JP7369098B2 - Tenant eviction prediction device, tenant eviction prediction method, and tenant eviction prediction system - Google Patents

Description

The present invention relates to a tenant leaving prediction device, a tenant leaving prediction method, and a tenant leaving prediction system.

So-called real estate tenant businesses, such as real estate developers, act as landlords themselves, leasing real estate such as land and buildings to others, and earning rent.
In this type of business, it is desirable for properties to be filled with tenants on a continuous basis and rental income to be stabilized. Therefore, techniques have been proposed to stabilize rental income.

Such conventional technology includes, for example, a system that sequentially evaluates variations in rental income, and as an initial setting, the initial value of the correlation coefficient of rent fluctuations among multiple tenants in real estate and the rent for each tenant are used. a means for setting the volatility of the rental income of the entire real estate, a means for calculating the volatility related to the rental income of the entire real estate, a means for adding and updating a time series data file related to rent on a monthly basis, and a means for each time the time series data file is updated. A means to recalculate the correlation coefficient between tenants regarding rent changes, update the overall rental income volatility using the updated correlation coefficient, and calculate the rent and vacancy rate based on the updated conditions. A real estate time-series risk management system (see Patent Document 1) has been proposed, which has a means for calculating a probability distribution of revenue prices as a random variable and a means for calculating a risk management index based on the probability distribution of revenue prices of the real estate. has been done.

Japanese Patent Application Publication No. 2005-258506

However, no technology has been proposed to deal with tenant evictions, which can have a major impact on stabilizing rental income. It is true that rental contracts generally stipulate the obligation to give notice of eviction within a certain period of time before the tenant's eviction date, and it is possible to receive such notice once the tenant has decided to vacate. be.
In other words, although it is possible to know about the tenant moving out, which has already been decided, there is no technology that can detect signs of eviction and take some kind of response before the eviction notice is issued.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a technology that detects signs of tenant leaving and supports appropriate measures in response.

A tenant eviction prediction device of the present invention that solves the above problems is provided with observation data observed regarding a predetermined event in real estate, eviction case information including at least the eviction history of the tenant in the real estate, and sales to tenants for each of the eviction history. By providing a storage device that holds activity guideline information and eviction case information including the observed data and the eviction process as input to a machine learning engine, it is possible to determine the relationship between the events indicated by the observation data and the eviction process. By applying the process of generating a model that defines the nature of the property and the observed data regarding the above-mentioned events in a given real estate to the model, the likely eviction history of the tenant in the real estate is identified as a sign of eviction , and the specified eviction history is applied to the model. The present invention is characterized by comprising a calculation device that executes a process of comparing prospects with the guideline information, identifying guidelines for business activities for the tenant, and outputting information on the guidelines.
Further, in the tenant eviction prediction method of the present invention, the information processing device collects observation data observed regarding a predetermined event in the real estate, eviction case information including at least the eviction history of the tenant in the real estate, and the tenant eviction history for each predicted eviction history. By storing the observation data and eviction case information including the eviction details in a storage device, and providing the machine learning engine with the observation data and the eviction case information including the eviction history , it is possible to calculate the events indicated by the observation data and the tenant eviction. By applying the process of generating a model that defines the relationship between the circumstances and the observed data regarding the above-mentioned events in a given real estate to the model, the likely exit history of the tenant in the real estate can be identified as a sign of eviction , and the above-mentioned The method is characterized in that the specified expected eviction history is checked against the guideline information, the guideline for business activities for the tenant is specified, and the information on the guideline is output.

Furthermore, the tenant eviction prediction system of the present invention includes observation data observed regarding a predetermined event in real estate, eviction case information including at least the history of the tenant's eviction from the real estate, and guidelines for sales activities for the tenant for each of the expected eviction history. A storage device that holds information , and eviction case information including the observation data and the eviction history are given to a machine learning engine as input, thereby defining the relationship between the events indicated by the observation data and the eviction history of the tenant. By applying observation data regarding the above-mentioned events in a predetermined real estate to the model, the expected eviction history of the tenant in the real estate is identified as a sign of eviction , and the identified expected eviction history is used as a sign of eviction. The present invention is characterized in that it includes a tenant eviction prediction device that includes a calculation device that executes processing of identifying a guideline for business activities for the tenant by comparing it with guideline information, and outputting information on the guideline.

