JP7109804B2 - Output program, output method and output device - Google Patents

Description

The present invention relates to an output program and the like for outputting property information of real estate.

A system has been proposed that provides property information (housing information) suitable for tenants who want to rent real estate (for example, Patent Document 1). However, if the borrower does not input their own wishes as accurate search conditions, there is a possibility that properties that do not match the search will be included in the search hits.

On the other hand, a system has been proposed that introduces suitable real estate properties by using a record of contract conclusions (for example, Patent Literature 2). The system described in Patent Literature 2 associates a user with one of a plurality of clusters, and outputs contract real estate included in contract results classified into the associated cluster. Since the properties output correspond to clusters, the user's specific preferences are not taken into account.

JP 2003-281175 A JP 2016-206783 A

The present invention has been made in view of such circumstances. The purpose is to provide an output program or the like for outputting the contract probability of a property in order to select property information of real estate suitable for a borrower .

An output program according to the present invention acquires desired conditions for a property of a borrower, acquires property information of a property that matches at least a part of the acquired desired conditions, attributes of the borrower and the acquired desired conditions, and The obtained property information is input into a decision tree in which the attribute of the borrower, the desired conditions for the property of the borrower, and the property information are used as explanatory variables, and the probability of closing the property is the objective variable, and the contract is closed for the property. It is characterized by causing a computer to perform a process of acquiring the probability and outputting the acquired contract probability.

According to the present invention, it is possible to output the contract probability of the property in order to select the property information of the real estate suitable for the borrower .

1 is an explanatory diagram showing a configuration example of a provision system; FIG. 3 is a block diagram showing a hardware configuration example of a providing server; FIG. FIG. 4 is an explanatory diagram showing an example of a property DB; FIG. 10 is an explanatory diagram showing an example of a basic room DB; It is explanatory drawing which shows the example of basic-equipment DB. FIG. 4 is an explanatory diagram showing an example of additional information DB; It is explanatory drawing which shows the example of contract conditions DB. FIG. 4 is an explanatory diagram showing an example of a peripheral DB; It is explanatory drawing which shows the example of special mention DB. It is explanatory drawing which shows the example of facility DB. It is explanatory drawing which shows the example of reply DB. It is explanatory drawing which shows the example of prospective customer DB. FIG. 4 is an explanatory diagram showing an example of a history DB; FIG. 4 is an explanatory diagram showing an example of an additional information DB; FIG. 10 is an explanatory diagram of processing for generating a room selection model; FIG. 11 is an explanatory diagram relating to room selection processing; 10 is a flow chart showing an example of the procedure of room selection model generation processing; 6 is a flow chart showing an example of the procedure of room selection processing; FIG. 11 is an explanatory diagram showing an example of a room selection result display screen; FIG. 11 is a flowchart showing an example of a procedure for preview room selection processing; FIG. FIG. 11 is an explanatory diagram showing an example of a preview room selection result display screen; It is explanatory drawing which shows the example of performance data DB. 10 is a flow chart showing an example of the procedure of decision tree generation processing; 9 is a flow chart showing an example of the procedure of performance data creation processing; 4 is a flowchart showing the procedure of decision tree analysis processing; FIG. 4 is an explanatory diagram showing an example of a decision tree; FIG. 11 is a flowchart showing another procedure example of room selection processing; FIG. FIG. 11 is a flowchart showing another procedure example of room selection processing; FIG. FIG. 11 is a flowchart showing another procedure example of room selection processing; FIG. 6 is a flow chart showing an example of the procedure of room search processing; FIG. 11 is an explanatory diagram showing an example of a search result display screen; 6 is a flow chart showing an example of the procedure of room search processing;

(Embodiment 1)
Embodiments will be described below with reference to the drawings. FIG. 1 is an explanatory diagram showing a configuration example of a provision system. The provision system 100 includes a provision server 1 , sales terminals 2 and user terminals 3 . The providing server 1, the sales terminal 2, and the user terminal 3 are connected by a network N so as to be able to communicate with each other. The providing server 1 mainly provides information on real estate rental properties. A sales terminal 2 is a terminal used by a salesperson of a real estate agent. The sales terminal 2 acquires property conditions, transmits the acquired property information to the provision server 1 , and receives property information from the provision server 1 . A user terminal 3 is a terminal used by a general user looking for a real estate property. Although three sales terminals 2 and three user terminals 3 are shown in FIG. 1, the present invention is not limited to this. The sales terminal 2 may be one or more. One or more user terminals 3 may be provided. However, if the providing server 1 does not provide information to general users, the user terminal 3 is not included in the providing system 100 .

FIG. 2 is a block diagram showing a hardware configuration example of a providing server. The providing server 1 is composed of a workstation, a server computer, a blade server, and the like. The providing server 1 ( ) includes a control unit 11 , a main storage unit 12 , an auxiliary storage unit 13 , a communication unit 15 and a reading unit 16 . Each section is connected by a bus B. Note that the functions of the providing server 1 may be realized by a cloud service. Also, the control device may be composed of a multicomputer composed of a plurality of computers, a virtual machine or a quantum computer virtually constructed by software.

The control unit 11 has one or more arithmetic processing units such as a CPU (Central Processing Unit), an MPU (Micro-Processing Unit), and a GPU (Graphics Processing Unit). The control unit 11 reads and executes a control program 1P (output program) stored in the auxiliary storage unit 13 to perform various information processing, control processing, and the like related to the providing server 1 . The control unit 11 operates as a functional unit such as an acquisition unit or an output unit by executing the control program 1P.

The main storage unit 12 is SRAM (Static Random Access Memory), DRAM (Dynamic Random Access Memory), flash memory, or the like. The main storage unit 12 mainly temporarily stores data necessary for the control unit 11 to execute arithmetic processing.

The auxiliary storage unit 13 is a hard disk, SSD (Solid State Drive), or the like, and stores the control program 1P and various DBs (Databases) necessary for the control unit 11 to execute processing. Auxiliary storage unit 13 stores property DB 131, basic room DB 132, basic equipment DB 133, additional information DB 134, contract terms DB 135, peripheral DB 136, special mention DB 137, facility DB 138, reply DB 139, prospective customer DB 13A, history DB 13B, and additional information DB 13C. do. Also, the auxiliary storage unit 13 stores a room selection model 141 . The auxiliary storage unit 13 may be an external storage device connected to the providing server 1 . Various DBs and the like stored in the auxiliary storage unit 13 may be stored in the cloud storage.

The communication unit 15 communicates with the sales terminal 2 and the user terminal 3 via the network N. FIG. Alternatively, the control unit 11 may use the communication unit 15 to download the control program 1P from another computer via the network N or the like and store it in the auxiliary storage unit 13 .

The reader 16 reads portable storage media 1a including CD (Compact Disc)-ROM and DVD (Digital Versatile Disc)-ROM. The control unit 11 may read the control program 1P from the portable storage medium 1a via the reading unit 16 and store it in the auxiliary storage unit 13 . Alternatively, the control unit 11 may read the control program 1P from the semiconductor memory 1b.

FIG. 3 is an explanatory diagram showing an example of the property DB. The property DB 131 stores information on property buildings. The property DB 131 stores information on the collective housing when the rental property belongs to the collective housing. When the rental property is a detached house, the property DB 131 stores information on the detached house. The property DB 131 includes a property ID column, a property name column, a property address column, an area column, a nearest station column, a precinct column, a management type column, a structure column, a building age column, and a common area column. The property ID column stores property IDs that can uniquely identify buildings. A property ID is issued when a new record is registered. The property name column stores the names of buildings. The property address column stores the location of the building. The area column stores the area of the location. The area may be determined and input by the salesperson, or may be uniformly determined based on the correspondence table prepared in advance between the town-chome and the area. The nearest station column stores the nearest station to the building and the means of transportation from the nearest station. The school district column stores the school district of public elementary schools and the school districts of public junior high schools corresponding to the location of the building. Specifically, the names of elementary schools and junior high schools are stored. The management type column stores the management type of the building. The management type is, for example, owned by the company, managed by the company, collectively leased, or general. Owned by the company indicates that the building owned by the real estate company is managed by the company itself. In-house management indicates that the property is managed by a real estate company under consignment. Collective lease indicates that the entire building is collectively leased from the building owner and managed. The structure column stores the structure of the building. The structure of the building is, for example, wooden, steel, lightweight steel, or RC. Wooden construction refers to structures that use wood for the main structural parts of the building, such as columns, beams, walls, and floors. Steel frame construction refers to a structure that uses a steel frame for the framework of a building. Steel frame construction is also called heavy steel construction. Light steel construction refers to a steel construction that uses steel materials with a thickness of less than 6 mm. RC construction is also called reinforced concrete construction. RC structure refers to a structure in which reinforcing bars are embedded in concrete. The building age column stores the building age of the building. Building age refers to the number of years that have passed since the building was completed. If the building is an apartment complex, the common area column stores facilities shared by residents. The information about the property building is also called property building information.

FIG. 4 is an explanatory diagram showing an example of the basic room DB. The basic room DB 132 stores basic information about rooms (basic room information). If the building is an apartment complex, the basic room DB 132 stores basic information about each dwelling unit. If the building is a detached house, the basic room DB 132 stores information on the building that is not included in the property DB 131 . The basic room DB 132 includes a room ID column, a dwelling unit number column, a floor plan column, a property ID column, a guarantee category column, a rent column, an administration cost column, a vacant room period column, and a situation column. The room ID column stores room IDs that can uniquely identify rooms. A room ID is issued when a new record is registered. The housing unit number column stores the housing unit number. The house number is also called the room number. If the building is a detached house, the house number is not stored. The floor plan column stores a character string representing the floor plan. The floor plan is, for example, 2LDK, 3DK, and the like. The property ID column stores the property ID of the corresponding building. The guarantee classification column stores whether or not it is necessary to join a rent guarantee company and conclude a rent guarantee contract. If a subscription is required, the warranty category column stores the subscription. If subscription is not required, the warranty category column stores "not required". A rent guarantee company is also called a rent guarantee company. The rent column stores the monthly rent for the room. The maintenance cost column stores the monthly maintenance cost of the room. The vacancy period column stores the vacancy period when the room is vacant. The status column stores information about room occupancy. For example, the status column stores the following values. If there is a resident in the room, the status column stores occupied. If there is a resident in the room but is scheduled to move out, the status column stores the move-out schedule.

