Systems, programs, information processing methods, machine learning models, and computers

Abstract

The objective of this invention is to provide a novel system, program, information processing method, machine learning model, and computer capable of predicting the move-out date of tenants in real estate. [Solution] One or more computers according to the present invention input at least one of the tenant information of the target tenant, the contract information of the target tenant, the property information of the target property in which the target tenant resides, or economic information to a machine learning model, causing the machine learning model to predict the date on which the target tenant will vacate the target property, or vacancy parameters related to the possibility of the target tenant vacating the target property, and display calculation result information showing the vacancy date or vacancy parameters predicted by the machine learning model.

Description

【Technical Field】 【0001】 The present invention relates to a system, a program, an information processing method, a machine learning model, and a computer. 【Background Art】 【0002】 As an invention related to a conventional system, for example, a tenant departure prediction device described in Patent Document 1 is known. In a tenant departure prediction system in which this tenant departure prediction device is communicably connected to a data providing device and an owner device via a network, the tenant departure prediction device includes a storage device that holds observation data observed regarding a predetermined event in a real estate and tenant departure case information in the real estate, and by providing the observation data and the departure case information as inputs to a machine learning engine, a process of generating a model that defines the relationship between the event indicated by the observation data and tenant departure, and a process of applying the observation data regarding the event in a predetermined real estate to the model to specify the tenant departure sign in the real estate and output information on the departure sign. 【Prior Art Documents】 【Patent Documents】 【0003】 【Patent Document 1】 Japanese Patent Application Laid-Open No. 2022-013307 【Summary of the Invention】 【Problems to be Solved by the Invention】 【0004】 As described above, it is desired to predict the departure date of the occupant of a real estate. 【0005】 Therefore, an object of the present invention is to provide a new system, program, information processing method, machine learning model, and computer capable of predicting the departure date of the occupant of a real estate. 【Means for Solving the Problems】 【0006】 The first form is, The system consists of one or more computers. Tenant change information is information relating to tenants who were living in the property, and which changes over time while the tenants were living in the property. The resident information includes resident change information for at least a portion of the period during which the resident resides in the property, Contract change information is information relating to the contract between the tenant who was residing in the property and the landlord of the property, and is information that changes over time while the tenant is residing in the property. The contract information includes contract change information covering at least a portion of the period during which the tenant resides in the property. Property change information is information relating to the property in which the tenant resided, and which changes over time while the tenant resided in the property. The property information includes property change information for at least a portion of the period during which the tenant resides in the property. The economic information includes economic indicators covering at least a portion of the period during which the tenant resides in the property. The aforementioned one or more computers By inputting at least one of the tenant information of the target tenant, the contract information of the target tenant, the property information of the target property in which the target tenant resides, or the economic information into a machine learning model, the machine learning model is made to predict the date on which the target tenant will vacate the target property, or a vacancy parameter related to the likelihood of the target tenant vacating the target property. The system displays calculation result information indicating the move-out date or move-out parameters predicted by the machine learning model. It is a system. 【0007】 The second form is, The machine learning model is a trained model that has been trained using training data that combines at least one of the tenant information, contract information, property information, or economic information with move-out time information regarding the timing when the tenant will move out of the property. This is the system described in the first form. 【0008】 The third form is, The aforementioned tenant change information includes at least one of the following: the tenant's sales, sales growth rate, number of employees, employee growth rate, number of new graduates hired, turnover rate, presence or absence of teleworking, stock price, average wage, average age, total number of locations, personnel information, fundraising information, average length of employment, number of job openings by area or job type, M&A information, gross profit, rent ratio to profit or sales, total area of ​​locations, number of locations per prefecture, or area per employee. The system is as described in the first or second form. 【0009】 The fourth form is, The aforementioned contract change information includes, in the contract between the tenant and the landlord of the property, at least one of the remaining contract period, contract area, contract rent, floor plan data of the property the tenant is occupying, or the rent-free period. The system is one of the first, second, or third forms described. 【0010】 The fifth form is, The aforementioned property change information includes at least one of the following: the age of the property, security, renovation date, vacancy rate, asking rent, total available area, or property owner. The system is one of the first to fourth forms. 【0011】 The sixth form is, The aforementioned economic indicators include at least one of the following: stock price index, long-term interest rate, short-term interest rate, GDP, employment statistics, price index, and exchange rate. The system is one of the first to fifth forms. 【0012】 The seventh form is, The supplementary information indicates a predicted value of the restoration cost when the occupant moves out of the property, or indicates whether the occupant has moved into the property as an unoccupied property. The teacher data includes the supplementary information. It is the system according to the second aspect. 【0013】 The eighth aspect is The one or more computers When the number of days until the move-out date indicated by the calculation result information is less than a first predetermined value, or when the move-out parameter indicated by the calculation result information indicates that it is higher than a second predetermined value, notify the user that the target occupant may move out of the target property. It is the system according to any one of the first aspect to the seventh aspect. 【0014】 The ninth aspect is The one or more computers Obtain target property information indicating a plurality of the target properties specified by the user. The one or more computers execute a process of causing the machine learning model to make a prediction for the plurality of target properties. It is the system according to the eighth aspect. 