimate that at least one of household movements to multiple vacant houses with predetermined attributes and movements from existing one-person households to two-person households will occur in a number corresponding to the predicted number of wide-area marriages.

[0076] 8 is a block diagram showing the detailed configuration of the moving-in population prediction unit 404. The moving-in population prediction unit 404 constructs machine learning models and predicts the macro-moving population and micro-moving population, distinguishing between cases where all household members move into vacant houses, resulting in the disappearance of a household and the creation of a new household, and cases where individuals move into existing households without moving into vacant houses.

[0077] The moving population prediction unit 404 includes, as a macro moving population prediction unit, a city / ward / town / village unit disappearance / new household number prediction model construction unit 801, a city / ward / town / village unit disappearance / new household number prediction unit 802, a city / ward / town / village unit disappearance / new population prediction model construction unit 803, and a city / ward / town / village unit disappearance / new population prediction unit 804.

[0078] The city / ward / town / village unit number of disappearing / new households prediction unit 802 assumes that, for example, if the new residence is a vacant house, all household members will move (a household will disappear and a new household will be created at the same time), and predicts the number of disappearing and new households in a wide area using a city / ward / town / village unit number of disappearing households prediction model and a city / ward / town / village unit number of new households prediction model (referred to as city / ward / town / village unit number of disappearing / new households prediction model) 821. The number of disappearing and new households in a wide area may also be predicted by time series analysis of the Basic Resident Register.

[0079] The city / ward / town / village-level disappearance / new population prediction unit 804 estimates that, for example, if the new residence is not vacant, some of the household members will move (no new household will be created, but the number of household members will increase), and predicts the relocating population over a wide area using the city / ward / town / village-level disappearance / new population prediction model 841.

[0080] 5. The moving population prediction unit 404 also includes, as a micro moving population prediction unit, a household unit disappearance / new household number prediction model construction unit 805, a household unit disappearance / new household number prediction unit 806, an individual unit disappearance / new population prediction model construction unit 807, and an individual unit disappearance / new population prediction unit 808. The moving population prediction unit 404 also includes an adjustment unit 809 having the same function as the adjustment unit 505 in FIG.

[0081] The city / ward / town / village-level number of households disappearing / emerging prediction model construction unit 801 extracts the past number of households disappearing / emerging at city / ward / town / village level when people move into vacant houses from the cleansed basic resident register data. Then, using the past number of households disappearing / emerging at city / ward / town / village level as training data, it constructs a city / ward / town / village-level number of households disappearing / emerging prediction model 821 for predicting the future number of households disappearing / emerging at city / ward / town / village level.

[0082] The city, ward, town, and village level population disappearance / new population prediction model construction unit 803 constructs a city, ward, town, and village level population disappearance / new population prediction model 841 for predicting future city, ward, town, and village level disappearance / new population using the city, ward, town, and village level disappearance / new population in the case where people do not move into vacant houses in the resident register data as training data.

[0083] The household unit disappearance / emergence number prediction model construction unit 805 extracts the number of households that will disappear / emerge on a household unit basis when a person moves into a vacant house from the cleansed resident register data, and constructs a household unit disappearance / emergence number prediction model 861 for predicting the number of households that will disappear / emerge on a household unit basis when all household members move in the future.

[0084] The household unit extinction / new population prediction model construction unit 807 extracts the household unit extinction / new population in the case where people do not move to vacant houses from the cleansed resident register data, and constructs an individual unit extinction / new population prediction model 881 for predicting the individual unit extinction / new population in the case where some household members move in the future.

[0085] For example, when the new residence is a vacant house, the city / ward / town / village unit household number prediction unit 802 presumes that all household members will move, and predicts the number of households that will disappear / appear in a wide area using city / ward / town / village unit household number prediction model 821. The number of households that will disappear / appear in a wide area may also be predicted by time series analysis of the basic resident register.

[0086] The city / ward / town / village-level disappearance / new population prediction unit 804 estimates that some household members will move if the new residence is not vacant, and predicts the disappearance / new population over a wide area using the city / ward / town / village-level disappearance / new population prediction model 841.