FIG. 5 is an explanatory diagram showing an example of the basic equipment DB. The basic equipment DB 133 stores information (basic equipment information) on the basic equipment of the room. The basic facility DB 133 includes a room ID column, a floor plan detail column, an entrance column, and a bathroom column. The room ID column stores room IDs. The floor plan details column stores the details of the room layout. FIG. 5 shows a detailed example of a 2DK floor plan. DK6.4 indicates that the dining kitchen is 6.4 tatami mats wide. Western-style room 6 indicates that one of the two rooms is a Western-style room of 6 tatami mats. Japanese-style room 7 indicates that the other room is a Japanese-style room of 7 tatami mats. The porch row remembers the facility of the porch. The entrance row includes a wide row, an auto-lock row, a TV row, and a shoe cupboard row. The width column stores the width of the hallway. The auto-lock queue stores whether or not the common building entrance is auto-locked. If the building entrance is auto-locking, store the auto-locking queue yes. If the building entrance is not auto-locking, store no auto-locking queue. The TV column stores whether or not an intercom with a TV monitor is installed at the entrance of the room. When the interphone with a TV monitor is installed, the presence of the TV channel is stored, and when the intercom with the TV monitor is not installed, the absence of the TV channel is stored. The shoe rack row stores whether or not a shoe rack is provided at the entrance. When the shoe rack is installed, the shoe rack row is stored, and when the shoe cabinet is not installed, the shoe cabinet row is not stored. The plumbing column stores plumbing facilities. The plumbing line includes a gas line, a washing pan line, a reheating line, a thermo faucet line, a shower toilet line, a bathroom dryer line, a kitchen hot water supply line, and a BT line. A gas string stores the gas that is being supplied to the room. If city gas is supplied, the gas train stores city gas. If propane gas is supplied, the gas train stores propane gas. The laundry pan column stores the size of the laundry pan. For example, if the size of the laundry pan is 640 mm wide×640 mm deep, 640 are stored as the laundry pan row. When the size of the laundry pan is 740 mm wide×640 mm deep, 740 are stored as the laundry pan row. When the size of the laundry pan is 800 mm wide×640 mm deep, 800 sizes are stored in the row of laundry pans. The reheating column stores whether or not the bathtub has a reheating function. If there is a reheating function, the reheating column stores yes. When there is no reheating function, no reheating column is stored. The thermofaucet column stores whether or not the bathroom faucet is a thermostatic faucet. If the bathroom faucet is a thermostatic faucet, the thermofaucet string stores Yes. If the bathroom faucet is not a thermostatic faucet, store No Thermofaucet Row. A thermostat type faucet mixes hot and cold water, and when the temperature is set with a temperature control handle, the amount of hot and cold water mixed is automatically adjusted by the function of a thermostat cartridge. As a result, it is possible to supply hot water at a stable temperature without hot water becoming hot or cold. The shower-toilet column stores whether or not the toilet has a shower-toilet function. If the shower/toilet function is on, the shower/toilet row stores "Yes". When the shower toilet function is not provided, no shower toilet row is stored. The bathroom dryer column stores whether or not the bathroom is equipped with a clothes dryer. If a clothes dryer is installed, the bathroom dryer row stores yes. If no clothes dryer is installed, store No Bathroom Dryer Row. The kitchen hot water supply column stores whether or not hot water is supplied in the kitchen. If hot water is available in the kitchen, the kitchen hot water supply column stores yes. If there is no hot water in the kitchen, the no kitchen hot water supply column is stored. The BT column stores whether or not the bathroom and toilet are separate rooms. If the bathroom and the toilet are separate rooms, the BT column stores the difference. In the case of a three-point unit bath in which the bathroom and the toilet are integrated, the BT row stores the unity.

FIG. 6 is an explanatory diagram showing an example of the additional information DB. Additional information DB 134 stores additional information about rooms. In the case of collective housing, the additional information also includes information regarding the building. The additional information DB 134 includes a room ID column, an EL column, a pet column, a musical instrument column, a net column, a storage column, an outdoor warehouse column, a parking lot column, and a bicycle parking lot column. The room ID column stores room IDs. The EL column stores whether or not there is an elevator in the case of an apartment house. If there is an elevator, the EL column stores Yes. If there is no elevator, the EL column stores None. The pet column stores whether or not pets such as dogs and cats can be kept. If pets can be kept, the pet column stores "Yes". If pet keeping is prohibited, the pet column stores "impossible". The instrument column stores whether musical instruments are allowed to be played in the room. If playing musical instruments is allowed in the room, the musical instrument column stores yes. If musical instruments are not allowed to be played in the room, the musical instrument column stores no. The net column stores whether or not the internet line is connected. If the internet line is hooked up, the net queue remembers yes. When the Internet line is connected and no additional usage is generated, the net line stores free of charge. If the internet line is not hooked up, the net queue stores no. The storage column stores the number of storage spaces such as closets and closets. If there is no storage, the storage row stores 0 or none. The outdoor warehouse column stores whether or not there is an outdoor warehouse available to the resident. If there is an outdoor warehouse, it stores that there is an outdoor warehouse column. If there is no outdoor warehouse, "None" is stored in the outdoor warehouse column. The parking lot row stores the conditions for renting the parking lot at the same time. For example, if the monthly fee is 3000 yen, the parking lot row stores 3000 yen (one car). If the rent includes parking for 2 cars, the parking row stores 2 cars included. The bicycle parking lot column stores information such as the presence or absence of a bicycle parking lot. When the rent includes the usage fee, the bicycle parking line is included. If the usage fee for the bicycle parking lot is 250 yen per month, 250 yen is stored for the bicycle parking line. When there is no bicycle parking lot, no bicycle parking row is stored.

FIG. 7 is an explanatory diagram showing an example of the contract terms DB. The contract terms DB 135 stores the main terms (contract terms) of the tenancy contract. The contract terms DB 135 includes a room ID column, a free rent column, an immediate move-in column, a security deposit column, a key money column, and a guarantor column. The room ID column stores room IDs. The free rent column stores whether or not a free rent for which rent is free for a certain period of time is included. If free rent is included, the free rent column stores the period of free rent. The immediate move-in column stores whether or not the move-in is immediate. If immediate occupancy is possible, the immediate occupancy column stores "possible". If immediate occupancy is not possible, the immediate occupancy column stores "impossible". The security deposit column stores security deposits to be paid at the time of move-in. For example, if the security deposit is one month's rent, the security deposit column stores one month. If the security deposit is not per rent, the security deposit column stores the specific amount. The key money string stores the key money to be paid at the time of moving in. The contents stored in the key money series are the same as those in the security deposit series. The guarantor column stores whether or not it is necessary to set up a joint guarantor at the time of contract. If it is necessary to set up a joint guarantor, the guarantor column stores the necessity. If there is no need to set up a joint guarantor, the guarantor column is stored as unnecessary.

FIG. 8 is an explanatory diagram showing an example of the peripheral DB. The peripheral DB 136 stores information (environmental information) such as facilities around the property. The peripheral DB 136 includes a property ID column, a convenience store column, a highway column, a station column, a hospital column, and a supermarket column. The property ID column stores property IDs. The convenience store string stores information of a convenience store (hereinafter abbreviated as convenience store) located near the property. The convenience store string includes a distance string and a store string. The distance column stores the distance from the property to the nearest convenience store. The store column stores the name of the convenience store chain and the store ID. The trunk road column stores the information of the trunk road closest to the property. The trunk road column includes a distance column and a road name column. The distance column stores the distance from the object to the nearest highway. The road name column stores the names of trunk roads. The station train stores information about the nearest station to the object. The station column includes a distance column and a station name column. The distance column stores the distance from the object to the nearest station. The station name column stores route names and station names. The hospital column stores information on hospitals closest to the property. The hospital column includes a distance column and a name column. The distance column stores the distance from the property to the nearest hospital. The name column stores hospital names. The supermarket queue stores information about the supermarket closest to the property. The super-column includes a distance column and a store column. The distance column stores the distance from the property to the nearest supermarket. The store column stores the names of supermarkets and store IDs. It is desirable that the distances stored in the distance column are distances rather than straight-line distances. Both the route distance and the straight line distance may be stored in the distance column. Determination of whether or not there is a convenience store or the like nearby is performed by determining a threshold value. When determining the required time on foot, for example, the walking speed is set to 80 m/min, and the required time is calculated and determined from the distance. Further, it may be determined whether or not there is a facility such as a convenience store nearby based on the time required by car. The speed in the case of a car is, for example, 400m/min. A default value is set for the determination threshold. For example, within 10 minutes is defined as walking distance.