According to the present invention, it is possible to detect a sign of a tenant moving out and support appropriate measures in response.

FIG. 1 is a network configuration diagram including a tenant departure prediction device according to the present embodiment. It is a diagram showing an example of the hardware configuration of a tenant leaving prediction device in this embodiment. It is a diagram showing an example of the hardware configuration of a data providing device in this embodiment. It is a diagram showing an example of the hardware configuration of the owner device in this embodiment. It is a figure showing an example of a data composition of observation data DB of this embodiment. It is a diagram showing an example of the data structure of a withdrawal case DB according to the present embodiment. FIG. 3 is a diagram showing an example of the data structure of a model DB according to the present embodiment. It is a figure showing an example of data composition of analysis result DB of this embodiment. FIG. 3 is a diagram showing an example of the data structure of a recommendation table according to the present embodiment. It is a diagram showing an example of the data structure of a sales result DB according to the present embodiment. It is a figure showing an example of a flow of a tenant leaving prediction method in this embodiment. It is a figure showing an example of a flow of a tenant leaving prediction method in this embodiment. It is a figure which shows the example of a screen in this embodiment. It is a figure which shows the example of a screen in this embodiment. It is a figure which shows the example of a screen in this embodiment.

<<Network configuration>>
Embodiments of the present invention will be described in detail below with reference to the drawings. FIG. 1 is a network configuration diagram including a tenant departure prediction device 100 of this embodiment. The tenant eviction prediction device 100 shown in FIG. 1 is a computer device that detects signs of tenant eviction in real estate 20 and supports appropriate measures in response.

As shown in the network configuration of FIG. 1, the tenant departure prediction device 100 of this embodiment is communicably connected to the data providing device 200 and the owner device 300 via the network 1.

Further, it is assumed that the tenant eviction prediction device 100, the data providing device 200, and the owner device 300, which are connected to each other in the network 1, constitute the tenant eviction prediction system 10.

Among these, the tenant departure prediction device 100 is an information processing device that mainly executes the tenant departure prediction method of this embodiment. This tenant eviction prediction device 100 generates a model of a sign of a tenant leaving in machine learning, and provides information on a sign of a tenant leaving the real estate 20 based on this model.

The operator of such a tenant eviction prediction device 100 can be assumed to be, for example, the owner of the real estate 20 or the provider of observation data regarding the real estate 20, in addition to the above-mentioned business operator that provides information provision service for signs of tenant eviction. (Of course, it is not limited to this).

Further, the data providing device 200 is a device that observes various events related to the real estate 20 described above using an observation device 30 such as a sensor, and provides the observed data. Such data providing device 200 can be mainly assumed to be operated by a real estate management company, for example. Of course, the operator may be a business operator that provides the above-mentioned tenant eviction sign information provision service or a real estate owner.

Further, the owner device 300 is an information processing device operated by a real estate owner, and serves as an information processing device that receives information on signs of tenant eviction from the tenant eviction prediction device 100.
<<Hardware configuration>>
Further, the hardware configuration of the tenant leaving prediction device 100 of this embodiment is as shown in FIG. 2. The tenant eviction prediction device 100 includes a storage device 101, a memory 103, a calculation device 104, an input device 105, an output device 106, and a communication device 107.

Among these, the storage device 101 is configured with an appropriate nonvolatile storage element such as an SSD (Solid State Drive) or a hard disk drive.

Furthermore, the memory 103 is composed of a volatile storage element such as a RAM.

Further, the arithmetic device 104 is a CPU that reads the program 102 held in the storage device 101 into the memory 103 and executes it, performs overall control of the device itself, and performs various judgments, calculations, and control processing.

The program 102 described above includes a machine learning engine 110 in addition to implementing an operating system, a file system, etc. that are common to information processing devices. This machine learning engine 110 receives real estate observation data provided from the data providing device 200 and eviction case information provided from the owner device 300 as input, and defines the relationship between events indicated by the observation data and tenant evictions. It will generate a model that

The machine learning engine 110 continues to tune the model by acquiring, for example, feedback from the owner device 300 regarding whether or not the tenant actually vacates and the circumstances behind the eviction regarding the signs of tenant eviction estimated by the above-described model. Existing techniques may be appropriately applied to the method of generating and tuning such a model.