[0087] The household unit disappearance / creation number prediction unit 806 estimates that, for example, if the new residence is a vacant house, all household members will move, and predicts the local number of disappearance / creation households using the household unit disappearance / creation number prediction model 861.

[0088] For example, if the new residence is not a vacant house, the individual-based extinction / new population prediction unit 808 estimates that some of the household members will move, and predicts the local extinction / new population using the individual-based extinction / new population prediction model 881.

[0089] The local future relocation population 810 can be calculated by multiplying the number of local disappearance / formation households predicted by the household unit disappearance / formation household number prediction unit 806 by the number of household members, and then adding the local disappearance / formation population predicted by the individual unit disappearance / formation population prediction unit 808.

[0090] The adjustment unit 809 adjusts the prediction parameters of the household-based disappearance / new population prediction unit 806 using the predictions of the city / ward / town / village-based disappearance / new household number prediction unit 802. The adjustment unit 809 also adjusts the prediction parameters of the individual-based disappearance / new population prediction unit 808 using the predictions of the city / ward / town / village-based disappearance / new population prediction unit 804.

[0091] Specifically, the threshold value of the probability of relocation is adjusted so that a relocation flag is assigned to each household in descending order of the probability of relocation calculated by the household unit number of disappearance / formation prediction unit 806 until the total number of disappearance / formation households reaches a certain number of disappearance / formation households at the city / ward / town / village unit predicted by the city / ward / town / village unit number of disappearance / formation prediction unit 802. This makes it possible to match the number of population change events at the administrative ward unit under the total amount constraint.

[0092] The predicted future population change 810 may be used to predict the population change for an even further future period. For example, the population change for next year may be predicted, and the predicted population change for next year may be used to predict the population change for the year after next. By repeatedly predicting the population change for a certain period in the future in this way, it is possible to accurately predict the population change for a long time in the future, such as 20 years from now.

[0093] The moving population prediction unit 404 may predict the probability and number of people moving for each building or area, according to patterns such as single people, marriage, nuclear families, etc. Specifically, the number of future wide-area marriages may be predicted using the number of past wide-area marriages, and at least one of household movements to multiple vacant houses with predetermined attributes and movements from existing one-person households to two-person households may be estimated for a number corresponding to the predicted number of wide-area marriages.

[0094] Moving data can be extracted using the date of change, reason for change, address, and PLATEAU from the Basic Resident Register. The address of a resident who moves will become a vacant house. Macro predictions are also made for moving, and the number of moves across a wide area is calculated, but moving is not a factor that causes population change on a wide scale. By generating micro moving events based on the predicted number of moves across a wide area, local population changes can be predicted more accurately.

[0095] 9 is a diagram showing the detailed configuration of the birth population prediction unit 405. The birth population prediction unit 405 predicts the probability of births in each residence for each individual and estimates the birth population. The birth population prediction unit 405 includes a city / ward / town / village-level birth prediction model construction unit 901 and a city / ward / town / village-level birth prediction unit 902 as a macro birth population prediction unit, and an individual-level birth prediction model construction unit 903, a local birth prediction unit 904, and an adjustment unit 905 as a micro birth population prediction unit.

[0096] The city / ward / town / village-level birth prediction model construction unit 901 constructs a city / ward / town / village-level birth prediction model 921, which is a machine learning model for predicting the number of births at the city / ward / town / village level. Here, the city / ward / town / village-level birth prediction model 921 is a machine learning model, but it may also be a mathematical model. The city / ward / town / village-level birth prediction unit 902, which serves as a macro population prediction unit that predicts future population over a wide area, predicts the number of births over a wide area using the city / ward / town / village-level birth prediction model 921.

[0097] The individual-based birth prediction model construction unit 903 constructs an individual-based birth prediction model 941, which is a machine learning model for predicting the number of births on an individual basis. Here, the individual-based birth prediction model 941 is a machine learning model, but it may also be a mathematical model. The local birth prediction unit 904, which serves as a micro population prediction unit that predicts local future populations, predicts the local number of births using the individual-based birth prediction model 941.