FIG. 9 is an explanatory diagram showing an example of the special mention DB. The special note DB 137 stores items (special notes) that the real estate agent considers to be particularly important for the property. The special note DB 137 includes a room ID column, a company management column, a renovation column, a long-term vacant room column, an initial cost reduction column, and a customer base column. The room ID column stores room IDs. The company management column stores whether or not the property is owned by a real estate agent. In the case of a room owned by a real estate agent (property owned by the company), ◯ is stored in the company management column. In the case of a room that is not owned by a real estate agent but is managed under consignment (property managed by the company), △ is stored in the company management column. In the case of a room that is only brokered by a real estate agent (property that the real estate agent neither entrusts management nor owns), X is stored in the own management column. The remodel column stores whether or not the room has been remodeled. When the room is remodeled, ◯ or applicable is stored in the remodeling column. If the room has not been remodeled, the remodel column stores x or not applicable. The long-term vacant room column stores whether or not the room is vacant for a long time. If the room is vacant for a long time, the long-term vacant room column stores ◯ or applicable. If the room is not vacant for a long period of time, x or non-applicable is stored in the long-term vacant room column. The criteria for judging whether or not a room is vacant for a long period of time are determined by the real estate agent as appropriate. The initial cost reduction column stores whether or not the room is subject to initial cost reduction. If the room is subject to initial cost reduction, ◯ or applicable is stored in the initial cost reduction column. If the room is not subject to the initial cost reduction, x or non-applicable is stored in the initial cost reduction column. Initial cost reduction means, for example, that by adding a small amount to the monthly rent, for example, about 1,500 yen, you do not need to pay the security deposit, key money, key exchange fee, and brokerage fee when you move in, and the house cleaning fee when you move out. It is. The customer class column stores the customer class suitable for the room or the customer class desired by the owner of the room. In addition, you may memorize|store the information of a psychological defect property in special mention DB137.

FIG. 10 is an explanatory diagram showing an example of the facility DB. The facility DB 138 stores detailed information (facility information) of facilities stored in the above-described peripheral DB 136 . The facility DB 138 includes a number column, a type column, a chain name column, a store number column, a store name column, a business hours column, a regular holiday column, and a point column. The number column stores sequential numbers that are primary keys. A number is issued when adding a record. The type column stores the type of facility. For example, the types are convenience stores, supermarkets, drug stores, family restaurants (meaning family restaurants), fast food, and the like. The chain name column stores chain names and common store names when the facility is a chain store or a store of a company that has multiple stores. The store number column stores numbers that can identify each store in the same chain. The store name column stores the unique name of each store. The business hours column stores the business hours of the store. The regular holiday column stores regular holidays of the store. In the case of a convenience store, 24 hours is stored in the business hours column, and 7 days a week in the regular holiday column. The point column stores the names of points that can be used at the store. Information on electronic money that can be used at the store may be stored in the facility DB 138 .

FIG. 11 is an explanatory diagram showing an example of an answer DB. The response DB 139 stores the content (response information) of the prospective customer's response to the questionnaire prepared by the real estate agent. The answers to the questionnaire may be entered by the salesperson using the sales terminal 2, or may be entered by the prospective customer using the user terminal 3. FIG. The response DB 139 includes a case ID column, a lower rent column, an upper rent column, an area column, a floor plan column, a nearest station column, a school district column, an auto lock column, a TV column, a BT column, a pet allowed column, a musical instrument allowed column, Including net free row and row with warehouse. The case ID column stores a case ID column that can uniquely identify a case. A case ID is issued when adding a new record. The Minimum Rent column stores the minimum rent desired by the prospective customer. The lower limit amount is divided into an amount range of 5,000 yen, and a numerical value is attached to each amount range. For example, 0: No lower limit, 1: 20,000 yen or more, 2: 25,000 yen or more, 3: 30,000 yen or more, ..., 13: 80,000 yen or more. Since it is 3 in FIG. 11, the customer wants 30,000 yen or more, as shown in parenthesis. The Rent Limit column stores the maximum rent desired by the prospective customer. As with the lower limit, a numerical value is attached to each amount range. For example, 1: 20,000 yen or less, 2: 25,000 yen or less, ..., 13: 80,000 yen or less, 14: no upper limit. Since it is 7 in FIG. 11, the customer desires 50,000 yen or less as shown in parenthesis. The area column stores customer preferences for areas. The area will be determined by the real estate agent as appropriate. FIG. 11 shows an example divided into six areas. ○ indicates the desired area. X indicates an area that is not desired. △ indicates an area that is allowed depending on other conditions. The floor plan column stores a desired floor plan. The floor plans are divided into groups, and numerical values are assigned to each group. For example, 1: 1R/1K/1DK, 2: 1LDK, 3: 2K/2DK, 4: 2LDK, 5: 3DK/3LDK, 6: 4DK or more, 7: others. Since it is 4 in the example shown in FIG. 11, it indicates that the prospective customer desires 2LDK. The nearest station column stores the desired nearest station. Although station names are stored in FIG. 11, station numbers, symbols, and the like may be used to specify stations instead of station names. The school district column stores elementary school districts and junior high school districts. Elementary school districts and junior high school districts respectively store, for example, corresponding elementary school names and junior high school names. The auto-lock column indicates whether or not the common entrance is desired to be equipped with an auto-lock when the desired property is a room in a collective housing. If you want to have auto-lock, store O in the auto-lock column. If auto-lock is desired, the auto-lock column stores Δ. If you do not specifically want to have auto-lock, the auto-lock column stores x. The TV line stores whether or not the user wishes to be equipped with an intercom with a TV monitor. The content stored by the TV row is similar to the auto-lock row. The separate BT column stores whether or not the user desires that the bathroom and toilet be separate rooms. The contents stored in the BT-by-BT column are the same as those in the auto-lock column. The pet-allowed column stores whether or not a property in which pets such as dogs and cats are permitted is desired. If you want a property that allows you to keep pets, ◯ is stored in the Pet Allowed column. As to whether or not pets are allowed, if no particular request is made, the pet allowed column stores Δ. If you want a property that does not allow pets, the Pets Allowed column stores an X. The musical instrument allowed column stores whether or not a property in which musical instruments are permitted in the room is desired. The contents stored in the instrument allowable column are the same as those of the pet allowable column. The free-of-net column stores whether or not a property with an internet connection and no additional usage is desired. The content stored by the net free queue is similar to the autolock queue. The row with warehouse stores whether or not the resident desires a property with an outdoor warehouse that can be used by the resident. The contents stored in the storage column are the same as those in the auto-lock column. The above case starts when a prospective customer looking for real estate visits a real estate agent and answers a questionnaire. Alternatively, it starts when the prospective customer accesses the web page of the real estate agent, answers the questionnaire on the questionnaire screen, and sends the answer. The case ends when the prospective customer makes an application, or when there are no more inquiries, such as when the prospective customer signs a contract with another real estate agent.

FIG. 12 is an explanatory diagram showing an example of the prospective customer DB. The prospective customer DB 13A stores information (personal information) about prospective customers. The prospective customer DB 13A includes a prospective customer ID column, a name column, a furigana column, an age column, a gender column, an occupational classification column, an email column, a telephone column, a FAX column, a postal code column, an address column, a contact method column, a faculty column, and a contract. Including planned person column, customer category column, occupant number column, family type column, foreign nationality column and Japanese column. The prospective customer ID column stores prospective customer IDs that can uniquely identify prospective customers. A prospective customer ID is issued when adding a new record. The name column stores the name of the prospective customer. The furigana column stores the furigana of the name. The Age column stores the age of the prospective customer. The Gender column stores the gender of the prospective customer. The occupational classification column stores the occupational classification of the prospective customer. Occupation categories are, for example, full-time employees, part-time workers, students, and others. The email column stores the prospective customer's contact email address. The phone column stores the contact phone number of the prospective customer. The FAX column stores the prospective customer's contact facsimile number. The Postal Code column stores the postal code of the prospective customer's current address. The address column stores the prospective customer's current address. The contact method column stores the prospective customer's preferred contact method. The contact method column stores, for example, phone, e-mail, and fax. The contact method column may store a number of prioritized contact methods. The Faculty column stores the faculty to which the prospective customer belongs if he is a college student. The prospective contract person column stores the information of the prospective contract person when the person who is expected to conclude the contract is other than the customer. For example, the prospective customer is a college student and the contractor is his guardian. The prospective contract person column includes an occupation column and an annual income column. The occupation column stores the occupation of the prospective contract. The annual income column stores the most recent annual income of the contract candidate. The customer segment stores the segment of the customer. When a corporation is the contractor, the customer category stores "corporation". If the resident is a student, the customer category stores "student". When a member of society signs a contract and moves in, "general" is stored. The number of residents column stores the number of people who are scheduled to move in. The family type column stores the type of family that is scheduled to move in. The family type column includes a household column, a children column, and an infant column. The household column stores household types. Household types include single-person households, husband-and-wife households, and husband-and-wife households with children. The children column stores the number of children. The infants column stores whether infants are included in the children. If infants are included, ◯, applicable or 1, etc. are stored in the infants column. If no infants are included, x, not applicable, or 0 is stored in the infants column. The foreign nationality column stores whether or not the prospective customer is of foreign nationality. If the prospective customer is of foreign nationality, 1 or ○ is stored in the foreign nationality column. If the prospective customer is of Japanese nationality, 0 or x or no value is stored in the foreign nationality column. The country of nationality may be stored in the foreign nationality column. For example, if the prospective customer has Chinese nationality, CN indicating China is stored in the foreign nationality column. The Japanese column stores the prospective customer's ability to communicate in Japanese. For example, according to the degree of communication ability, it is classified into three stages (○, △, ×). Further, the prospective customer DB 13A may store the religion, creed, ethnicity, and race of the prospective customer. However, this information is sensitive information and should be handled with care. Information about prospective customers is hereinafter also referred to as customer information.