Note that the above-described machine learning engine 110 may be provided in the tenant eviction prediction device 100 itself, or may be called and used from an external device on the network 1.

Further, the input device 105 is an appropriate device such as a keyboard, a mouse, a microphone, etc. that accepts key input and voice input from the user.

Further, the output device 106 is an appropriate device such as a display or a speaker that displays the data processed by the arithmetic device 104.

Further, the communication device 107 is a network interface card that connects to the network 1 and handles communication processing with other devices such as the data providing device 200 and the owner device 300.

The storage device 101 includes, in addition to the program 102 for implementing the functions necessary for the tenant eviction prediction device 100 of this embodiment, an observation data DB 125, an eviction case DB 126, a model DB 127, an analysis result DB 128, and a recommendation. At least a table 129 is stored. Details of these databases and tables will be described later.

FIG. 3 shows an example of the hardware configuration of the data providing device 200 of this embodiment. The data providing device 200 of this embodiment includes a storage device 201, a memory 203, an arithmetic device 204, and a communication device 205.

Among these, the storage device 201 is configured with an appropriate nonvolatile storage element such as an SSD (Solid State Drive) or a hard disk drive.

Furthermore, the memory 203 is composed of a volatile storage element such as a RAM.

Further, the arithmetic device 204 is a CPU that reads out and executes the program 202 held in the storage device 201 into the memory 203, performs overall control of the device itself, and performs various judgments, calculations, and control processing.

Further, the communication device 205 is a network interface card that connects to the network 1 and handles communication processing with other devices such as the observation device 30 of the real estate 20 to be observed and the tenant eviction prediction device 100.

Note that the above-mentioned observation device 30 can monitor, for example, the number and weight of people getting on and off elevators, the amount of usage of various infrastructures such as electricity, gas, water and sewage, sales at vending machines and shops, and elevators in the real estate 20 that is the observation target. This is a sensor that observes various phenomena such as the number of people entering and exiting a hall, entrance, office, or conference room, and is equipped with a communication function with the data providing device 200. The observation devices 30 for such phenomena include, for example, elevator cage weight sensors, people counters using infrared rays, etc., usage fee meters for infrastructure facilities, payment machines for vending machines and shops, etc., which are operated daily by the management company of the real estate 20 for maintenance management. Equipment, etc. can be assumed.

Furthermore, in the storage device 201, at least observation data 225 is stored in addition to a program 202 (including the machine learning engine 110) for implementing functions necessary for the data providing device 200 of this embodiment.

This observation data 225 becomes the source data of the observation data DB 125 held by the tenant departure prediction device 100. Therefore, the observation data 225 will be specifically explained in the explanation regarding the observation data DB 125.

FIG. 4 shows an example of the hardware configuration of the owner device 300 of this embodiment. The owner device 300 of this embodiment includes a storage device 301, a memory 303, an arithmetic device 304, and an input device 30.
5, an output device 306, and a communication device 307.

Among these, the storage device 301 is composed of an appropriate nonvolatile storage element such as an SSD (Solid State Drive) or a hard disk drive.

Furthermore, the memory 303 is composed of a volatile storage element such as a RAM.

Further, the arithmetic device 304 is a CPU that reads a program 302 held in the storage device 301 to the memory 303 and executes it, performs overall control of the device itself, and performs various judgments, calculations, and control processing.

Further, the input device 305 is an appropriate device such as a keyboard, mouse, microphone, etc. that accepts key input and voice input from the user.

Further, the output device 306 is an appropriate device such as a display or a speaker that displays the data processed by the arithmetic device 304.

Further, the communication device 307 is a network interface card that connects to the network 1 and handles communication processing with other devices such as the tenant departure prediction device 100.

Note that the storage device 301 stores at least a move-out case DB 325 and a business results DB 326 in addition to a program 302 for implementing functions necessary for the owner device 300 of this embodiment.