[0098] The adjustment unit 905 compares the wide-area future population predicted based on past wide-area population fluctuations with the local future population predicted by the mathematical model or machine learning model, and adjusts the parameters of the mathematical model or machine learning model.

[0099] Specifically, the adjustment unit 905 adjusts the threshold value of the probability of birth so that a birth flag is assigned to each individual, starting with the individual with the highest probability of birth calculated by the individual-based birth prediction model 941, until the number of births reaches the macro-predicted number of births at the municipality level. This makes it possible to align the number of population change events at the administrative district level under the total quantity constraint.

[0100] The local future birth population output from the local birth prediction unit 904 may be used to predict the local birth population further into the future. For example, the birth population for next year may be predicted, and the predicted birth population for next year may be used to predict the birth population for the year after next. By repeatedly predicting the birth population for a certain period in the future in this way, it is possible to accurately predict the birth population for a long future period, such as 20 years from now.

[0101] The future birth prediction unit extracts addresses of married households with women in the childbearing age group and calculates the birth probability according to the number of years since marriage. For example, for a 25-year-old married couple with no children, the probability of birth within one year of marriage is very high (e.g., 80%), and the probability decreases as age and the number of children already present increase. Until the macro-prediction of births at the city / ward / town / village level is reached, a birth occurrence flag is assigned to individuals (women) with the highest probability of birth (i.e., they are counted in the predicted number of births).

[0102] Thus, predicting the number of marriages is necessary to accurately predict the number of births. In this embodiment, the marriage prediction unit 906 predicts the number of marriages by considering the probability of a pattern in which a single-person household at an existing address evolves into a two-person household and a pattern in which a two-person household emerges at a new address (moving in or relocating). From the number of households that change from a single-person household to a two-person household and the number of households that emerge at a new address (moving in or relocating), the number of households that are likely to have married over a wide area is estimated using past probabilities and ages, and allocated to a local area. In this case, a large number of marriages may be allocated to vacant houses in areas with a high rate of migration due to past marriages, and a small number of marriages may be allocated to vacant houses in areas with a low rate of such migration. The predicted married households may be randomly allocated to each vacant house, or the marriage suitability of each vacant house may be evaluated, and future predicted married households may be allocated in descending order of marriage suitability.

[0103] Marriages can also be analyzed by individual attribute. For example, in the Basic Resident Register data, in addition to the number of people who are clearly registered as moving due to marriage, the number of people whose individual attributes change to head of household and spouse can be counted as the number of marriages. The number of marriages is counted by household.

[0104] 10 is a diagram showing an example of a display screen 1000 displayed by the business plan generation unit 304 of the information processing system 300 according to this embodiment. For example, when a postal code 1001 and a future year are entered on the display screen 1000, the business plan generation unit 304 displays a map of the area assigned with that postal code and displays the population of each building predicted by the micro population prediction unit 302 using shading and numbers. Here, an input field for the postal code is used, but a screen that allows direct input of an address may also be used.

[0105] When the user selects a project from the list displayed on the left, the project plan generation unit 304 aggregates the local future user population for the selected project and displays the project's future income and expenditure forecast and the project plan to be promoted. In Figure 10, a water project, i.e., a check mark 1011 for water pipe diameter, is selected as an example, and a comparison display 1012 of predicted future water usage and maintenance costs for each water pipe route is displayed. For routes where profitability is expected to decline, a cross 1013 is displayed overlaid. Routes where profitability is expected to decline and routes where it is not may be displayed in different colors. Using such a screen makes it possible to prioritize pipeline renewal and make decisions about pipeline withdrawal or reduction.

[0106] Note that local business plans are not limited to water pipe routes. For example, in Figure 10, it is possible to check gas pipe diameters, schools, libraries, hospitals, etc. on the same screen. For schools, libraries, hospitals, etc., by predicting changes in the number of users, decisions such as consolidation and closure can be made quickly.

[0107] According to the above embodiment, future local population changes can be visualized. This allows appropriate decisions to be made regarding downsizing and withdrawal of businesses. This in turn allows for reductions in business operating costs and sustainable business management without lowering the standard of living of residents.