FIG. 13 is an explanatory diagram showing an example of the history DB. History DB 13B stores activities for prospective customers. The history DB 13B includes a case ID column, a prospective customer ID column, a reference property column, a viewing property column, a status column, and a result column. The case ID column stores case IDs. The prospective customer ID column stores prospective customer IDs. The reference property column stores the room ID of the room referred to by the prospective customer. The room ID of the room listed on the property sheet presented to the prospective customer at the store by the salesperson is stored. Further, the room ID of the room referred to by the prospective customer may be obtained from the history of the prospective customer's referral to the property introduction Web page provided by the real estate agent and stored. The preview property column stores the room ID of the room previewed by the prospective customer. The preview property column includes preview 1 column, preview 2 column, preview 3 column, and other columns. In many cases, the salesperson presents three properties to be previewed, so the three room IDs are stored in the preview 1 column, preview 2 column, and preview 3 column. If more than 3 rooms have been previewed, the remaining room IDs are stored in the others column. Nairan refers to the actual observation and investigation of a real estate property, and is also called a preview. The status column stores the status of the case. For example, the status is in progress, preview completed, customer follow-up, finished, lost order. In progress indicates the stage before previewing. "Previewed" indicates that the property has been viewed but has not yet been applied for. Purchasing is the stage after contacting prospective customers who have seen the property but have not yet applied for it by telephone or e-mail. End indicates that the application has been reached. A lost order indicates that it has been confirmed that the application will not be made. The results column stores the status or results for closed cases. The result columns include a contract column, a room ID column, and a contract rent column. The contract column indicates the status of the contract. For example, the status is under review, waiting for payment, and closed. During examination indicates that the tenant examination is being performed. Waiting for payment indicates a situation in which the customer is waiting for payment of security deposit, key money, initial rent, and the like. Closing indicates that the conclusion of the contract has been completed. The room ID column stores room IDs of rooms for which an application has been made or a contract has been concluded. The contracted rent column stores the contracted rent.

FIG. 14 is an explanatory diagram showing an example of the additional information DB. The additional information DB 13C stores additional information about prospective customers obtained by sales representatives. The additional information DB 13C shows a case ID column, a current house structure column, a working hours column, a self-cooking column, a property column, a holiday column, and a bathing time column. The case ID column stores case IDs. The current dwelling structure column stores the dwelling structure in which the prospective customer is currently living. The Hours of Work column stores the hours that the prospective customer is working. The working hours column includes a start column and an end column. The start column stores the work start time. The end column stores the work end time. If the prospective customer is a student or the like, the time period during which the prospective customer is at school or the like is stored. The self-cooking column stores whether or not to cook for oneself, and if so, the day of the week, time zone, etc. for self-cooking. The property column stores the property if the prospective customer owns something that needs to be considered when choosing a room. For example, the property column stores drum washing machines. The holiday column stores holidays of prospective customers. If the prospective customer is not working, the holiday column stores an empty string, blank, or the like. The bath time column stores the time the prospective customer takes a bath. As described above, the information stored in the additional information DB 13C includes information that has little relation to room selection conditions, but stores additional information obtained by the salesperson in order to grasp the lifestyle of the prospective customer.

FIG. 15 is an explanatory diagram relating to the room selection model generation process. FIG. 15 conceptually illustrates the process of performing machine learning to generate the room selection model 141 . The process of generating the room selection model 141 will be described with reference to FIG.

The providing server 1, as a room selection model 141, based on past contract information, deep do the learning. The contracted property information includes property building information stored in the property DB 131 , basic room information stored in the basic room DB 132 , and basic facility information stored in the basic facility DB 133 . Additional information stored in the additional information DB 134, contract conditions stored in the contract condition DB 135, peripheral information stored in the peripheral DB 136, special notes stored in the special note DB 137, facilities stored in the facility DB Information (however, limited to information relating to facilities included in the surrounding information) may be included in the contracted property information. Personal information includes customer information stored in the prospective customer DB 13A. The personal information may include additional information stored in the additional information DB 13C. Through learning, a neural network is generated that receives information about customers as input and outputs information about rooms that are suitable for customers.

The neural network is, for example, a CNN (Convolution Neural Network), and includes an input layer that receives input of information about rooms and information about customers, an output layer that outputs information about rooms suitable for customers, and information about rooms and information about customers. and an intermediate layer that associates the

The input layer has a plurality of neurons that receive as inputs values of a plurality of data items included in customer-related information, and passes the input item values to the intermediate layer. For example, when the room selection model 141 is a CNN, the intermediate layer alternately has a convolution layer that convolves the values of each data item input from the input layer and a pooling layer that maps the values convoluted in the convolution layer. It has a concatenated structure, and extracts the final feature amount while compressing the input information. The output layer has a plurality of neurons that output the values of each item regarding the room suitable for the customer, and outputs the item value regarding the room suitable for the customer based on the feature amount output from the intermediate layer.

In this embodiment, the room selection model 141 is described as being a CNN, but the room selection model 141 is not limited to CNN, and other learning algorithms such as neural networks, Bayesian networks, and decision trees other than CNN can be used. It may be a constructed trained model.

The providing server 1 performs learning using teacher data. The training data is data (contract history) in which contracted property information, customer information, and reply information are associated with each other. The training data is created on the basis of the history data of contract conclusions stored in the history DB 13B.

The providing server 1 inputs the customer information and the answer information included in the training data into the input layer, performs arithmetic processing in the intermediate layer, and acquires information on the room suitable for the customer from the output layer. Since the information about the room includes multiple data items, a set of values in which the values for each data item are combined is obtained from each output node. Each set is assigned a probability value. The probability value indicates the likelihood of the room information output with respect to the input data. The sum of the probability values is one. Each element of the set of values output from the output layer may be a discrete value (for example, the room layout is 1K, 1DK, 2DK, 2LDK, . . . ) or a continuous value (for example, the rent is value in the range of 0 yen to 200,000 yen). The output room-related information corresponds to property building information, basic room information, and basic equipment information.

The providing server 1 compares the information about the room output from the output layer with the contracted property information included in the teacher data, that is, the correct value, and performs calculations in the intermediate layer so that the output value from the output layer approaches the correct value. Optimize the parameters used for processing. The parameters are, for example, weights (coupling coefficients) between neurons, coefficients of activation functions used in each neuron, and the like. Although the parameter optimization method is not particularly limited, for example, the provision server 1 optimizes various parameters using the error backpropagation method.

The provision server 1 performs the above-described learning process using all the teacher data, and generates a learned room selection model 141 . The room selection model 141 when the provision system 100 is constructed becomes a learned room selection model 141 through learning processing. Note that the output layer of the room selection model 141 is not limited to that shown in FIG. For example, different item information may be output for each node in the output layer. One output node outputs the floor plan, and another output node outputs the rent. Each output node may output a plurality of data in which values and probabilities are paired. In the case of the floor plan, the output is (2LDK, 0.85), (2DK, 0.09), (3K, 0.06). If the correct floor plan included in the teacher data is 2LDK, learning is performed so that the probability of 2LDK increases.

A description will be given of a business in which a salesperson uses the provision system 100 to determine a room to propose to a prospective customer. A salesperson acquires customer information such as name, address, and occupation from prospective customers who visit the store. The salesperson uses the sales terminal 2 to register customer information in the prospective customer DB 13A. In addition, the prospective customer is requested to answer the questionnaire, and the answer information is acquired. The salesperson uses the sales terminal 2 to register the reply information in the reply DB 139 . It should be noted that a tablet terminal or the like may be prepared so that the customer can input the customer information and the answer information by himself/herself. Via the sales terminal 2, the salesperson designates the customer information and the reply information, and requests the offer server 1 to select a room to propose. Triggered by the reception of the request from the sales terminal 2, the providing server 1 performs room selection processing.

FIG. 16 is an explanatory diagram relating to room selection processing. FIG. 16 outlines the contents of the processing. The provision server 1 acquires the specified customer information from the prospective customer DB 13A and the specified reply information from the reply DB 139 . The providing server 1 inputs the acquired customer information and answer information to the room selection model 141 as input data. As an output, the room selection model 141 outputs information (proposed room information, property information) regarding the room to be proposed. The proposed room information includes property building information, basic room information, and basic equipment information. The provision server 1 includes the proposed room information output from the room selection model 141, the property building information stored in the property DB 131, the room basic information stored in the room basic DB 132, and the basic equipment DB 133. The basic facility information is referred to, and the room most similar to the proposed room information is specified. If the specified room is not available, the next room will be selected. The providing server 1 uses the room ID of the identified room to acquire basic room information (property information) from the basic room DB 132 . The providing server 1 transmits the room ID, room basic information, and similarity to the sales terminal 2 . The degree of similarity here can be obtained by a method similar to the degree of similarity of vectors. For example, cosine similarity, Pearson's correlation coefficient, or deviation pattern similarity is used as the similarity. Calculation of the degree of similarity is publicly known, so the explanation is omitted.

Based on the room selection result displayed on the sales terminal 2, a proposal is made to the prospective customer. When a contract is concluded as a result of the proposal, the information on the new contract (contract history) becomes teaching data for re-learning the room selection model 141 . Re-learning of the room selection model 141 may be performed each time a new contract is concluded, batch processing may be performed at predetermined intervals, or after a certain number of contracts have been accumulated.

Next, the processing for generating the room selection model 141 will be described again using the flowchart. FIG. 17 is a flow chart showing a procedure example of room selection model generation processing. The control unit 11 of the providing server 1 acquires the contract history from the history DB 13B (step S1). The control unit 11 acquires contracted property information (property building information, basic room information, basic facility information) using the room ID included in the contract history (step S2). The control unit 11 acquires customer information using the prospective customer ID included in the contract history (step S3). The control unit 11 acquires reply information using the case ID included in the contract history (step S4). The control unit 11 generates teacher data including contracted property information, customer information, and answer information (step S5). Using the teacher data, the control unit 11 generates a room selection model 141 (learned model) that outputs proposed room information when customer information and answer information are input (step S6). Specifically, the control unit 11 inputs the customer information and answer information included in the teacher data to the input layer of the neural network, and acquires the proposed room information from the output layer. The control unit 11 compares the obtained proposed room information with the correct value (contract property information) of the teacher data, and adjusts the parameters (weights) used in the arithmetic processing in the intermediate layer so that the information output from the output layer approaches the correct value. etc.). The control unit 11 stores the generated room selection model 141 in the auxiliary storage unit 13 . At the stage of initial construction of the room selection model 141, steps S2 to S6 are repeated for each contract history acquired in step S1, and a series of processing ends.