Among these databases, the eviction case DB 325 will be described later along with the explanation of the eviction case DB 126 held by the tenant eviction prediction device 100. Further, details of the sales result DB 326 will be described later.
<<Data structure example>>
Next, information used by the tenant eviction prediction device system 100 of this embodiment will be explained. FIG. 5 shows an example of the observation data DB 125 in this embodiment.

The observation data DB 125 of this embodiment is a database in which observation data 225 distributed from the data providing device 200 is accumulated.

The data structure is a collection of records consisting of data such as observation period and observed values, with identification information and events of the real estate to be observed as keys.

FIG. 6 shows an example of the withdrawal case DB 126 in this embodiment. The move-out case DB 126 of this embodiment is, for example, a database in which data from the move-out case DB 325 managed by the owner device 300 is acquired and accumulated.

The data structure uses the identification information of the real estate as a key, and includes the identification information of the tenant who left the property, the time of departure, the circumstances of the departure (e.g., leaving due to a request for an increase in floor space, leaving due to a request for a reduction in floor space, facilities, services, and location). Leaving due to dissatisfaction, leaving due to business closure, etc.), tenant attributes (e.g. industry type, business type, company size, number of employees, busy season, fiscal year-end, financial situation, average age of employees, average annual income, average length of service, etc.), eviction It is a collection of records consisting of data such as attributes of real estate (e.g. building structure, number of floors, floor area, rent per tsubo, equipment, age, nearest station, distance from the nearest station, etc.).

FIG. 7 shows an example of the model DB 127 in this embodiment. Model DB of this embodiment
Reference numeral 127 denotes a database that has accumulated prediction models of signs of tenant leaving, generated by the machine learning engine 110.

The data structure is a collection of records consisting of data such as applicable attributes (eg, real estate attributes, tenant attributes), and (entities of) models, using a model ID that uniquely identifies a model as a key.

FIG. 8 shows an example of the analysis result DB 128 in this embodiment. The analysis result DB 128 of this embodiment is a database that accumulates information on signs of tenant eviction in real estate predicted by the above-described model.

The data structure is a collection of records consisting of data such as the target real estate and target tenant as keys, information on signs of eviction predicted by the above-mentioned model (prospective eviction history), and actual eviction results.

FIG. 9 shows an example of the recommendation table 129 in this embodiment. The recommendation table 129 of this embodiment is a table that defines the content of recommended business activities in accordance with the signs of departure, that is, the expected circumstances of departure, predicted by the model.

The data structure is a collection of records consisting of data such as sales activities corresponding to the expected departure history, with the expected departure history as a key. Information on business activities specified based on this recommendation table 129 is notified to the owner device 300 of the real estate where the tenant is expected to leave, and can serve as a reference for actual business activities such as following up on tenants and developing new tenants. Furthermore, it is even more preferable that the information on sales activities includes extraction criteria for other real estate properties (described as matching candidate extraction processing summary in the diagram) to be presented during sales activities for the tenant.

FIG. 10 shows an example of the sales result DB 325 in this embodiment. The sales result DB 325 of this embodiment is a database that accumulates information on the results of the sales activities for the above-mentioned tenants.

The data structure is a collection of records consisting of data such as the target real estate and tenant as keys, reference sales activity information (identified from the recommendation table 129 above), and sales activity results. .

The results of sales activities accumulated here can serve as a basis for determining the appropriateness of sales activities based on signs of departure. Therefore, it depends on whether the result of this sales activity was favorable, such as successfully following up with the tenant in response to signs of eviction and avoiding cancellation, or unfavorable, such as not being able to avoid such cancellation. This can be used as an opportunity to review the contents of the recommendation table 129.
<<Flow example: Learning model generation>>
Hereinafter, the actual procedure of the tenant leaving prediction method in this embodiment will be explained based on the drawings. Various operations corresponding to the tenant eviction prediction method described below are realized by a program that the tenant eviction prediction device 100 reads into a memory or the like and executes. This program is composed of codes for performing various operations described below.

FIG. 11 is a diagram illustrating a flow example of the tenant leaving prediction method in this embodiment. Here, first, the tenant eviction prediction device 100 acquires observation data from the data providing device 200, and stores this in the observation data DB 125 (s10).

Furthermore, the tenant eviction prediction device 100 acquires eviction case information from the owner device 300 and stores this in the eviction case DB 126 (s11).