【0015】 The tenth aspect is A program, The occupant change information is information about the occupant who has moved into the property, and is information that changes over time while the occupant is living in the property. The occupant information includes the occupant change information over at least a part of the period during which the occupant is living in the property. The contract change information is information about the contract between the occupant who has moved into the property and the landlord of the property, and is information that changes over time while the occupant is living in the property. The contract information includes the contract change information over at least a part of the period during which the occupant is living in the property. Property change information is information relating to the property in which the tenant resided, and which changes over time while the tenant resided in the property. The property information includes property change information for at least a portion of the period during which the tenant resides in the property. The economic information includes economic indicators covering at least a portion of the period during which the tenant resides in the property. The program is installed in the control circuits of the one or more computers. By inputting at least one of the tenant information of the target tenant, the contract information of the target tenant, the property information of the target property in which the target tenant resides, or the economic information into a machine learning model, the machine learning model is made to predict the date on which the target tenant will vacate the target property, or a vacancy parameter related to the likelihood of the target tenant vacating the target property. The machine learning model displays calculation result information indicating the predicted move-out date or move-out parameters. It is a program. 【0016】 The 11th form is, Information processing method, Tenant change information is information relating to tenants who were living in the property, and which changes over time while the tenants were living in the property. The resident information includes resident change information for at least a portion of the period during which the resident resides in the property, Contract change information is information relating to the contract between the tenant who was residing in the property and the landlord of the property, and is information that changes over time while the tenant is residing in the property. The contract information includes contract change information covering at least a portion of the period during which the tenant resides in the property. Property change information is information relating to the property in which the tenant resided, and which changes over time while the tenant resided in the property. The property information includes property change information for at least a portion of the period during which the tenant resides in the property. The economic information includes economic indicators covering at least a portion of the period during which the tenant resides in the property. In the aforementioned information processing method, one or more computers, By inputting at least one of the tenant information of the target tenant, the contract information of the target tenant, the property information of the target property in which the target tenant resides, or the economic information into a machine learning model, the machine learning model is made to predict the date on which the target tenant will vacate the target property, or a vacancy parameter related to the likelihood of the target tenant vacating the target property. The system displays calculation result information indicating the move-out date or move-out parameters predicted by the machine learning model. It is an information processing method. 【0017】 The 12th form is, It is a machine learning model, Tenant change information is information relating to tenants who were living in the property, and which changes over time while the tenants were living in the property. The resident information includes resident change information for at least a portion of the period during which the resident resides in the property, Contract change information is information relating to the contract between the tenant who was residing in the property and the landlord of the property, and is information that changes over time while the tenant is residing in the property. The contract information includes contract change information covering at least a portion of the period during which the tenant resides in the property. Property change information is information relating to the property in which the tenant resided, and which changes over time while the tenant resided in the property. The property information includes property change information for at least a portion of the period during which the tenant resides in the property. The economic information includes economic indicators covering at least a portion of the period during which the tenant resides in the property. One or more computer memory circuits store machine learning models. The machine learning model is configured in the control circuits of one or more computers. Based on at least one of the tenant information of the tenant, the contract information of the tenant, the property information of the property in which the tenant resides, or the economic information, predict the date on which the tenant will vacate the property, or predict the likelihood of the tenant vacating the property. Output calculation result information indicating the aforementioned move-out date or the aforementioned move-out parameters. This is a machine learning model. 【0018】 The 13th form is, It is a computer, Tenant change information is information relating to tenants who were living in the property, and which changes over time while the tenants were living in the property. The resident information includes resident change information for at least a portion of the period during which the resident resides in the property, Contract change information is information relating to the contract between the tenant who was residing in the property and the landlord of the property, and is information that changes over time while the tenant is residing in the property. The contract information includes contract change information covering at least a portion of the period during which the tenant resides in the property. Property change information is information relating to the property in which the tenant resided, and which changes over time while the tenant resided in the property. The property information includes property change information for at least a portion of the period during which the tenant resides in the property. The economic information includes economic indicators covering at least a portion of the period during which the tenant resides in the property. The memory circuit of the aforementioned computer stores a machine learning model. The control circuit of the aforementioned computer is Based on at least one of the tenant information of the tenant, the contract information of the tenant, the property information of the property in which the tenant resides, or the economic information, the machine learning model is made to predict the date on which the tenant will vacate the property, or the likelihood of the tenant vacating the property. Output calculation result information indicating the aforementioned move-out date or the aforementioned move-out parameters. It is a computer. [Effects of the Invention] 【0019】 According to this invention, it is possible to predict the move-out date of tenants in a property. [Brief explanation of the drawing] 【0020】 [Figure 1] Figure 1 is a block diagram of System 1. [Figure 2] Figure 2 is a block diagram of terminal 10. [Figure 3] Figure 3 is a block diagram of server 110. [Figure 4] Figure 4 is a resident table showing resident information I1. [Figure 5] Figure 5 is a contract table showing contract information I2. [Figure 6] Figure 6 is a property table showing property information I3. [Figure 7] Figure 7 is an economic table showing economic information I4. [Figure 8] Figure 8 is a flowchart showing the actions performed by the control circuit 112 of the server 110 during the learning process. [Figure 9] Figure 9 is a flowchart showing the operations performed by the control circuit of terminal 10 and the control circuit 112 of server 110 during the prediction process. [Figure 10] Figure 10 shows the image displayed on the display 20 of terminal 10. [Modes for carrying out the invention] 【0021】 (Embodiment) System 1 according to an embodiment of this disclosure will be described with reference to the drawings. 