Next, the room selection process will be explained again using the flowchart. FIG. 18 is a flow chart showing an example of a procedure for room selection processing. The sales terminal 2 acquires customer information (step S11). The sales terminal 2 acquires the reply information (step S12). The sales terminal 2 transmits the customer information and the reply information to the providing server 1 (step S13). The sales terminal 2 may individually transmit the customer information and the reply information to the providing server 1 as soon as they are acquired. The control unit 11 of the providing server 1 receives the customer information and the reply information, and stores the customer information in the prospective customer DB 13A and the reply information in the reply DB 139 (step S14). The control unit 11 inputs the customer information and the reply information to the room selection model 141, and acquires the proposed room information output by the room selection model 141 (step S15). The control unit 11 refers to the proposed room information, property building information stored in the property DB 131, basic room information stored in the basic room DB 132, and basic facility information stored in the basic facility DB 133, and makes a proposal. A degree of similarity between the room information and each room information is calculated (step S16). The control unit 11 identifies a room most similar to the proposed room information (step S17). The control unit 11 transmits basic room information about the identified room to the sales terminal 2 (step S18). The sales terminal 2 receives and displays the room basic information (step S19).

If the room specified in step S17 cannot be moved into, the second room is selected. In addition, if the specified room is a room managed by another real estate agent, if the second and subsequent rooms are managed by the company and the difference in similarity with the specified room is within a predetermined range , it may be a property managed by the company.

FIG. 19 is an explanatory diagram showing an example of a room selection result display screen. The room selection result display screen d01 includes a prospective customer name display d011, a property name display d012, a room information display field d013, a next button d014, a previous button d015, and a viewing recommendation button d016. The prospective customer name display d011 displays the name of the prospective customer. The property name display d012 displays the name of the building and the like. The room information display column d013 displays information on the room. The next button d014 is a button for switching the display to the next room of the displayed room. The front button d015 is a button for switching the display to the room of the previous rank when the room being displayed is second or lower. The viewing recommendation button d016 is a button for displaying a room suitable for viewing together with a recommended room.

This embodiment has the following effects. Since the room selection model 141 learned from the contract history is used to determine the room to be proposed to the customer, even a salesperson with little work experience can propose a room suitable for the customer in the same way as a seasoned salesperson. becomes.

When calculating the degree of similarity, each data item may be weighted. For example, a customer is asked to answer a questionnaire on items that they particularly want to focus on. For items specified by the customer, the weight is increased to calculate the degree of similarity. Further, there may be a plurality of pieces of suggested room information, which may be ranked according to similarity and displayed on the sales terminal 2 .

Personal information to be input to the room selection model 141 in the room selection process is customer information and answer information, but is not limited to this. The personal information to be input may include additional information stored in the additional information DB 13C. In this case, when the room selection model 141 is generated, the additional information is included in the teacher data and used as one of the input data to the room selection model 141 .

Also, although the proposed room information output to the room selection model 141 includes property building information, room basic information, and basic equipment information, it is not limited to this. The proposed room information may include additional information, contract conditions, peripheral information, special notes, and facility information. In this case, when calculating the degree of similarity for selecting a room, the additional information stored in the additional information DB 134 includes additional information, contract conditions, peripheral information, special notes, and facility information included in the proposed room information. , contract terms stored in the contract terms DB 135, peripheral information stored in the peripheral DB 136, special notes stored in the special note DB 137, facility information stored in the facility DB 138 (however, facilities included in the peripheral information Contrast with Further, additional information, contract conditions, peripheral information, special notes, and facility information are included in the contracted property information used when generating the room selection model 141 .

By increasing the types of input data to the room selection model 141 and the types of output data, the rooms selected by the room selection process will be closer to what prospective customers want, and an improvement in contract rate can be expected.

Although it is assumed that the sales terminal 2 acquires the customer information and the reply information, the present invention is not limited to this. The user terminal 3 may transmit the customer information and the reply information to the providing server 1 via the network N. Furthermore, the user terminal 3 may be permitted to execute the room selection process. The providing server 1 inputs the customer information and answer information transmitted from the user terminal 3 to the room selection model 141 and transmits to the user terminal 3 basic room information based on the proposed room information output by the room selection model 141 .

A plurality of room selection models 141 may be provided according to the attributes of prospective customers. A room selection model for medical students that selects rooms for medical students and a selection model for foreign nationals that selects rooms for foreign nationals may be provided.

(Embodiment 2)
This embodiment further includes a function of suggesting a room suitable for viewing (also referred to as a preview) together with the selected room in the first embodiment. Even if a customer who is looking for a room is introduced to only one room that best suits his/her wishes, it is normal for the customer to hesitate whether or not to make a contract. For this reason, real estate sales representatives usually recommend two more rooms in addition to the room they most want to recommend to the customer. Then, the salesperson invites the customer to preview the three rooms and guides the customer to select the most recommended room. If the two rooms to be added here are too similar to the most recommended room, the customer will rather be confused. However, what kind of room is most suitable has conventionally depended on the experience of the salesperson. This embodiment has a function of proposing a room to be inspected together with the most recommended room. In the following description, the room most recommended to the customer is called the recommended room, and the room suitable for viewing together with the recommended room is called the viewing room.

FIG. 20 is a flowchart illustrating an example of a procedure for preview room selection processing. The salesperson inputs the room ID of the recommended room to the sales terminal 2. - 特許庁When the salesperson determines the recommended room by the room selection process in Embodiment 1, if the recommended room displayed on the sales terminal 2 is selected, input of the room ID is unnecessary. The sales terminal 2 transmits the room ID to the providing server 1 (step S31). The control unit 11 of the providing server 1 receives the room ID (step S32). The control unit 11 acquires the contract history of the contract for the recommended room from the history DB 13B using the room ID (step S33). The control unit 11 analyzes the acquired contract history (step S34). For example, association analysis or market basket analysis. This analysis is a method of finding out a rule such as, in many cases, when product A sells, product B sells together. By applying this, it is possible to obtain rooms that have been previewed many times together with the recommended rooms in the past contract history. The control unit 11 selects the top two rooms that have been previewed together with the recommended room as preview rooms (step S35). The control unit 11 transmits basic room information about the selected viewing room to the sales terminal 2 (step S36). The sales terminal 2 receives and displays the room basic information (step S37). The sales terminal 2 terminates the processing.

FIG. 21 is an explanatory diagram showing an example of a preview room selection result display screen. The preview room selection result display screen d02 includes a prospective customer name display d021, a recommended property name display d022, a preview room information display column d023, an OK button d024, an add button d025, and an edit link d026. The prospective customer name display d021 displays the name of the prospective customer. The recommended property name display d022 displays the name of the building including the room recommended to the prospective customer. The viewing room information display field d023 displays information on rooms recommended for viewing together with recommended rooms. The decision button d024 is a button for inputting that the viewing of the displayed room has been decided. By operating the enter button d024, the combination of rooms to be inspected is stored in the history DB 13B. The add button d025 is selected when adding a property to view. The edit link d026 is a hyperlink that displays an operation menu in the displayed room. A pull-down menu displaying an edit link d026 is displayed. The pull-down menu includes, for example, detailed reference, deletion, replacement, and the like. Detailed reference is a menu for displaying detailed information (a floor plan, a surrounding map, etc.) about the room. Delete is a menu for deleting the displayed room from the room to be previewed. "Replace" is a menu for deleting the displayed room from the rooms to be previewed and adding another room.

In this embodiment, the sales terminal proposes a room to be inspected together with the room most recommended to the customer. According to the proposal, the salesperson can increase the contract rate for the most recommended room by having the customer conduct a preview. In addition, regardless of the experience of the salesperson, it is possible to select a room to be inspected.

Although market basket analysis is used to select a contracted room, a trained model similar to the rules obtained from the analysis may be used. When the trained model inputs the recommended room information, it outputs the preview room information. The training data of the trained model is a combination of the information of the contracted room acquired from the contract contract history and the information of the room that was previewed together. Information about the contracted room included in the training data is input to the input layer of the neural network, and information about the room to be viewed is obtained from the output layer. The obtained room information to be viewed together is compared with the correct value of the teacher data (information on the room to be viewed), and the parameters (weight, etc.) used for arithmetic processing in the intermediate layer are used so that the result output from the output layer approaches the correct value. ) to optimize. By performing such learning, a trained model is generated from the neural network.

(Embodiment 3)
This embodiment relates to a mode of determining a room to be proposed to a customer using a decision tree. Decision trees are one of the supervised learning algorithms for solving classification and regression problems. This is a method of classifying data step by step by setting conditions one after another for given data. In the present embodiment, a decision tree is generated in which the attribute of the customer, the desired condition of the customer's real estate property, and the information of the real estate property are used as explanatory variables, and the probability of closing a deal is used as an objective variable. The customer's attributes include, for example, age, sex, occupation, contracting entity (individual/corporation), number of tenants, presence/absence of guarantor, and the like. The desired conditions are the range of rent, floor plan, area, and the like. The real estate property information includes, for example, rent, room layout, area, and the like.