Next, the tenant eviction prediction device 100 inputs the data stored in the observation data DB 125 and the eviction case DB 126 to the machine learning engine 110, thereby creating a model that defines the relationship between the event indicated by the observation data and tenant eviction. is generated (s12).

In this case, the above-mentioned phenomena include the number and weight of people getting on and off the elevator, the amount of usage of various infrastructure such as electricity, gas, water and sewage, sales at vending machines and shops, elevator hall and Assume various events such as the number of people entering and leaving the entrance, office, and conference room.

In addition, the relationship between such events and tenant departures includes, for example, the rate of increase/decrease in the observed value of the event in a unit period, the presence or absence of the observed value, and the status of the tenant who moves into the real estate in the relevant state (a certain period has passed since the observation). (later) withdrawal probability. Note that if the eviction case information given to the machine learning engine 110 when generating the above-mentioned model includes the eviction process in addition to whether or not the tenant vacates, the model will calculate the relationship between the events shown by the observed data and the tenant eviction process. A specified model will be generated.

Note that when generating the above-mentioned model, a process may be performed in advance to classify the eviction case information by tenant attribute and to classify the observed data by real estate in which a tenant with this tenant attribute resided.

In that case, the tenant eviction prediction device 100 can generate a model for each tenant attribute by providing the eviction case information and observed data obtained in each classification to the machine learning engine 110 as input. For example, models are generated for each attribute such as industry type, business type, company size, number of employees, busy season, fiscal year end, financial situation, average age of employees, average annual income, and average length of service.

Alternatively, when generating the above-mentioned model, a process may be performed in advance to classify observation data for each real estate attribute, and to classify eviction case information for each tenant who lived in a real estate with this real estate attribute.

In that case, the tenant eviction prediction device 100 can generate a model for each real estate attribute by providing the eviction case information and observed data obtained in each classification to the machine learning engine 110 as input. For example, models are generated for each attribute such as building structure, number of floors, floor area, rent per tsubo, facilities, age, nearest station, and distance from the nearest station.

Such a plurality of model generation methods can be assumed to be specified in advance by the operator of the tenant eviction prediction device 100 or the owner device 300, and selectively executed by the tenant eviction prediction device 100 according to the designation. Alternatively, it is also conceivable that the tenant eviction prediction device 100 selectively executes the process depending on the presence or absence of an item, ie, an attribute, included in observation data or eviction case information.

Subsequently, the tenant eviction prediction device 100 stores the model generated in s12 in the model DB 126 (s13), and ends the process.
<<Flow example: Determining signs of eviction, etc.>>
FIG. 12 is a diagram illustrating a flow example of the tenant leaving prediction method in this embodiment. Here, a process for identifying signs of tenant eviction using the model generated in the above-described flow will be described.

In this case, the tenant departure prediction device 100 receives, for example, from the owner device 300 a designation (screen 1300 in FIG. 13) regarding the tenant or real estate to be predicted (s20). This designation includes tenant or real estate identification information or attribute designation.

Furthermore, the tenant departure prediction device 100 extracts observation data regarding the target real estate or the real estate in which the target tenant is occupying from the observation data DB 125 based on the specification received in s20 (s21).

Next, the tenant eviction prediction device 100 identifies signs of tenant eviction in the real estate by applying the observed data extracted in s21 to the model held in the model DB 126 (s22), and transmits this to the owner device 300. A response is made (screen 1400 in FIG. 14). This eviction sign includes at least one of the probability of eviction and the expected circumstances of eviction.

Among these, the expected reasons for leaving include leaving due to a desire to increase floor space, leaving due to requesting a reduction in floor space, leaving due to dissatisfaction with facilities, services, or location, and leaving due to going out of business.

In addition, as a preliminary study, the applicant used approximately 150 actual eviction cases and verified whether it was possible to capture the remaining 25% of the signs using machine learning with 75% of them as correct data. We succeeded in detecting signs of eviction and issuing an alarm with a probability of 6 months before the receipt of the cancellation notice.

Furthermore, the tenant eviction prediction device 100 compares the expected eviction history included in the information on the eviction sign obtained in s22 with the recommendation table 129, specifies guidelines for business activities for tenants who are expected to vacate based on the eviction sign, and Information on the guideline (screen 1500 in FIG. 15) is responded to the owner device 300 (s23), and the process ends.