【0022】 [System 1 Structure] First, the overall configuration of System 1 will be explained with reference to the diagrams. Figure 1 is a block diagram of System 1. Figure 2 is a block diagram of Terminal 10. Figure 3 is a block diagram of Server 110. 【0023】 System 1, shown in Figure 1, comprises terminals 10 (one or more computers) and servers 110 (one or more computers). Terminals 10 and servers 110 are connected to each other via a communication network, which can be the internet or an intranet. 【0024】 Terminal 10 is a computer. Terminal 10 is an information processing device used by a user. Terminal 10 is, for example, a smartphone, a tablet, or a personal computer. As shown in Figure 2, terminal 10 includes a control circuit 12, a memory circuit 14, a network interface 16, a graphics processing unit 18, a display 20, and an operation unit 26. 【0025】 The memory circuit 14 stores the program PG and data. The memory circuit 14 is, for example, a combination of ROM (Read Only Memory), RAM (Random Access Memory), and storage (for example, flash memory or hard disk). 【0026】 The program PG includes, for example, the following programs: • Programs for the OS (Operating System) • Programs for applications that perform information processing (e.g., web browsers, or target applications described later) 【0027】 The data includes, for example, the following: • Databases referenced in information processing • Data obtained by performing information processing (i.e., the results of information processing) 【0028】 The control circuit 12 implements the functions of the terminal 10 by executing the program PG stored in the memory circuit 14. The control circuit 12 is, for example, a circuit that includes at least one of the following: ·CPU(Central Processing Unit) ·GPU(Graphic Processing Unit) ·ASIC(Application Specific Integrated Circuit) ·FPGA(Field Programmable Array) 【0029】 The control circuit 12 includes, as functional blocks, a prediction instruction means 42, a display control means 44, a notification means 46, and a target property information acquisition means 48. 【0030】 The network interface 16 controls communication between the terminal 10 and external devices. The external device is server 110. 【0031】 The graphics processing unit 18 displays an image on the display 20 based on the image data generated by the control circuit 12. The display 20 is either a liquid crystal display or an organic EL (Electro-Luminescence) display. 【0032】 The operation unit 26 generates operation signals based on user input and outputs these signals to the control circuit 12. The operation unit 26 can be a touch panel, keyboard, mouse, or the like. 【0033】 System 1 may include multiple terminals 10. 【0034】 Server 110 is a computer. Server 110 is a web server that stores data for web pages. As shown in Figure 3, Server 110 includes a control circuit 112, a storage circuit 114, and a network interface 116. 【0035】 The memory circuit 114 stores programs and data. The memory circuit 114 is, for example, a combination of ROM (Read Only Memory), RAM (Random Access Memory), and storage (for example, flash memory or hard disk). 【0036】 The control circuit 112 implements the functions of the server 110 by executing the program stored in the memory circuit 114. The control circuit 112 is a circuit that includes, for example, at least one of the following: ·CPU(Central Processing Unit) ·GPU(Graphic Processing Unit) ·ASIC(Application Specific Integrated Circuit) ·FPGA(Field Programmable Array) 【0037】 The control circuit 112 includes a predictive control means 60 and an output means 62 as functional blocks. 【0038】 The memory circuit 114 stores the machine learning model 120. The machine learning program is a program that executes a machine learning algorithm to find certain rules from the training data and generate a trained machine learning model 120 that represents the found rules. When the control circuit 112 of the server 110 executes the machine learning program, multiple training data are trained and the parameters of the inference program are adjusted. As a result, the trained machine learning model 120 is generated. 【0039】 The machine learning algorithm is not particularly limited as long as it is supervised learning, and may include, for example, decision trees, nearest neighbors, Naive Bayesian classifiers, support vector machines, or neural networks. Therefore, the trained machine learning model 120 includes decision trees, nearest neighbors, Naive Bayesian classifiers, support vector machines, or neural networks. Backpropagation may be used in the machine learning process that generates the trained machine learning model 120. 【0040】 For example, a neural network includes an input layer, one or more hidden layers, and an output layer. Specifically, a neural network is a deep neural network. A deep neural network, also known as a recurrent neural network or convolutional neural network, performs deep learning. A deep neural network typically includes, for example, an input layer, multiple hidden layers, and an output layer. 【0041】 The network interface 116 controls communication between the server 110 and an external device. The external device is the terminal 10. 【0042】 [System 1 Operation] Next, we will explain the operation of System 1. Figure 4 is the tenant table showing tenant information I1. Figure 5 is the contract table showing contract information I2. Figure 6 is the property table showing property information I3. Figure 7 is the economy table showing economic information I4. 【0043】 The control circuit 12 of terminal 10 reads the program PG stored in the memory circuit 14, causing the program PG to execute the operations described below in the control circuit 12 of terminal 10. The program PG then causes the control circuit 12 of terminal 10 to function as a predictive instruction means 42, a display control means 44, a notification means 46, and a target property information acquisition means 48. In the information processing method, the control circuit 12 of terminal 10 executes the functions of the predictive instruction means 42, the display control means 44, the notification means 46, and the target property information acquisition means 48. 【0044】 The control circuit 112 of the server 110 reads the programs stored in the memory circuit 114, causing these programs to perform the operations described below in the control circuit 112 of the server 110. The programs then cause the control circuit 112 of the server 110 to function as a predictive control means 60 and an output means 62. In the information processing method, the control circuit 112 of the server 110 performs the functions of the predictive control means 60 and the output means 62. 【0045】 First, let's explain each table. The memory circuit 114 of server 110 stores the tenant table shown in Figure 4, the contract table shown in Figure 5, the property table shown in Figure 6, and the economy table shown in Figure 7. 【0046】 Tenant Information I1, shown in Figure 4, contains information about tenants who were residing in the property. Tenant Information I1 includes semi-annual (half-year) data on sales, sales growth rate, number of employees, employee growth rate, number of new graduates hired, turnover rate, industry, founding date, listing date, telework status, average wage, average length of service, average age, reason for relocation, total number of locations, personnel information, and fundraising information. Sales are the tenant's sales. Sales growth rate is the percentage increase in sales in the current half-year compared to sales in the first half-year. Number of employees is the total number of employees of the tenant. Employee growth rate is the percentage increase in the number of employees in the current half-year compared to the number of employees in the first half-year. Number of new graduates hired is the number of new graduates hired by the tenant each year. Turnover rate is the tenant's semi-annual turnover rate. Industry is the industry to which the tenant belongs. Founding date is the tenant's founding date. Listing date is the tenant's listing date. The "Telework Status" indicates whether or not the tenant is teleworking. Average wage is the average wage of the tenant's employees. Average length of service is the average number of years an employee has been employed by the tenant company. Average age is the average age of the tenant's employees. Reason for relocation is the reason the tenant moved its office. Total number of locations is the number of locations the tenant has. Personnel information is information about personnel changes that occurred at the tenant during each half-year period. Funding information is information about fundraising conducted by the tenant during each half-year period. 