In the present embodiment, in the learning stage, a decision tree is created based on past activity information for prospective customers (hereinafter also referred to as "activity information"), and in the operation stage, the created decision tree is used to Decide which properties to propose. CART (Classification And Regression Trees) is used in the following decision tree creation. First, based on past activity information stored in the history DB 13B, a group of performance data combining customer attributes, real estate property information, and sales results (whether or not a contract has been concluded) is It is divided into two performance data groups according to the customer's attribute, the customer's desired conditions for the real estate property, or one value (feature amount) included in the real estate property information. At this time, the feature values are selected so that one of the performance data groups after the division is biased toward the data that resulted in a contract, and the other performance data group is biased toward the data that did not lead to a contract. The greater the bias after the division compared to before the division, the greater the influence of the feature amount on the sales results. First, the performance data group is divided into two performance data groups according to the feature value that has the greatest impact on sales results. Repeat the division of In this way, a decision tree, which is a model for classifying performance data in a tree-like manner, is generated.

The influence of the feature amount on the sales results is determined by calculating the information gain when the performance data group is divided according to the value of the feature amount. Assume that the performance data group is divided into a first classification group and a second classification group. Let _N0 be the total number of data included in the performance data group before division, N1 be the number of data included in the _first classification group, and _N2 be the number of data included in the second classification group. N ₀ =N ₁ +N ₂ . Let the information gain be IG. The information gain IG is represented by the following formula (1).
IG = (impurity of the actual data group before division) - {(N ₁ /N ₀ ) × (impurity of the first taxonomy) + (N ₂ /N ₀ ) × (impurity of the second taxonomy Purity)} (1)

Impurity is a quantified bias of data included in the performance data group. Specifically, the Gini impurity is used as the impurity. Let _n0 be the total number of data included in the performance data group, n1 be the _number of data that have closed deals, and n2 be the _number of data that have not closed deals. n ₀ =n ₁ +n ₂ . Let Gi be the Gini impurity. The Gini impurity Gi is represented by the following formula (2).
Gi=1−{(n ₁ /n ₀ ) ² +(n ₂ /n ₀ ) ² } (2)

The value of the Gini impurity increases as the data included in the performance data group is biased towards one of the data for which a deal is closed or the data for which a deal is not closed. Information gain indicates the difference between the impurity before splitting and the weighted average of the impurities after splitting. The value of the information gain increases as the bias before division is smaller and the bias after division is greater. Therefore, the greater the information gain, the greater the effect of the feature amount on the sales results. By selecting the feature amount that increases the information gain, it is possible to select the feature amount that greatly affects the sales results.

Note that instead of the information gain, an improve value, which is a value obtained by multiplying the information gain value by _N0 , may be calculated. As with the information gain, the larger the improve value, the greater the influence of the feature amount on the sales result. Assuming that the improve value is Imp, the improve value Imp is expressed by the following equation (3).
Imp = N ₁ × {(impurity of performance data group before division) - (impurity of first classification group)} + N ₂ × {(impurity of performance data group before division) - (second classification group impurity)} (3)

CART was used as an algorithm for decision tree analysis, but not limited to it, ID3 (Iterative Dichotomiser 3), C4.5, or C5.0 using entropy as impurity, or CHAID (Chi- Squared Automatic Interaction Detection) etc. may be used.

The performance data used when generating the decision tree will be explained. FIG. 22 is an explanatory diagram showing an example of the performance data DB. The performance data DB 13D stores performance data to be subjected to decision tree analysis. The performance data DB 13D is stored in the auxiliary storage unit 13, for example. A decision tree is generated by analyzing the performance data. The performance data DB 13D includes an ID column, a branch number column, a sales result column, a gender column, an age column, a desired floor plan column, a desired rent column, a desired area column, a room ID column, a contract rent column, a floor plan column, and an area column. The ID column stores an ID that identifies performance data. The branch number column stores branch numbers. Achievement data can be uniquely identified by the ID and branch number. The business result column stores whether or not the contract for the proposed room has been concluded. If the contract is concluded, the business result column stores "contract concluded". If the contract is not concluded, the business result column stores "lost order". When proposing a room to a prospective customer, it is normal to propose a plurality of rooms, so even in cases where a contract is concluded, a plurality of performance data are created. Performance data with the same ID indicates that the performance data is created from the same case. Of the performance data with the same ID, one performance data results in a closed contract, and the other performance data results in a lost order. The Gender column stores the gender of the prospective customer. The age column stores the age group of the prospective customer. If the prospective customer's age is between 20 and 29, the age column stores "twenties". The desired floor plan column stores the floor plan desired by the prospective customer. The desired rent column stores the rent requested by the prospective customer. A rent band specifying the desired lower limit and upper limit of rent may be stored in the desired rent column. The desired area column stores the area desired by the prospective customer. The Room ID column stores the room ID of the room proposed to the prospective customer. The contract rent column stores the contract rent for the room. The floor plan row stores the layout of the rooms. The area column stores the area in which the room is located.

In performance data, gender and age are examples of attributes of prospective customers (borrowers). The attributes may include occupation, annual income, number of residents, etc. stored in the prospective customer DB 13A. Desired floor plan and desired rent are examples of desired conditions. The desired conditions may include the nearest station stored in the response DB 139, the school district, and whether auto-locking is desired. Contract rent and room layout are examples of property information. The property information includes the structure and age of the building stored in the property DB 131, information such as the entrance and plumbing facilities stored in the basic facility DB 133, and surrounding environment information such as the nearest convenience store and the nearest arterial road stored in the surrounding DB 136. , special notes stored in the special note DB 137, and facility information stored in the facility DB 138 may be included.

Next, processing for generating a decision tree will be described. FIG. 23 is a flow chart showing an example of the procedure of decision tree generation processing. The control unit 11 of the providing server 1 creates performance data (step S51). The control unit 11 performs decision tree analysis using the created performance data (step S52) to generate a decision tree.

FIG. 24 is a flow chart showing an example of the procedure of performance data creation processing. The processing shown in FIG. 24 corresponds to step S51 in FIG. The control unit 11 of the providing server 1 acquires activity information from the history DB 13B (step S61). The activity information to be acquired excludes those in progress, and acquires those whose results have been determined and whose value in the status column is completed or lost. The control unit 11 selects one to be processed from the acquired activity information (step S62).

The control unit 11 acquires customer information (step S63). The control unit 11 searches the prospective customer DB 13A using the prospective customer ID included in the activity information as a key to acquire customer information. The control unit 11 acquires the reply information (step S64). Using the case ID included in the activity information as a key, the control unit 11 searches the response DB 139 and acquires response information. The control unit 11 acquires property information (step S65). The control unit 11 searches the basic room DB 132 using, as keys, the room IDs stored in the room ID column of the result column and each column of the preview property column included in the activity information, and acquires the basic room information. Using the property ID included in the basic room information as a key, the control unit 11 searches the property DB 131 to acquire information about the property building.

The control unit 11 acquires other information (step S66). Using the room ID as a key, the control unit 11 searches the basic facility DB 133 for basic facility information, the additional information DB 134 for additional information about the room, the contract terms DB 135 for contract terms, and the special notes DB 137. Search for to get specials. Also, the control unit 11 searches the peripheral DB 136 using the property ID as a key to obtain environment information. Further, the control unit 11 searches the additional information DB 13C using the case ID as a key to obtain additional information. Which information is acquired as the other information is determined by the data items included in the performance data.

Based on the activity information to be processed, the acquired customer information, response information, property/room information, basic equipment information, additional information, contract conditions, special notes, environmental information, and additional information, the control unit 11 outputs performance data. It is generated and stored in the performance data DB 13D (step S67).

Attributes of prospective customers included in performance data are obtained from customer information. The desired conditions included in the performance data are obtained from the answer information and additional information. Property information included in the performance data is obtained from basic room information, property building information, basic facility information, contract conditions, special notes, and environmental information.

The control unit 11 determines whether or not there is unprocessed activity information (step S68). When control unit 11 determines that there is unprocessed activity information (YES in step S68), the process returns to step S62 to process the unprocessed activity information. If it is determined that there is no unprocessed activity information (NO in step S68), the performance data creation process is terminated.

FIG. 25 is a flow chart showing the procedure of decision tree analysis processing. The processing shown in FIG. 25 corresponds to step S52 in FIG. The control unit 11 of the providing server 1 calculates the information gain by round robin for the performance data group stored in the performance data DB 13D (step S81). The control unit 11 calculates information gain for all items included in the performance data. The control unit 11 individually calculates the information gain when the performance data group is divided according to the content of each item included in the performance data.

Further, when the performance data can be classified in a plurality of ways according to the contents of the items included in the performance data, the control section 11 calculates the information gain for each of the plurality of classification methods. Multiple classification methods for which the information gain should be calculated may be predetermined for each item. Note that the control unit 11 may calculate an improve value instead of the information gain.

Next, the control unit 11 selects one item value (characteristic value) for use in classifying the performance data from among the items included in the performance data (step S82). The control unit 11 selects an item that provides the maximum information gain among the calculated information gains. Note that the control unit 11 may use the improve value instead of the information gain to make the selection. Next, the control unit 11 classifies a plurality of performance data by dividing the performance data group into two performance data groups according to the selected feature amount (step S83). The control unit 11 classifies the performance data according to the classification method selected in step S82. By classifying the performance data according to the value of the item that maximizes the information gain, the performance data are sorted according to the value of the item in descending order of the impact on the business result.

Next, the control unit 11 determines whether or not there is a performance data group that does not satisfy the classification end condition among the plurality of performance data groups (step S84). The classification end condition is predetermined and stored in advance in the auxiliary storage unit 13 or included in the control program 1P. An example of a classification end condition is that the number of times of classification has reached a predetermined number. Another example of the classification termination condition is that the ratio of data indicating that the sales result is a closed contract or data indicating that the sales result is a lost order in the performance data group exceeds a predetermined ratio. Another example of a classification termination condition is that the number of performance data included in the performance data group is less than a predetermined number. Another example of a classification termination condition is that the number of performance data groups exceeds a predetermined upper limit.