[0108] This embodiment can be applied to a wide range of public policies and private business areas, such as local government infrastructure planning, urban development, measures against vacant houses, medical demand forecasting, PPP projects, etc. In particular, it can provide scientific forecast information that serves as the basis for implementing infrastructure downsizing measures as a financial measure for local governments in an era of population decline.

[0109] [Other embodiments] Although the present invention has been described above with reference to the embodiments, the present invention is not limited to the above embodiments. Various modifications can be made to the configuration and details of the present invention that are understandable to those skilled in the art within the technical scope of the present invention. Furthermore, systems or devices that combine the separate features included in each embodiment in any manner are also included in the technical scope of the present invention.

[0110] The present invention may also be applied to a system consisting of multiple devices or to a single device. Furthermore, the present invention may also be applied when an information processing program that realizes the functions of the embodiments is supplied to a system or device and executed by a built-in processor. The technical scope of the present invention also includes a program installed on a computer to realize the functions of the present invention, a medium storing the program, a server from which the program is downloaded, and a processor that executes the program. In particular, the technical scope of the present invention includes at least a non-transitory computer-readable medium storing a program that causes a computer to execute the processing steps included in the above-described embodiments.

[0111] [Other expressions of embodiments] Some or all of the above-described embodiments can be described as, but are not limited to, the following supplementary notes. (Appendix 1) A macro mortality prediction unit that predicts future mortality over a wide area using a machine learning model for calculating the probability of death on a regional basis, which is generated using basic resident register data; and A micro mortality prediction unit that predicts local future mortality populations using a machine learning model for calculating the probability of death at the individual level, which is generated using resident registration data; and a mortality prediction adjustment unit that compares the wide-area future mortality predicted by the macro mortality prediction unit with the local future mortality predicted by the micro mortality prediction unit and adjusts prediction parameters of the micro mortality prediction unit; a macro birth population forecasting unit that forecasts future birth populations over a wide area using a machine learning model for calculating the probability of births occurring on a regional basis, which is generated using basic resident register data; and a micro birth population forecasting unit that forecasts local future birth populations using a machine learning model for calculating the probability of births on an individual basis, which is generated using basic resident register data; a birth population prediction adjustment unit that compares the wide-area future birth population predicted by the macro birth population prediction unit with the local future birth population predicted by the micro birth population prediction unit, and adjusts prediction parameters of the micro birth population prediction unit; a local future population calculation unit that calculates a local future population by integrating the local future death population and future birth population predicted by the micro death population prediction unit and the micro birth population prediction unit after the prediction parameters have been adjusted by the death population prediction adjustment unit and the birth population prediction adjustment unit; and An information processing system comprising: (Appendix 2) The information processing system described in Appendix 1, wherein the micro birth population prediction unit predicts the future number of wide-area marriages using the past number of wide-area marriages, and generates at least one of movements to multiple vacant houses with specified attributes and changes from existing one-person households to two-person households, a number corresponding to the predicted number of wide-area marriages, thereby predicting the future number of births. (Appendix 3) a macro-outmigration population forecasting unit that forecasts future outmigration populations over a wide area using a machine learning model for calculating the probability of outmigration on a regional basis, which model is generated using basic resident register data; a micro-outmigration population prediction unit that predicts future local outmigration populations using a machine learning model for calculating the probability of outmigration at the residence level, which is generated using resident registration data; an out-migration population prediction adjustment unit that compares the wide-area future out-migration population predicted by the macro out-migration population prediction unit with the local future out-migration population predicted by the micro out-migration population prediction unit and adjusts prediction parameters of the micro out-migration population prediction unit; Furthermore, 3. The information processing system according to claim 1, wherein the local future population calculation unit calculates the local future population by further integrating the local future out-migration population predicted by the micro out-migration population prediction unit after the out-migration population prediction adjustment unit adjusts the prediction parameters. (Appendix 4) a macro-inflow population forecasting unit that forecasts future inflow populations over a wide area using a machine learning model for calculating the probability of inflows occurring on a regional basis, which is generated using basic resident register data; and a micro-inflow population forecasting unit that forecasts future local inflow populations using a machine learning model for calculating the probability of inflows at the residential level, which is generated using basic resident register data; an in-migration population prediction adjustment unit that compares the wide-area future in-migration population predicted by the macro in-migration population prediction unit with the local future in-migration population predicted by the micro in-migration population prediction unit and adjusts prediction parameters of the micro in-migration population prediction unit; Furthermore, 3. The information processing system according to claim 1, wherein the local future population calculation unit calculates the local future population by further integrating the local future in-migration population