【0047】 Tenant information I1 includes tenant change information I11 and tenant non-change information I12. Tenant change information I11 is information about tenants who were residing in the property and that changes over time while the tenant is residing in the property. In the tenant table shown in Figure 4, tenant change information I11 includes sales, sales growth rate, number of employees, employee growth rate, number of new graduates hired, turnover rate, presence or absence of teleworking, average wage, average length of service, average age, reason for relocation, total number of locations, personnel information, and funding information. Tenant non-change information I12 is information about tenants who were residing in the property and that does not change over time while the tenant is residing in the property. In the tenant table shown in Figure 4, tenant non-change information I12 includes industry, founding date, and listing date. As shown in Figure 4, the tenant information I1 includes tenant change information I11 for at least a portion of the period during which the tenant resides in the property. In this embodiment, the tenant information I1 includes tenant change information I11 for the entire period during which the tenant resides in the property. 【0048】 The contract information I2 shown in Figure 5 contains information about the contract between the tenant who was occupying the property and the property owner. Contract information I2 includes information on the building name, move-in date, contract rent, contract area, contract duration, floor plan data, and the building's age at the time of move-in, updated every six months. The building name is the name of the building in which the tenant is occupying the property. The move-in date is the date the tenant moved into the property. The contract rent is the rent for the property in which the tenant is occupying the property. The contract area is the area of ​​the property in which the tenant is occupying the property. The contract duration is the remaining length of the contract period in which the tenant has an agreement with the property owner. The "Floor Plan Data" indicates the URL where the floor plan data for the property the resident is living in is stored. The "Year Built at Move-in" is the age of the property at the time the resident moved in. 【0049】 Contract information I2 includes contract change information I21 and contract non-change information I22. Contract change information I21 is information about the contract between the tenant who was residing in the property and the property owner, and is information that changes over time while the tenant is residing in the property. In the contract table shown in Figure 5, contract change information I21 includes contract rent, contract area, contract duration, and drawing data. Contract non-change information I22 is information about the contract between the tenant who was residing in the property and the property owner, and is information that does not change over time while the tenant is residing in the property. In the contract table shown in Figure 5, contract non-change information I22 includes building name, move-in date, and age of the building at the time of move-in. As shown in Figure 5, contract information I2 includes contract change information I21 for at least a portion of the period during which the tenant is residing in the property. In this embodiment, contract information I2 includes contract change information I21 for the entire period during which the tenant is residing in the property. 【0050】 The property information I3 shown in Figure 6 contains information about the property the tenant was occupying. Property information I3 includes the address, submarket, nearest station, travel time from terminal station, walking time to nearest station, year built, standard floor area, ceiling height, building grade, security, renovation date, vacancy rate, asking rent, total available area, and semi-annual information on the building owner. The address is the property's address. The submarket is the trade area to which the property belongs. The nearest station is the station closest to the property. Travel time from terminal station is the travel time from the terminal station to the property. Walking time to nearest station is the time it takes to walk to the nearest station. Year built is the year the property was built. Standard floor area is the total floor area including the private and common areas of one floor, corresponding to a "standard floor" in high-rise and mid-rise buildings. Ceiling height is the height from the finished floor to the bottom of the finished ceiling. Building grade refers to the grade of the property. Security refers to the security installed in the property. The Renewal Date is the date the property was renovated. The Vacancy Rate is the vacancy rate of the property. The Asking Rent is the rent for the rooms currently available for rent in the property. The Total Available Area is the total area of ​​the rooms currently available for rent in the property. The Building Owner is the owner of the property. 【0051】 Property information I3 includes property change information I31 and property non-change information I32. Property change information I31 is information about the property the tenant was occupying and that changes over time while the tenant is occupying the property. In the property table shown in Figure 6, property change information I31 includes the year of construction, security, renovation date, vacancy rate, asking rent, total available area, and building owner. Property non-change information I32 is information about the property the tenant was occupying and that does not change over time while the tenant is occupying the property. In the property table shown in Figure 6, property non-change information I32 includes the address, submarket, nearest station, travel time from terminal station, walking time to nearest station, standard floor area, ceiling height, and building grade. As shown in Figure 6, property information I3 includes property change information I31 for at least a portion of the period the tenant is occupying the property. In this embodiment, property information I3 includes property change information I31 covering the entire period during which the tenant resides in the property. 【0052】 Economic information I4 includes economic indicators covering at least a portion of the period during which the tenant resides in the property. In this embodiment, economic information I4 includes economic indicators covering the entire period during which the tenant resides in the property. Economic information I4 includes semi-annual information on stock price indices, long-term interest rates, short-term interest rates, GDP, employment statistics, price indices, and exchange rates. The stock price index is the Nikkei 225. The long-term interest rate is Japan's long-term interest rate. The short-term interest rate is Japan's short-term interest rate. GDP is Japan's Gross Domestic Product. Employment statistics are Japan's unemployment rate. The price index is the Consumer Price Index. The exchange rate shows how many yen it costs to buy one dollar. 【0053】 Furthermore, the memory circuit 114 of server 110 also stores tables for tenants other than Company M, contract tables other than Company M, and property tables other than Building X. 【0054】 Furthermore, the memory circuit 114 of the server 110 stores the move-out date information I5. The move-out date information I5 indicates the date on which the tenant, as indicated by the tenant information I1, moved out of the property contracted in the contract information I2. In this embodiment, the move-out date information indicates that Company M, as indicated by the tenant information I1, moved out of Building X, as indicated in the contract information I2, on July 1, 2025. However, in addition to the move-out date information I5 for Company M, the memory circuit 114 of the server 110 also stores the move-out date information I5 for tenants other than Company M. 