Although the best mode for carrying out the present invention has been specifically described above, the present invention is not limited thereto and can be modified in various ways without departing from the gist thereof.

According to this embodiment, it is possible to efficiently detect signs of cancellation by a tenant. This enables efficient sales activities aimed at ensuring the occupancy rate of real estate properties, such as shortening the lead time between canceling a rental contract with a tenant and acquiring a rental contract with a new tenant. It can be said. In other words, it is possible to detect signs of tenant eviction and support appropriate responses accordingly.

The description of this specification clarifies at least the following. That is, in the tenant eviction prediction device of the present embodiment, when generating the model, the calculation device performs a process of classifying the eviction case information for each tenant attribute, and a process for classifying the eviction case information for each tenant attribute, and for each real estate in which a tenant with the tenant attribute has moved. Then, the model for each tenant attribute is generated by executing a process of classifying the observed data, and providing the eviction case information and the observed data obtained in each of the classifications to the machine learning engine as input. , when identifying the signs of eviction, applying observed data regarding the phenomenon in the real estate where a tenant with a predetermined attribute resides to the model according to the attribute; It is also possible to specify the .

According to this, signs of eviction are determined based on various tenant attributes such as industry, type of business, company size, number of employees, busy season, fiscal year end, financial situation, average age of employees, average annual income, average years of service, etc. It becomes possible to generate a model that specifies the In turn, it will be possible to detect signs of tenant eviction with higher accuracy and support appropriate responses accordingly.

Furthermore, in the tenant eviction prediction device of the present embodiment, when generating the model, the calculation device performs a process of classifying the observation data for each real estate attribute, and a process for classifying the observation data for each real estate property with the real estate attribute. , execute a process of classifying the eviction case information, and provide the eviction case information and observation data obtained in each classification to the machine learning engine as input, thereby generating the model for each real estate attribute. , when identifying the signs of eviction, the observed data regarding the phenomenon in the real estate with a predetermined attribute is applied to the model according to the attribute, thereby identifying the sign of tenant eviction in the real estate with the property attribute. It may be said that it is.

According to this, a model is generated that identifies signs of eviction based on various real estate attributes such as building structure, number of floors, floor area, rent per tsubo, equipment, age, nearest station, distance from nearest station, etc. It becomes possible. In turn, it will be possible to detect signs of tenant eviction with higher accuracy and support appropriate responses accordingly.

Furthermore, in the tenant eviction prediction device of the present embodiment, the arithmetic device further acquires eviction case information including at least the history of the tenant's eviction from the real estate from the real estate management device and stores it in the storage device. When the model is generated, the observation data and the eviction case information including the eviction process are input to the machine learning engine, thereby specifying the relationship between the events indicated by the observation data and the tenant eviction process. By applying observation data related to the phenomenon in a predetermined real estate to the model, the process of generating a model that generates a model may be applied to the model, thereby specifying the likely eviction history of the tenant in the real estate as the sign of eviction.

According to this, a model is created that identifies signs of eviction according to various eviction circumstances, such as eviction due to a desire to increase floor space, eviction due to a desire to reduce floor space, eviction due to dissatisfaction with facilities, services, or location, eviction due to going out of business, etc. It becomes possible. In turn, it will be possible to detect signs of tenant eviction with higher accuracy and support appropriate responses accordingly.

Further, in the tenant eviction prediction device of the present embodiment, the storage device holds guideline information for sales activities for the tenant for each of the expected eviction circumstances, and the arithmetic device stores the identified expected eviction circumstances in the guidelines. It may also be possible to further execute a process of collating the information to identify guidelines for business activities for the tenant, and outputting information on the guidelines.

According to this, guidelines regarding sales activities such as proposing another floor or other section of the same real estate, or starting the search for a new tenant in anticipation of the tenant's departure, are set out according to the circumstances (prospects) of the tenant's departure. It will be possible to present the information to the company that operates the property. This can lead to efficient sales activities aimed at securing the occupancy rate of the property, such as shortening the lead time from canceling the rental contract with the tenant to signing a new contract with the next tenant. In turn, it will be possible to detect signs of tenant eviction with higher accuracy and support appropriate responses accordingly.