predicted by the micro in-migration population prediction unit after the in-migration population prediction adjustment unit adjusts the prediction parameters. (Appendix 5) a macro-relocation population prediction unit that predicts the future relocation population over a wide area using a first machine learning model for calculating the probability of relocation on a regional basis, the first machine learning model being generated using resident registration data; and a micro-relocation population prediction unit that predicts the local future relocation population using a second machine learning model for calculating the probability of relocation on a residential basis, the second machine learning model being generated using the resident registration data; a moving-in population prediction adjustment unit that compares the wide-area future moving-in population predicted by the macro moving-in population prediction unit with the local future moving-in population predicted by the micro moving-in population prediction unit and adjusts prediction parameters of the micro moving-in population prediction unit; Furthermore, An information processing system according to claim 1 or 2, wherein the local future population calculation unit calculates the local future population by further integrating the local future relocation population predicted by the micro relocation population prediction unit after the relocation population prediction adjustment unit adjusts the prediction parameters. (Appendix 6) 2. The information processing system according to claim 1, further comprising a business plan generation unit that generates a local business plan based on the local future population calculated by the local future population calculation unit. (Appendix 7) a macro mortality prediction step in which the macro mortality prediction unit predicts future mortality over a wide area using a machine learning model for calculating the probability of death on a regional basis, which model is generated using resident registration data; a micro mortality prediction step in which the micro mortality prediction unit predicts local future mortality using a machine learning model for calculating the probability of death at the residence level, which model is generated using resident registration data; a mortality prediction adjustment step in which a mortality prediction adjustment unit compares the wide-area future mortality predicted by the macro mortality prediction unit with the local future mortality predicted by the micro mortality prediction unit, and adjusts prediction parameters of the micro mortality prediction unit; a macro birth population prediction step in which the macro birth population prediction unit predicts future birth populations over a wide area using a machine learning model for calculating birth probabilities on a regional basis, the model being generated using basic resident register data; a micro birth population prediction step in which the micro birth population prediction unit predicts a local future birth population using a machine learning model for calculating the probability of births occurring in each residential unit, which model is generated using resident basic register data; a birth population forecast adjustment step in which a birth population forecast adjustment unit compares the wide-area future birth population predicted by the macro birth population forecast unit with the local future birth population predicted by the micro birth population forecast unit, and adjusts the forecast parameters of the micro birth population forecast unit; an integration step in which the local future population calculation unit integrates the local future death population and future birth population predicted by the micro death population prediction unit and the micro birth population prediction unit after the prediction parameters have been adjusted by the death population prediction adjustment unit and the birth population prediction adjustment unit, thereby calculating the local future population; An information processing method including: (Appendix 8) A macro mortality prediction step that predicts future mortality over a wide area using a machine learning model for calculating the probability of death on a regional basis, which is generated using basic resident register data; A micro-mortality prediction step predicts future local mortality using a machine learning model for calculating the probability of death at the residential level, which is generated using basic resident register data; a mortality prediction adjustment step of comparing the wide-area future mortality predicted in the macro mortality prediction step with the local future mortality predicted in the micro mortality prediction step and adjusting prediction parameters of the micro mortality prediction step; a macro-fertility prediction step that predicts future birth populations over a wide area using a machine learning model for calculating the probability of births occurring on a regional basis, which is generated using basic resident register data; a micro-birth population forecasting step that forecasts local future birth populations using a machine learning model for calculating the probability of births occurring at residential units, which is generated using resident registration data; a birth population prediction adjustment step of comparing the wide-area future birth population predicted in the macro birth population prediction step with the local future birth population predicted in the micro birth population prediction step, and adjusting prediction parameters of the micro birth population prediction unit; an integration step of integrating the local future death population and future birth population predicted in the micro death population prediction step and the micro birth population prediction step after adjusting the prediction parameters in the death population prediction adjustment step and the birth population prediction adjustment step to calculate a local future population; An information processing program that causes a computer to execute the above. (Appendix 9) a micro-prediction unit that predicts the number of local and future population change events using a machine learning model for calculating the probability of occurrence of population change events at the residence level, which is generated using resident registration data; an adjustment unit that compares the predicted occurrence number of the wide-area population change event with the occurrence number of local and future population change events predicted by the micro prediction unit, and adjusts prediction parameters of the micro prediction unit; a local future population calculation unit that calculates a local future population using the number of occurrences of local population change events predicted by the micro prediction unit after the prediction parameters are adjusted by the adjustment unit; and An information processing system comprising:

Claims

1. A macro mortality prediction unit that predicts future mortality over a wide area using a machine learning model for calculating the probability of death on a regional basis, which is generated using basic resident register data; and A micro mortality prediction unit that predicts local future mortality populations using a machine learning model for calculating the probability of death at the individual level, which is generated using resident registration data; and a mortality prediction adjustment unit that compares the wide-area future mortality predicted by the macro mortality prediction unit with the local future mortality predicted by the micro mortality prediction unit and adjusts prediction parameters of the micro mortality prediction unit; a macro birth population forecasting unit that forecasts future birth populations over a wide area using a machine learning model for calculating the probability of births occurring on a regional basis, which is generated using basic resident register data; and a micro birth population forecasting unit that forecasts local future birth populations using a machine learning model for calculating the probability of births on an individual basis, which is generated using basic resident register data; a birth population prediction adjustment unit that compares the wide-area future birth population predicted by the macro birth population prediction unit with the local future birth population predicted by the micro birth population prediction unit, and adjusts prediction parameters of the micro birth population prediction unit; a local future population calculation unit that calculates a local future population by integrating the local future death population and future birth population predicted by the micro death population prediction unit and the micro birth population prediction unit after the prediction parameters have been adjusted by the death population prediction adjustment unit and the birth population prediction adjustment unit; and An information processing system comprising:

2. 2. The information processing system according to claim 1, wherein the micro birth population prediction unit predicts the future number of wide-area marriages using the past number of wide-area marriages, and generates at least one of a number of movements to a plurality of vacant houses having predetermined attributes and a number of changes from existing one-person households to two-person households corresponding to the predicted number of wide-area marriages, thereby predicting the future number of births.

3. a macro-outmigration population forecasting unit that forecasts future outmigration populations over a wide area using a machine learning model for calculating the probability of outmigration on a regional basis, which model is generated using basic resident register data; a micro-outmigration population prediction unit that predicts future local outmigration populations using a machine learning model for calculating the probability of outmigration at the residence level, which is generated using resident registration data; an out-migration population prediction adjustment unit that compares the wide-area future out-migration population predicted by the macro out-migration population prediction unit with the local future out-migration population predicted by the micro out-migration population prediction unit and adjusts prediction parameters of the micro out-migration population prediction unit; Furthermore, 3. The information processing system according to claim 1, wherein the local future population calculation unit calculates the local future population by further integrating the local future out-migration population predicted by the micro out-migration population prediction unit after the out-migration population prediction adjustment unit adjusts the prediction parameters.

4. a macro-inflow population forecasting unit that forecasts future inflow populations over a wide area using a machine learning model for calculating the probability of inflows occurring on a regional basis, which is generated using basic resident register data; and a micro-inflow population forecasting unit that forecasts future local inflow populations using a machine learning model for calculating the probability of inflows at the residential level, which is generated using basic resident register data; an in-migration population prediction adjustment unit that compares the wide-area future in-migration population predicted by the macro in-migration population prediction unit with the local future in-migration population predicted by the micro in-migration population prediction unit and adjusts prediction parameters of the micro in-migration population prediction unit; Furthermore, 3. The information processing system according to claim 1, wherein the local future population calculation unit calculates the local future population by further integrating the local future in-migration population predicted by the micro in-migration population prediction unit after the in-migration population prediction adjustment unit adjusts the prediction parameters.