If the control unit 11 determines that there is a performance data group that does not satisfy the classification termination condition (YES in step S84), it selects one pass/fail data group that does not satisfy the classification termination condition (S85). The control unit 11 returns the process to step S81 and executes the processes of steps S81 to S83 for the selected result data group. In this way, the control unit 11 recursively repeats the processing of steps S81 to S85, thereby classifying the performance data step by step according to the item values in descending order of influence on the business result.

If there is no result data group that does not satisfy the classification end condition (NO in step S84), the control unit 11 creates a decision tree that classifies the plurality of result data in stages, and stores the created decision tree in the auxiliary storage unit 13. (step S86). The control unit 11 terminates the decision tree analysis process and returns the process to the caller.

FIG. 26 is an explanatory diagram showing an example of a decision tree. The decision tree shown in FIG. 26 is an example of classifying performance data by two items, desired rent and property rent. If the customer's desired rent exceeds 50,000 yen and the introduced property rent exceeds 50,000 yen, the probability of losing the order is 80% and the probability of closing the contract is 20%. If the rent desired by the customer exceeds 50,000 yen and the rent of the introduced property is 50,000 yen or less, the probability of losing the order is 20% and the probability of closing the contract is 80%. If the rent desired by the customer is 50,000 yen or less and the rent of the introduced property exceeds 50,000 yen, the probability of losing the order is 80%, and the probability of closing the contract is 20%. If the rent desired by the customer is 50,000 yen or less and the rent of the introduced property is 50,000 yen or less, the probability of losing the order is 20% and the probability of closing the deal is 80%. Although FIG. 26 shows an example of classification by two items, the result of classification by other items included in the performance data may be reflected in the decision tree. For example, the classification is based on the desired floor plan and the floor plan of the introduced property.

Processing for selecting a room to be proposed to a new prospective customer using the generated decision tree will be described. In the following description, it is assumed that customer information, reply information, and additional information about a new prospective customer have been obtained and have already been stored in the prospective customer DB 13A, reply DB 139, and additional information DB 13C, respectively. FIG. 27 is a flow chart showing another procedure example of room selection processing. The control unit 11 of the providing server 1 acquires customer information from the prospective customer DB 13A (step S91). The control unit 11 acquires desired information from the reply DB 139 and the additional information DB 13C (step S92). The control unit 11 refers to the basic room DB 132 and selects a room to be processed (step S93). The control unit 11 selects rooms whose status column values are vacant, expected to move out, etc., and excludes rooms whose status column values are occupied, applied for, and the like. The control unit 11 acquires property information from the property DB 131 and the basic room DB 132 (step S94). The control unit 11 acquires other information from the basic equipment DB 133, the additional information DB 134, the contract terms DB 135, the peripheral DB 136, the special notes DB 137, the additional information DB 13C, and the like (step S95). Based on the information obtained in steps S91 to S95, the control unit 11 searches the decision tree and obtains the contract probability (step S96). The obtained contract probability is stored in a temporary storage area provided in the main storage unit 12, the auxiliary storage unit 13, etc., in association with the property ID and the room ID. The control unit 11 determines whether or not there is an unprocessed room (step S97). When control unit 11 determines that there is an unprocessed room (YES in step S97), the process returns to step S93, and the unprocessed room is processed. When the control unit 11 determines that there is no unprocessed room (NO in step S97), the control unit 11 narrows down the rooms based on the contract probability (step S98). For example, the control unit 11 narrows down the rooms to three rooms with the highest contract probability. The control unit 11 outputs information about the narrowed down rooms (step S99), and terminates the process. Since the room information output by the control unit 11 is displayed on the sales terminal 2, the salesperson follows it and proposes a room to the prospective customer.

The present embodiment has the following effects. A decision tree generated based on past sales activities is used to determine the room to be proposed to the customer, so even a salesperson with little work experience is predicted to have a high probability of closing a deal with a prospective customer. It becomes possible to propose a room that

Although this embodiment has been described on the premise that one decision tree is generated, the present invention is not limited to this. The performance data may be divided according to some items included in the performance data, and a decision tree may be generated for each divided performance data group. For example, the customer's occupation may be divided into four cases of student, office worker, business owner, and others, and a decision tree may be generated for each of the divided performance data. In the room selection process, a decision tree to be used is selected according to the attributes of the prospective customer, and the selected decision tree is used for processing.

(Embodiment 4)
The present embodiment relates to narrowing down the rooms for which the contract probability is calculated. FIG. 28 is a flow chart showing another procedure example of room selection processing. The room selection process according to the present embodiment is obtained by modifying the room selection process according to the third embodiment shown in FIG. FIG. 28 shows the modified portion. After executing step S92 in FIG. 27, the control unit 11 of the providing server 1 extracts narrowing-down information (step S100). The information for narrowing down is, for example, the desired floor plan, desired rent, and desired area included in the desired information. The control unit 11 uses the information for narrowing down to narrow down the rooms to be processed (step S101). The control unit 11 executes steps from step S93 in FIG. 27 on the narrowed-down rooms as processing targets.

Since the present embodiment narrows down the rooms to be processed, it is possible to reduce the processing amount of the room selection process.

(Embodiment 5)
The present embodiment, like the fourth embodiment, narrows down the rooms to be processed and also uses a plurality of decision trees. FIG. 29 is a flow chart showing another procedure example of room selection processing. The room selection process according to the present embodiment is obtained by modifying the room selection process according to the third embodiment shown in FIG. FIG. 29 shows the changed part. In FIG. 29, the processes in steps S100 and S101 are the same as those in the fourth embodiment, so description thereof will be omitted. After executing step S101, the control unit 11 extracts selection information for selecting a decision tree (step S102). The information for selection is, for example, the prospective customer's occupation, family type, and the like. The control unit 11 uses the selection information to select a decision tree to be used (step S103). Using the selected decision tree, the control unit 11 executes steps after step S93 in FIG. 27 with the narrowed-down rooms as processing targets.

In this embodiment, since the properties and rooms to be processed are narrowed down, it is possible to reduce the processing amount of the room selection process, and by selectively using a plurality of decision trees, it is predicted that the probability of closing the contract is high. It is possible to select a room.

(Embodiment 6)
This embodiment relates to a form in which a prospective customer accesses the provision server 1 via a network and searches/refers to property/room information. The providing server 1 provides prospective customers with property/room information that matches the search conditions, in the same manner as in conventional non-operation information sites. At that time, recommended room information determined using a decision tree is presented. In the following description, it is assumed that the customer terminal used by the prospective customer communicates with the provision server 1, but the communication is not limited to this. A web server that mediates between the customer terminal and the providing server 1 may be provided.

FIG. 30 is a flow chart showing a procedure example of room search processing. The control unit 11 of the providing server 1 acquires desired information from the customer terminal (step S111). Control unit 11 searches for a room that matches the desired information (step S112). The control unit 11 searches the property DB 131 and the basic room DB 132, as well as the basic equipment DB 133, the additional information DB 134, the contract conditions DB 135, the peripheral DB 136, the special note DB 137, or the facility DB 138 according to the conditions included in the desired information. If the desired information does not include any conditions, the control unit 11 searches for rooms that can be moved in immediately. The control unit 11 determines whether or not there is a hit in the search (step S113). When the control unit 11 determines that there is no hit in the search (NO in step S113), the process proceeds to step S121. When the control unit 11 determines that there is a hit in the search (YES in step S113), it determines whether or not information (customer information) on prospective customers has been obtained (step S114). The customer information is received from the customer terminal together with the desired information entered by the prospective customer himself/herself into the customer terminal. In addition, the customer terminal may store the prospective customer's Internet behavior history as a cookie or the like, and the customer information (age, gender, etc.) may be estimated from the behavior history. When determining that the customer information is not obtained (NO in step S114), the control unit 11 shifts the process to step S116. If the control unit 11 determines that customer information has been obtained (YES in step S114), it extracts information used for decision tree determination from the customer information (step S115). The control unit 11 selects a room to be processed from among the rooms hit by the search (step S116). The control unit 11 acquires information to be used for determination by the decision tree from the information regarding the selected room (step S117). The control unit 11 uses the decision tree to obtain the contract probability of the selected room (step S118). The control unit 11 determines whether or not there is an unprocessed room (step S119). When control unit 11 determines that there is an unprocessed room (YES in step S119), the process returns to step S116, and the unprocessed room is processed. When the control unit 11 determines that there is no unprocessed room (NO in step S119), it sorts the information of the rooms hit in step S111 (step S120). The control unit 11 generates a screen to be transmitted to the customer terminal (step S121). When the transition is made from step S113, the screen to be sent is generated to notify that the number of search hits is 0 (step S121). When the transition is made from step S120, a screen is generated to conspicuously display the several cases with the highest contract probability, separately from the list. The control unit 11 outputs the generated screen to the customer terminal (step S122), and terminates the process.

FIG. 31 is an explanatory diagram showing an example of a search result display screen. The search result display screen d03 includes a display area d031, a recommended area d032, a condition setting area d033, and a search button d034. The display area d031 displays the information of the rooms hit by the search. The recommended area d032 displays information on rooms determined by the decision tree to have a high contract probability. The condition setting area d033 is an area for setting search conditions. The search button d034 is a button for instructing a re-search under the search conditions set in the condition setting area d033. For example, the recommended area d032 has a frame different from that of the display area d031. In FIG. 31, it is shown as a shaded frame. Also, the recommended area d032 includes a recommended message d0321. In this way, the recommended area d032 is expected to catch the attention of prospective customers by including a display mode that is visually different from the display area d031.

The number of recommended areas d032 provided in the search result display screen d03 is arbitrary. Determined by the contract probability determined by the decision tree. For example, if a maximum of three rooms with a contract probability of 75% or higher are set to be displayed, the recommended area d032 is not displayed if there is no property with a contract probability of 75% or higher. Even if there are four or more properties with a contract probability of 75% or more, only three recommended areas d032 are displayed. In addition, although there is no particular limitation on how the display area d031 and the recommended area d032 are displayed, it is desirable to display the recommended area d032 as high as possible.