【0055】 The operation of System 1 includes the following: (1) learning process and (2) prediction process. (1) Learning process: This is the process in which a machine learning model 120 performs machine learning using training data, thereby generating a trained model. (2) Prediction process: This is the process in which the machine learning model 120 predicts the date when tenants will move out. 【0056】 (1) Learning process First, the learning process will be explained with reference to the diagram. Figure 8 is a flowchart of the control circuit 112 of the server 110 executed during the learning process. 【0057】 First, the control circuit 112 of the server 110 acquires tenant information I1, contract information I2, property information I3, economic information I4, and move-out date information I5 from the memory circuit 114 (step S111). In step S111, the control circuit 112 of the server 110 acquires not only tenant information I1 of company M, contract information I2 of company M, and property information I3 of building X from the memory circuit 114, but also tenant information I1 of companies other than company M, contract information I2 of companies other than company M, and property information I3 of buildings other than building X. In other words, the control circuit 112 of the server 110 acquires all of the available tenant information I1, contract information I2, property information I3, economic information I4, and move-out date information I5 from the memory circuit 114. 【0058】 Next, the control circuit 112 of the server 110 uses the combination of tenant information I1, contract information I2, property information I3, economic information I4, and move-out date information I5 as training data to cause the machine learning model 120 to perform machine learning processing (step S112). More specifically, the control circuit 112 of the server 110 generates training data consisting of the combination of tenant information I1 of Company M shown in Figure 4, contract information I2 of Company M shown in Figure 5, property information I3 of Building X shown in Figure 6, economic information I4 shown in Figure 7, and move-out date information I5 indicating the move-out date of Company M. At this time, the control circuit 112 of the server 110 sets the move-out risk for the half-year period to which the move-out date (July 1, 2025) belongs to 100. Move-out risk is a parameter that indicates the likelihood that the tenant (Company M) will move out of the property (Company X). A higher move-out risk indicates a higher likelihood that the tenant (Company M) will move out. Furthermore, the control circuit 112 of server 110 sets the move-out risk so that the move-out risk decreases as the time period moves further away from the half-year in which the move-out date (July 1, 2025) falls. The control circuit 112 of server 110 also generates training data for tenants other than those of Company M. This completes the process. 【0059】 Through the above process, the control circuit 112 of the server 110 performs machine learning on the machine learning model 120 to generate a trained model. Thus, the machine learning model 120 is a trained model that has been trained using a combination of tenant information I1, contract information I2, property information I3, and economic information I4, along with move-out timing information I5, which concerns when tenants move out of the property, as training data. 【0060】 (2) Prediction processing First, the prediction process will be explained with reference to the diagram. Figure 9 is a flowchart showing the operations performed by the control circuit of terminal 10 and the control circuit 112 of server 110 during the prediction process. Figure 10 shows the image displayed on the display 20 of terminal 10. 【0061】 The user inputs input information I0 by operating the operation unit 26 of terminal 10. Input information I0 is information about tenants whose move-out date is to be predicted. Hereinafter, the tenant whose move-out date is to be predicted will be referred to as the target tenant. The target tenant is Company N. The property in which the target tenant resides will be referred to as the target property. The target property is Building Y. Input information I0 consists of the target tenant's tenant information I1, the target tenant's contract information I2, and the target property's property information I3. Furthermore, the control circuit 12 of terminal 10 obtains economic information I4 from the memory circuit 114. At this time, the control circuit 12 of terminal 10 obtains economic information I4 for the period in which the target tenant resides in the target property. As a result, the control circuit 12 of terminal 10 obtains input information I0 including the target tenant's tenant information I1, the target tenant's contract information I2, the target property's property information I3, and the economic information I4 (step S1). Note that in the training data, the properties listed in tenant information I1, contract information I2, and property information I3 were properties where the tenants had resided, whereas in input information I0, the target properties listed in tenant information I1, contract information I2, and property information I3 are properties where the target tenants currently reside. 【0062】 Next, the control circuit 12 of terminal 10 transmits input information I0 to server 110 via network interface 16 (step S2). Accordingly, the network interface 116 of server 110 receives input information I0 and outputs input information I0 to control circuit 112. As a result, control circuit 112 of server 110 acquires input information I0 (step S101). 【0063】 Next, the control circuit 112 (predictive control means 60) of the server 110 causes the machine learning model 120 to predict the date on which Company N (target tenant) will vacate Building Y (target property), and the vacancy risk (vacancy parameters) related to the possibility of Company N (target tenant) vacating Building Y (target property), based on input information I0 including tenant information I1 of the target tenant, contract information I2 of the target tenant, property information I3 of the target property in which the target tenant is residing, and economic information I4 (step S102). As a result, the control circuit 112 of the server 110 generates calculation result information I6 that shows the date on which Company N (target tenant) will vacate Building Y (target property), and the vacancy risk (vacancy parameters) related to the possibility of Company N (target tenant) vacating Building Y (target property). 【0064】 Here, in step S1, the control circuit 12 of terminal 10 transmits input information I0 to server 110 via the network interface 16, and thereafter, the control circuit 112 of server 110 causes the machine learning model 120 to predict the move-out date and move-out risk based on the input information I0. In other words, in step S1, the control circuit 12 of terminal 10 (prediction instruction means 42) inputs input information I0, which includes tenant information I1 of company N (target tenant), contract information I2 of company N (target tenant), property information I3 of building Y (target property) where company N (target tenant) is located, and economic information I4, to the machine learning model 120, thereby causing the machine learning model 120 to predict the move-out date when company N (target tenant) will move out of building Y (target property), and the move-out risk (move-out parameter) regarding the possibility of company N (target tenant) moving out of building Y (target property). 【0065】 Next, the control circuit 112 of the server 110 transmits the calculation result information I6 to the terminal 10 via the network interface 116 (step S103). That is, the control circuit 112 (output means 62) of the server 110 outputs the calculation result information I6 indicating the departure date and departure risk (departure parameters). Accordingly, the network interface 16 of the terminal 10 receives the calculation result information I6 and outputs the calculation result information I6 to the control circuit 12. As a result, the control circuit 12 of the terminal 10 acquires the calculation result information I6 (step S3). 