5. a macro-relocation population prediction unit that predicts a wide-area future relocation population using a first machine learning model for calculating the probability of relocation on a regional basis, the first machine learning model being generated using resident registration data; a micro-relocation population prediction unit that predicts a local future relocation population using a second machine learning model for calculating the probability of relocation on a residential basis, the second machine learning model being generated using the resident registration data; a moving-in population prediction adjustment unit that compares the wide-area future moving-in population predicted by the macro moving-in population prediction unit with the local future moving-in population predicted by the micro moving-in population prediction unit and adjusts prediction parameters of the micro moving-in population prediction unit; Furthermore, 3. The information processing system according to claim 1, wherein the local future population calculation unit calculates the local future population by further integrating the local future relocation population predicted by the micro relocation population prediction unit after the relocation population prediction adjustment unit adjusts the prediction parameters.

6. 2. The information processing system according to claim 1, further comprising a business plan generation unit that generates a local business plan based on the local future population calculated by the local future population calculation unit.

7. a macro mortality prediction step in which the macro mortality prediction unit predicts future mortality over a wide area using a machine learning model for calculating the probability of death on a regional basis, which model is generated using resident registration data; a micro mortality prediction step in which the micro mortality prediction unit predicts local future mortality using a machine learning model for calculating the probability of death at the residence level, which model is generated using resident registration data; a mortality prediction adjustment step in which a mortality prediction adjustment unit compares the wide-area future mortality predicted by the macro mortality prediction unit with the local future mortality predicted by the micro mortality prediction unit, and adjusts prediction parameters of the micro mortality prediction unit; a macro birth population prediction step in which the macro birth population prediction unit predicts future birth populations over a wide area using a machine learning model for calculating birth probabilities on a regional basis, the model being generated using basic resident register data; a micro birth population prediction step in which the micro birth population prediction unit predicts a local future birth population using a machine learning model for calculating the probability of births occurring in each residential unit, which model is generated using resident basic register data; a birth population forecast adjustment step in which a birth population forecast adjustment unit compares the wide-area future birth population predicted by the macro birth population forecast unit with the local future birth population predicted by the micro birth population forecast unit, and adjusts the forecast parameters of the micro birth population forecast unit; an integration step in which the local future population calculation unit integrates the local future death population and future birth population predicted by the micro death population prediction unit and the micro birth population prediction unit after the prediction parameters have been adjusted by the death population prediction adjustment unit and the birth population prediction adjustment unit, thereby calculating the local future population; An information processing method including:

8. A macro mortality prediction step that predicts future mortality over a wide area using a machine learning model for calculating the probability of death on a regional basis, which is generated using basic resident register data; A micro-mortality prediction step predicts future local mortality using a machine learning model for calculating the probability of death at the residential level, which is generated using basic resident register data; a mortality prediction adjustment step of comparing the wide-area future mortality predicted in the macro mortality prediction step with the local future mortality predicted in the micro mortality prediction step and adjusting prediction parameters of the micro mortality prediction step; a macro birth population forecasting step that forecasts future birth populations over a wide area using a machine learning model for calculating birth probabilities on a regional basis, which is generated using basic resident register data; a micro-birth population forecasting step that forecasts local future birth populations using a machine learning model for calculating the probability of births occurring at residential units, which is generated using resident registration data; a birth population forecast adjustment step of comparing the wide-area future birth population predicted in the macro birth population forecast step with the local future birth population predicted in the micro birth population forecast step, and adjusting the forecast parameters of the micro birth population forecast step; an integration step of integrating the local future death population and future birth population predicted in the micro death population prediction step and the micro birth population prediction step after adjusting the prediction parameters in the death population prediction adjustment step and the birth population prediction adjustment step to calculate a local future population; An information processing program that causes a computer to execute the above.

Images