1 Network 10 Tenant eviction prediction system 20 Building 30 Observation device 100 Tenant eviction prediction device 101 Storage device 102 Program 103 Memory 104 Arithmetic device 105 Input device 106 Output device 107 Communication device 110 Machine learning engine 125 Observation data DB
126 Eviction case DB
127 Model DB
128 Analysis result DB
129 Recommendation table 200 Data providing device 201 Storage device 202 Program 203 Memory 204 Arithmetic device 205 Communication device 225 Observation data 300 Owner device 301 Storage device 302 Program 303 Memory 304 Arithmetic device 305 Input device 306 Output device 307 Communication device 325 Leaving case DB

Claims (5)

a storage device that holds observation data observed regarding a predetermined event in the real estate, eviction case information including at least the history of the tenant's eviction from the real estate, and guideline information for business activities for the tenant for each of the expected eviction history ;
The observation data and the eviction case information including the eviction process are input to a machine learning engine to generate a model that defines the relationship between the events indicated by the observation data and the tenant eviction process , and a predetermined process. By applying observed data regarding the above-mentioned events in the real estate to the model, the expected eviction history of the tenant in the real estate is identified as a sign of eviction , and the identified eviction history is compared with the guideline information to determine the expected eviction history for the tenant. a computing device that performs processing for identifying guidelines for business activities and outputting information on the guidelines ;
A tenant eviction prediction device comprising: The arithmetic device is
When generating the model, a process of classifying the eviction case information for each tenant attribute and a process of classifying the observed data for each real estate in which a tenant with the tenant attribute was occupied are executed, and each of the above classifications is performed. Generate the model for each tenant attribute by giving the eviction case information and the observation data obtained in to the machine learning engine as input,
When identifying the signs of eviction, the observed data regarding the phenomenon in the real estate where the tenant with the predetermined attribute moves is applied to the model according to the attribute, thereby identifying the sign of the eviction of the tenant with the attribute in the real estate. It specifies
The tenant eviction prediction device according to claim 1, characterized in that: The arithmetic device is
When generating the model, a process of classifying the observed data for each real estate attribute and a process of classifying the eviction case information for each tenant who lived in the real estate of the real estate attribute are executed, and each of the above classifications is performed. Generating the model for each real estate attribute by giving the eviction case information and the observed data obtained in to the machine learning engine as input,
When identifying the signs of eviction, the observed data regarding the phenomenon in real estate with a predetermined attribute is applied to the model according to the attribute, thereby identifying the sign of tenant eviction in real estate with the property attribute. be,
The tenant eviction prediction device according to claim 1, characterized in that: The information processing device
Observation data observed regarding predetermined events in the real estate, eviction case information including at least the eviction history of tenants in the real estate, and guideline information for sales activities for tenants for each of the eviction history are retained in a storage device. ,
The observation data and the eviction case information including the eviction process are input to a machine learning engine to generate a model that defines the relationship between the events indicated by the observation data and the tenant eviction process , and a predetermined process. By applying observed data regarding the above-mentioned events in the real estate to the model, the expected eviction history of the tenant in the real estate is identified as a sign of eviction , and the identified eviction history is compared with the guideline information to determine the expected eviction history for the tenant. A process of identifying guidelines for business activities and outputting information on the guidelines ;
A tenant eviction prediction method characterized by performing the following steps. a storage device that holds observation data observed regarding a predetermined event in the real estate, eviction case information including at least the history of the tenant's eviction from the real estate, and guideline information for business activities for the tenant for each of the expected eviction history ;
The observation data and the eviction case information including the eviction process are input to a machine learning engine to generate a model that defines the relationship between the events indicated by the observation data and the tenant eviction process , and a predetermined process. By applying observed data regarding the above-mentioned events in the real estate to the model, the expected eviction history of the tenant in the real estate is identified as a sign of eviction , and the identified eviction history is compared with the guideline information to determine the expected eviction history for the tenant. a computing device that performs processing for identifying guidelines for business activities and outputting information on the guidelines ;
A tenant eviction prediction system comprising a tenant eviction prediction device.

Images