【0066】 Next, the control circuit 12 (display control means 44) of terminal 10 displays calculation result information I6, which shows the move-out date or move-out parameters predicted by the machine learning model 120, on the display 20 (step S4). More specifically, the control circuit 12 of terminal 10 displays the calculation result image shown in Figure 10 on the display 20. The calculation result image shown in Figure 10 includes not only the move-out date and move-out parameters of company A, but also the move-out dates and move-out risks of other tenants residing in Building Y. The move-out dates and move-out risks of other tenants residing in Building Y are those previously predicted by the machine learning model 120. In this case, the memory circuit 14 of terminal 10 stores the calculation result information I6 showing the move-out dates and move-out risks of other tenants residing in Building Y. 【0067】 Here, in step S102, the control circuit 12 (notification means 46) of terminal 10 notifies the user that the target tenant may move out of the target property if the number of days until the move-out date indicated by the calculation result information I6 is less than a first predetermined value, or if the move-out risk (move-out parameter) indicated by the calculation result information I6 is high. The first predetermined value is, for example, 60 days. The second predetermined value is, for example, 70. In the calculation result image shown in Figure 10, the move-out date for company N is September 1, 2025. Today is July 10, 2025. Therefore, the number of days until company N's move-out date is less than 60 days. Also, the move-out risk for company N is 90. Therefore, the move-out risk for company N is higher than 70. Thus, the control circuit 12 of terminal 10 highlights company N's move-out date and move-out risk in the calculation result image shown in Figure 10. In this embodiment, the control circuit 12 of terminal 10 hatches the departure date and departure risk of company N. This allows the user to recognize that there is a high probability that company N will vacate the premises. 【0068】 [effect] In System 1, the control circuit 12 of terminal 10 inputs input information I0, which includes N company's tenant information I1, N company's contract information I2, property information I3 of Building Y where N company is located, and economic information I4, to the machine learning model 120, causing the machine learning model 120 to predict the date on which N company will vacate Building Y and the risk of N company vacating Building Y. Here, tenant information I1 includes tenant change information I11 for at least a portion of the period N company is occupying Building Y. Contract information I2 includes contract change information I21 for at least a portion of the period N company is occupying Building Y. Property information I3 includes property change information I31 for at least a portion of the period N company is occupying Building Y. Economic information I4 includes economic indicators for at least a portion of the period N company is occupying Building Y. Thus, the machine learning model 120 predicts the date on which Company N will vacate Building Y, and the risk of Company N vacating Building Y, based on information that changes during at least a portion of the period in which Company N is occupying Building Y, and that may affect Company N's departure from Building Y. As a result, System 1 can accurately predict the departure date of tenants in a property. 【0069】 In System 1, the control circuit 12 of terminal 10 notifies the user that the tenant may be moving out of the property if the number of days until the move-out date indicated by the calculation result information I6 is less than a first predetermined value, or if the calculation result information I6 indicates a high risk of move-out. This allows the user to learn early about the possibility of the tenant moving out of the property. 【0070】 (Other embodiments) The system according to the present invention is not limited to System 1, but can be modified within the scope of its gist. 【0071】 The memory circuit 114 of server 110 stores resident information I1, contract information I2, property information I3, and economic information I4. However, one or more servers other than server 110 may have a database, and the memory circuits of one or more servers may store resident information I1, contract information I2, property information I3, and economic information I4 in the database. In this case, in step S1, the control circuit 112 of server 110 may acquire at least a portion of the input information I0, including resident information I1, contract information I2, property information I3, and economic information I4, from one or more servers. 【0072】 Furthermore, the control circuit 12 of terminal 10 notifies the user that the tenant may vacate the property if the number of days until the move-out date indicated by the calculation result information I6 is less than a first predetermined value, or if the move-out risk (move-out parameter) indicated by the calculation result information I6 is high. This notification is presented to the user by the calculation result image shown in Figure 10. However, this notification may also be sent, for example, by email, a messaging function attached to system 1, or a messaging function such as SNS. 【0073】 Furthermore, the control circuit 112 of the server 110 may cause the machine learning model 120 to predict the date on which Company N (the target tenant) will vacate Building Y (the target property), and the risk of Company N (the target tenant) vacating Building Y (the target property), based on at least one of the following: tenant information I1 of the target tenant, contract information I2 of the target tenant, and property information I3 or economic information I4 of the target property in which the target tenant resides. 【0074】 Furthermore, the move-out date information I5 does not need to be limited to the move-out date; it can be any information regarding the time when the tenant moves out of the property. Therefore, the move-out date information may be the month in which the tenant moved out of the property, the day or month on which the tenant is presumed to have moved out of the property, or the fact that the tenant moved out. 【0075】 The eviction risk (eviction parameter) may also be a value expressed as a percentage of the probability of the tenant vacating within X years, where X is, for example, 1. 【0076】 Furthermore, in the calculation result image shown in Figure 10, the control circuit 12 of terminal 10 may, instead of hatching the departure date and departure risk of company N, use a different hatching color for company N's departure date and departure risk than the colors for the departure dates and departure risks of other tenants, or it may add a mark to draw attention to company N's departure date and departure risk. 【0077】 Furthermore, tenant change information I11 may include, in addition to sales, sales growth rate, number of employees, employee growth rate, number of new graduates hired, turnover rate, presence or absence of teleworking, stock price, average wage, average age, total number of locations, personnel information, fundraising information, or average length of employment, the number of job openings by area or occupation, M&A information, gross profit, rent ratio to profit or sales, total area of ​​locations, number of locations in each prefecture, or area per employee. In addition, tenant change information I11 only needs to include at least one of the following: sales, sales growth rate, number of employees, employee growth rate, number of new graduates hired, turnover rate, presence or absence of teleworking, stock price, average wage, average age, total number of locations, personnel information, fundraising information, average length of employment, number of job openings by area or occupation, M&A information, gross profit, rent ratio to profit or sales, total area of ​​locations, number of locations in each prefecture, or area per employee. 