This embodiment has the following effects. In the search result display screen d03, the information of the room determined by the decision tree to have a high probability of closing the deal is displayed, which is expected to improve the closing rate.

Note that the recommended area d032 is not displayed if there is no property whose contract probability is equal to or higher than a predetermined value, but the present invention is not limited to this. If there is no property with a contract probability higher than the specified value, delete some of the multiple conditions included in the desired information, or expand the range of values to increase the number of applicable rooms. . Then, with respect to the increased rooms, the probability of closing a deal may be determined by a decision tree, and rooms determined to have a high probability of closing a deal may be displayed in the recommended area d032. In this case, even if the prospective customer himself/herself misunderstands the conditions of the desired room, an improvement in the closing rate is expected. In addition, since the information of the rooms not hit by the search will be displayed in the recommendation area d032, the content of the recommendation message d0321 is changed to "How about a similar room like this?" It is desirable to make them equal.

In the above-described embodiment, it is assumed that one decision tree is used in the room selection process, but a plurality of decision trees may be used. For example, a random forest is adopted in which a plurality of decision trees are created from a plurality of items arbitrarily selected from all items included in the performance data. In the room selection process, a random forest is used to obtain the contract probability of each room.

(Modification)
This modification includes another example of the room search process shown in the sixth embodiment. FIG. 32 is a flow chart showing a procedure example of room search processing. The prospective customer operates the customer terminal to transmit an information request to the providing server 1 (step S131). The control unit 11 of the providing server 1 receives the request (step S132). The control unit 11 acquires the cookie from the transmitted request (step S133). If no cookie has been sent from the customer terminal, the cookie acquired by the control unit 11 is empty. The control unit 11 determines whether or not there is a room condition (step S134). Here, the control unit 11 determines that there is a condition in the following cases. In the search result display screen d03, after conditions are set in the condition setting area d033, the search button d034 is selected, and desired room information is transmitted from the customer terminal. It should be noted that if conditions similar to those in the condition setting area d033 can be set, the information may be sent from a web form other than the search result display screen d03. In addition, the desired room information is not sent to the customer terminal by a form, but is set in a cookie. The control unit 11 determines that there is a condition when the form including the desired room information has not been sent, or when the cookie is empty, or when the cookie is not empty but does not contain the desired room information. If the control unit 11 determines that there are room conditions (YES in step S134), it acquires the conditions from the form and the cookie (step S135). It should be noted that re-acquisition is not necessary if it has already been acquired when the determination in step S134 is made. The control unit 11 searches for a room that meets the conditions (step S136). The control unit 11 uses the decision tree to obtain the contract probability for each room hit in the search (step S137). If the cookie contains analogous attributes of prospective customers, such as gender, age group, occupation, and area of residence, a plurality of decision trees may be used according to these attributes. For example, a student decision tree is used for prospective customers who are likely to be students. The control unit 11 sorts the room information in descending order according to the contract probability (step S138). The control unit 11 generates a screen to be transmitted to the customer terminal (step S139). The control unit 11 transmits the generated screen to the customer terminal (step S140). The customer terminal receives and displays the screen (step S141). The screen displayed by the customer terminal is the search result display screen d03. The control unit 11 ends the processing. As described above, when the prospective customer selects the search button d034 on the search result display screen d03, the room search process is activated and executed from step S131.

When the control unit 11 determines that there is no room condition (NO in step S134), it determines whether or not the number of rooms browsed by prospective customers is k or more (step S143). A natural number is set in advance for k. The determination is made based on the browsing history of the prospective customer. The browsing history is included in the cookie sent by the customer terminal. Alternatively, the browsing history associated with the prospective customer's ID may be stored in the auxiliary storage unit 13, and the ID given by the providing server 1 may be acquired from the cookie sent by the customer terminal to acquire the browsing history. Note that the browsing history referred to here is different from the history stored in the history DB 13B. When it is determined that the number of rooms browsed by the prospective customer is k or more (YES in step S143), the room information based on the browsing history is acquired (step S144). The control unit 11 executes steps after step S139. When it is determined that the number of rooms browsed by prospective customers is not k or more (NO in step S143), it is determined whether or not there are k or more popular rooms (step S145). Although the same k as in step S143 is used for determination, other natural numbers may be used in place of k. A popular room is, for example, the room with the most page views. It can even be the room that has been favorited the most by a prospective customer. When determining that there are k or more popular rooms (YES in step S145), the control unit 11 acquires information on popular rooms (step S146). The control unit 11 executes steps after step S139. When determining that there are not more than k popular rooms (NO in step S145), the control unit 11 acquires information about a predetermined room (step S147). The predetermined room is, for example, a recommended room set in advance. The control unit 11 executes steps after step S139.

In this modified example, it is possible to obtain a list of rooms even if the prospective customer does not set room conditions (room request information). Since the prospective customer can obtain the information about the room with a few operations, the possibility of abandoning the search due to the complicated condition setting is reduced. By making the prospective customer set the conditions again after displaying the list, it is possible to prevent the prospective customer from feeling that there is no information on the room that he or she desires.

Note that the processing when the room conditions have not been acquired (NO in step S134) is merely an example, and other processing may be performed. For example, a list may be created and displayed in which all rooms are sorted according to a predetermined condition, in ascending order of rent, in ascending order of building age, in descending order of information (new arrival order), or the like. In addition, if the conditions of the room are not acquired, but the analogous attribute of the prospective customer is obtained, a list of rooms according to the attribute, for example, a list of rooms recommended for students, a list of rooms recommended for office workers in their twenties You may display a room list of The room list may be created in advance, or may be created each time using the value of the customer class column of the special DB 137 .

The technical features (components) described in each embodiment can be combined with each other, and new technical features can be formed by combining them.
The embodiments disclosed this time are illustrative in all respects and should not be considered restrictive. The scope of the present invention is indicated by the scope of the claims rather than the above-described meaning, and is intended to include all modifications within the scope and meaning equivalent to the scope of the claims.

100 Provision System 1 Provision Server 11 Control Unit 12 Main Storage Unit 13 Auxiliary Storage Unit 131 Property DB
132 Room Basic DB
133 Basic Equipment DB
134 additional information DB
135 Contract Terms DB
136 Peripheral DB
137 Special DB
138 Facility DB
139 Answer DB
13A Prospective Customer DB
13B History DB
13C Additional Information DB
141 room selection model 15 communication unit 1P control program 2 sales terminal 3 user terminal

Claims (11)

Acquire the desired conditions for the property of the lessee,
Acquiring property information of properties that match at least part of the acquired desired conditions,
The attribute of the borrower, the desired conditions obtained, and the property information obtained are used as explanatory variables, and the property contract probability is used as the objective variable. To the decision tree , input to obtain the closing probability of the property,
An output program characterized by causing a computer to perform a process of outputting the acquired contract probability. 2. The output program according to claim 1, wherein the attributes of the borrower include age, gender or classification. 3. The output program according to claim 1 or 2 , wherein the property information corresponding to a predetermined number of high order closing probabilities is output. the computer
Acquire the desired conditions for the property of the lessee,
Acquiring property information of properties that match at least part of the acquired desired conditions,
The attribute of the borrower, the desired conditions obtained, and the property information obtained are used as explanatory variables, and the property contract probability is used as the objective variable. To the decision tree , input to obtain the closing probability of the property,
An output method characterized by outputting the acquired contract probability. a first acquisition unit that acquires desired conditions for a tenant's property;
a second acquisition unit that acquires property information of properties that match at least part of the acquired desired conditions;
The attribute of the borrower, the desired conditions obtained, and the property information obtained are used as explanatory variables, and the property contract probability is used as the objective variable. a third acquisition unit that inputs to the decision tree and acquires the contract probability of the property;
An output device comprising: an output unit that outputs the obtained contract probability. Enter input data including answer information and personal information into a trained model trained with teacher data including answer information to questions about real estate properties, personal information, and contracted property information, acquire property information,
Acquire property information that matches the acquired property information,
Acquiring a contract history related to the matching property information,
Based on the acquired contract history, acquire information on properties that were previewed together with contracted properties but were not contracted,
Output the matching property information and information on properties that were previewed but did not result in a contract
An output program characterized by making a computer perform processing. 7. The output program according to claim 6 , wherein the teacher data and the input data contain additional information including environmental information around the property. 8. The output program according to claim 7, wherein the property information includes a plurality of pieces of property information suitable for the borrower. 9. The property information according to any one of claims 6 to 8 , wherein the property information includes a management type indicating whether or not the property is managed by the company, and the property information in which the management type is managed by the company is output. The described output program. A computer inputs input data, including answer information and personal information, into a trained model that has been trained with teacher data including answer information, personal information, and contracted property information for questions about real estate, acquires property information,
Acquire property information that matches the acquired property information,
Acquiring a contract history related to the matching property information,
Based on the acquired contract history, acquire information on properties that were previewed together with contracted properties but were not contracted,
Output the matching property information and information on properties that were previewed but did not result in a contract
An output method characterized by: a first acquisition unit that acquires property information by inputting input data, including answer information and personal information, into a learned model that has been trained using teacher data that includes answer information, personal information, and contracted property information for real estate questions; ,
a second acquisition unit that acquires property information that matches the acquired property information ;
a third acquisition unit that acquires a contract history related to the matching property information;
a fourth acquiring unit that acquires information on a property that was previewed together with a contracted property but did not result in a contract, based on the acquired contract conclusion history;
an output unit that outputs the matching property information and information on properties that were previewed but did not result in a contract;
An output device comprising:

Images