【0078】 Furthermore, contract change information I21 may include a rent-free period in addition to the remaining contract period, contract area, contract rent, or floor plan data of the property occupied by the tenant. In addition, contract change information I21 only needs to include at least one of the remaining contract period, contract area, contract rent, floor plan data of the property occupied by the tenant, or a rent-free period. A rent-free period is a period in a lease agreement during which rent payment is waived or reduced, for example, a period during which rent is free from the move-in date for a specified period. 【0079】 Furthermore, property change information I31 only needs to include at least one of the following: the age of the property, security, renovation date, vacancy rate, asking rent, total available area, or property owner. 【0080】 Furthermore, economic indicators only need to include at least one of the following: stock market index, long-term interest rate, short-term interest rate, GDP, employment statistics, price index, or exchange rate. 【0081】 The training data may also include supplementary information. This supplementary information may indicate the estimated cost of restoring the property to its original condition when a tenant moves out, or whether the tenant moved into the property as a "turnkey" property (with existing fixtures and fittings). In this case, input information I0 may also include supplementary information. 【0082】 Furthermore, the control circuit 112 (target property information acquisition means 48) of the server 110 may acquire target property information indicating multiple target properties specified by the user, and perform a process to have the machine learning model 120 make predictions for the multiple target properties. That is, in step S102, the control circuit 112 of the server 110 has the machine learning model 120 predict the date on which other tenants will vacate Y Building, and the risk of other tenants vacating Y Building, based on input information I0 which includes tenant information I1 of other tenants, contract information I2 of other tenants, property information I3 of Y Building where other tenants are located, and economic information I4. 【0083】 Furthermore, the machine learning model 120 may be a pre-trained model that has been trained using machine learning with training data consisting of a combination of at least one of tenant information I1, contract information I2, property information I3, or economic information I4, and move-out timing information I5, which concerns when tenants move out of the property. 【0084】 The memory circuit 114 of server 110 stores the machine learning model 120. However, the memory circuit 114 of a server other than server 110 may also store the machine learning model 120. 【0085】 The memory circuit 114 of the server 110 may store a Large Language Model (LLM) as the machine learning model 120. In this case, the terminal 10 sends a prompt along with the input information I0. The prompt includes instructions to cause the LLM to predict the move-out date on which the target tenant will vacate the property, or move-out parameters related to the likelihood of the target tenant vacating the property. 【0086】 Note that contract information I2 includes the contract rent. However, it may be difficult for the operator of server 110 to obtain information on the contract rent. In such cases, contract information I2 may include the advertised rent instead of the contract rent. The control circuit 112 of server 110 may estimate the contract rent based on this advertised rent. [Explanation of symbols] 【0087】 1: System 10: Terminal 12: Control circuits 14:Memory circuit 16: Network Interface 18: Graphics Processing Unit 20: Display 26:Operation unit 42: Predictive instruction means 44: Display control means 46: Notification means 48: Method for obtaining target property information 60: Predictive control means 62: Output means 110: Server 112: Control circuits 114:Memory circuit 116: Network Interface 120: Machine Learning Models I0: Input Information I1: Tenant Information I11: Information on changes in occupants I12: Information on no change in resident status I2: Contract Information I21: Contract Change Information I22: Contract Non-Change Information I3: Property Information I31: Property Change Information I32: Property Status Information I4: Economic Information I5:Moving out time information I6: Calculation result information PG: Program

Claims

[Claim 1] The system consists of one or more computers. Tenant change information is information relating to tenants who were living in the property, and which changes over time while the tenants were living in the property. The resident information includes resident change information for at least a portion of the period during which the resident resides in the property, Contract change information is information relating to the contract between the tenant who was residing in the property and the landlord of the property, and is information that changes over time while the tenant is residing in the property. The contract information includes contract change information covering at least a portion of the period during which the tenant resides in the property. Property change information is information relating to the property in which the tenant resided, and which changes over time while the tenant resided in the property. The property information includes property change information for at least a portion of the period during which the tenant resides in the property. The economic information includes economic indicators covering at least a portion of the period during which the tenant resides in the property. The one or more computers mentioned above are By inputting at least one of the tenant information of the target tenant, the contract information of the target tenant, the property information of the target property in which the target tenant resides, or the economic information into a machine learning model, the machine learning model is made to predict the date on which the target tenant will vacate the target property, or a vacancy parameter related to the likelihood of the target tenant vacating the target property. The system displays calculation result information indicating the move-out date or move-out parameters predicted by the machine learning model. system. [Claim 2] The machine learning model is a trained model that has been trained using training data that combines at least one of the tenant information, contract information, property information, or economic information with move-out time information regarding the timing when the tenant will move out of the property. The system according to claim 1. [Claim 3] The aforementioned tenant change information includes at least one of the following: the tenant's sales, sales growth rate, number of employees, employee growth rate, number of new graduates hired, turnover rate, presence or absence of teleworking, stock price, average wage, average age, total number of locations, personnel information, fundraising information, average length of employment, number of job openings by area or job type, M&A information, gross profit, rent ratio to profit or sales, total area of ​​locations, number of locations per prefecture, or area per employee. The system according to claim 1 or claim 2. [Claim 4] The aforementioned contract change information includes at least one of the following in the contract between the tenant and the landlord of the property: the remaining length of the contract period, the contract area, the contract rent, the floor plan data of the property the tenant is occupying, or the rent-free period. The system according to claim 1 or claim 2. [Claim 5] The aforementioned property change information includes at least one of the following: the age of the property, security, renovation date, vacancy rate, asking rent, total available area, or property owner. The system according to claim 1 or claim 2. [Claim 6] The aforementioned economic indicators include at least one of the following: stock price index, long-term interest rate, short-term interest rate, GDP, employment statistics, price index, and exchange rate. The system according to claim 1 or claim 2. [Claim 7] The supplementary information shows an estimated value of the restoration costs incurred when the tenant vacates the property, or indicates whether the tenant moved into the property as a "turnkey" property (with existing fixtures and fittings). The aforementioned training data includes the aforementioned auxiliary information. The system according to claim 2. [Claim 8] The one or more computers mentioned above are If the calculation result information indicates that the number of days until the move-out date is less than a first predetermined value, or if the calculation result information indicates that the move-out parameter is higher than a second predetermined value, the user is notified that the tenant may move out of the property. The system according to claim 1 or claim 2. [Claim 9] The one or more computers mentioned above are The system acquires target property information indicating a plurality of target properties specified by the user, The one or more computers mentioned above perform a process to cause the machine learning model to make predictions about the multiple target properties. The system according to claim 8. [Claim 10] It is a program, Tenant change information is information relating to tenants who were living in the property, and which changes over time while the tenants were living in the property. The resident information includes resident change information for at least a portion of the period during which the resident resides in the property, Contract change information is information relating to the contract between the tenant who was residing in the property and the landlord of the property, and is information that changes over time while the tenant is residing in the property. The contract information includes contract change information covering at least a portion of the period during which the tenant resides in the property. Property change information is information relating to the property in which the tenant resided, and which changes over time while the tenant resided in the property. The property information includes property change information for at least a portion of the period during which the tenant resides in the property. The economic information includes economic indicators covering at least a portion of the period during which the tenant resides in the property. The program is provided to the control circuits of the one or more computers. By inputting at least one of the tenant information of the target tenant, the contract information of the target tenant, the property information of the target property in which the target tenant resides, or the economic information into a machine learning model, the machine learning model is made to predict the date on which the target tenant will vacate the target property, or a vacancy parameter related to the likelihood of the target tenant vacating the target property. The machine learning model displays calculation result information indicating the predicted move-out date or move-out parameters. program. [Claim 11] Information processing method, Tenant change information is information relating to tenants who were living in the property, and which changes over time while the tenants were living in the property. The resident information includes resident change information for at least a portion of the period during which the resident resides in the property, Contract change information is information relating to the contract between the tenant who was residing in the property and the landlord of the property, and is information that changes over time while the tenant is residing in the property. The contract information includes contract change information covering at least a portion of the period during which the tenant resides in the property. Property change information is information relating to the property in which the tenant resided, and which changes over time while the tenant resided in the property. The property information includes property change information for at least a portion of the period during which the tenant resides in the property. The economic information includes economic indicators covering at least a portion of the period during which the tenant resides in the property. In the aforementioned information processing method, one or more computers, By inputting at least one of the tenant information of the target tenant, the contract information of the target tenant, the property information of the target property in which the target tenant resides, or the economic information into a machine learning model, the machine learning model is made to predict the date on which the target tenant will vacate the target property, or a vacancy parameter related to the likelihood of the target tenant vacating the target property. The system displays calculation result information indicating the move-out date or move-out parameters predicted by the machine learning model. Information processing methods. [Claim 12] It is a machine learning model, Tenant change information is information relating to tenants who were living in the property, and which changes over time while the tenants were living in the property. The resident information includes resident change information for at least a portion of the period during which the resident resides in the property, Contract change information is information relating to the contract between the tenant who was residing in the property and the landlord of the property, and is information that changes over time while the tenant is residing in the property. The contract information includes contract change information covering at least a portion of the period during which the tenant resides in the property. Property change information is information relating to the property in which the tenant resided, and which changes over time while the tenant resided in the property. The property information includes property change information for at least a portion of the period during which the tenant resides in the property. The economic information includes economic indicators covering at least a portion of the period during which the tenant resides in the property. One or more computer memory circuits store machine learning models. The machine learning model is configured in the control circuits of one or more computers. Based on at least one of the tenant information of the tenant, the contract information of the tenant, the property information of the property in which the tenant resides, or the economic information, predict the date on which the tenant will vacate the property, or predict the likelihood of the tenant vacating the property. Output calculation result information indicating the aforementioned move-out date or the aforementioned move-out parameters. Machine learning models. [Claim 13] It is a computer, Tenant change information is information relating to tenants who were living in the property, and which changes over time while the tenants were living in the property. The resident information includes resident change information for at least a portion of the period during which the resident resides in the property, Contract change information is information relating to the contract between the tenant who was residing in the property and the landlord of the property, and is information that changes over time while the tenant is residing in the property. The contract information includes contract change information covering at least a portion of the period during which the tenant resides in the property. Property change information is information relating to the property in which the tenant resided, and which changes over time while the tenant resided in the property. The property information includes property change information for at least a portion of the period during which the tenant resides in the property. The economic information includes economic indicators covering at least a portion of the period during which the tenant resides in the property. The memory circuit of the aforementioned computer stores a machine learning model. The control circuit of the aforementioned computer is Based on at least one of the tenant information of the tenant, the contract information of the tenant, the property information of the property in which the tenant resides, or the economic information, the machine learning model is made to predict the date on which the tenant will vacate the property, or the likelihood of the tenant vacating the property. Output calculation result information indicating the aforementioned move-out date or the aforementioned move-out parameters. computer.

Images