JP7807782B2 - Store property recommendation system and its program - Google Patents

Description

The present invention relates to a system that recommends store properties to businesses wishing to open stores, and in particular to a store property recommendation system and program that recommends store opening areas and recommends high-quality store properties in those areas.

[Prior Art]
Businesses (prospective store owners) hoping to open stores would search for store properties in their desired area and with the desired conditions on real estate agency websites.

[Related Technology]
Incidentally, relevant prior art documents include Japanese Patent Laid-Open No. 2002-169875 "Real Estate Business Information Provision Processing System" (Patent Document 1), Japanese Patent Laid-Open No. 2002-259731 "Property Brokerage Method and Property Brokerage System" (Patent Document 2), Japanese Patent Laid-Open No. 2002-366681 "Franchise Business Support System..." (Patent Document 3), and Japanese Patent Laid-Open No. 2008-065607 "Store Opening Support Method" (Patent Document 4).

Patent Document 1 describes an information providing processing system that allows users to easily search for desired properties at any time.
Patent Document 2 describes a system that provides property information that meets desired conditions in order to act as an intermediary between those wishing to open a store and property providers.

Patent Document 3 describes a system that assists those who wish to become franchisees in finding a franchisor and a store that meets their requirements.
Patent document 4 describes a method of constructing a market selection model based on past store performance data and statistical data, and then revising the initial candidate sites selected based on the model through a questionnaire to select the final candidate sites.

Japanese Patent Application Laid-Open No. 2002-169875 Japanese Patent Application Laid-Open No. 2002-259731 Japanese Patent Application Laid-Open No. 2002-366681 Japanese Patent Application Laid-Open No. 2008-065607

However, with the above-mentioned conventional store property search systems, the store location and store conditions must be set by the prospective store owner, and whether the searched store is located in a location that will result in strong business performance must be determined solely by the prospective store owner's experience and intuition, resulting in the inability to effectively find suitable store properties.

Patent Documents 1 to 3 simply search for stores that wish to open based on certain conditions, while Patent Document 4 only selects candidate locations for opening a store, and does not describe a configuration for recommending areas that are suitable for the store's business format and have strong business performance, or recommending store properties in those areas that are also performing well.

The present invention was made in consideration of the above situation, and aims to provide a store property recommendation system and program that can recommend areas that are suitable for a store's business model and where business performance can be improved, and further recommend store properties in those areas that will perform well.

The present invention, which aims to solve the problems of the above-mentioned conventional examples, is a store property recommendation system that includes a store property recommendation server that recommends store properties, and the store property recommendation server uses a trained regional model that has learned data on the regional characteristics of stores that have performance above a specific threshold based on the regional characteristics of areas where stores with performance above a specific threshold are located from a plurality of store data including the area and performance for a specific business type, to determine whether the regional characteristics of a candidate area are regional characteristics for stores that will perform above a specific threshold, and uses the determination result to recommend an area for a store of the specific business type, and uses a trained store model that has learned store data of stores that have performance above a specific threshold based on the store data of stores that have performance above a specific threshold to determine whether a store property that is the subject of real estate brokerage is a store property for a store that will perform above a specific threshold, and uses the determination result to recommend store properties for stores of the specific business type, and generates a store opening report that includes a ranking of combinations of recommended areas and recommended store properties for a specific business type based on information on the regional characteristics for stores that will perform above a specific threshold and store properties for stores that will perform above a specific threshold .

In the above-mentioned store property recommendation system, the store property recommendation server, when determining whether the regional characteristics of a candidate area are regional characteristics suitable for a store with performance above a specific threshold, generates a trained regional model by training a training regional model with data on the regional characteristics of stores with performance above a specific threshold, and inputs the data on the regional characteristics of the candidate area into the trained regional model to determine whether the regional characteristics of the candidate area are regional characteristics suitable for a store with performance above a specific threshold.

In the above-mentioned store property recommendation system, the present invention is characterized in that, when the store property recommendation server determines whether a store property subject to real estate brokerage is a store property intended for a store that will achieve performance above a specific threshold, it trains a training store model with store data of stores that have performance above a specific threshold to generate a trained store model, and inputs data about the store property subject to real estate brokerage into the trained store model to determine whether the store property is a store property intended for a store that will achieve performance above a specific threshold.

The present invention is characterized in that, in the above-mentioned store property recommendation system, when a wide area made up of multiple recommended area units is specified by a prospective store opening candidate, the store property recommendation server recommends areas within that wide area suitable for a specific type of store.

The present invention is characterized in that, in the above-mentioned store property recommendation system, when the size of the store to be recommended is specified by an applicant for store opening, the store property recommendation server recommends store properties for a specific store within the range of the store size specified.

The present invention is a computer program executed on a store property recommendation server that recommends store properties, and is characterized in that the store property recommendation server is configured to: determine whether the regional characteristics of a candidate area are regional characteristics for stores that will perform at or above a specific threshold, based on the regional characteristics of areas where stores with performance above a specific threshold are located from multiple store data including the area and performance, for a specific business type, using a trained regional model that has learned data on the regional characteristics of stores with performance above a specific threshold; use the determination result to recommend an area for a store of the specific business type; determine whether a store property that is a target for real estate brokerage is a store property for a store that will perform at or above a specific threshold, based on store data of stores with performance above a specific threshold, using a trained store model that has learned store data of stores with performance above a specific threshold; use the determination result to recommend store properties for stores of the specific business type; and generate a store opening report including a ranking of combinations of recommended areas and recommended store properties for a specific business type, based on information on the regional characteristics for stores with performance above a specific threshold and store properties for stores with performance above a specific threshold .

According to the present invention, a store property recommendation server determines whether the regional characteristics of a candidate region are suitable for a store whose performance is above a specific threshold, based on the regional characteristics of a region where a store whose performance is above a specific threshold is located from a plurality of store data including the region and performance, using a trained regional model that has trained data on the regional characteristics of stores whose performance is above a specific threshold, and recommends a region suitable for a store of a specific business type using the determination result. The trained store model is trained based on the store data of stores whose performance is above a specific threshold. This store property recommendation system uses the above to determine whether a store property that is the subject of real estate brokerage is a store property for a store that will perform at or above a specific threshold, and uses the determination result to recommend store properties for stores of a specific business type, and generates a store opening report that includes a ranking of combinations of recommended areas and recommended store properties for a specific business type based on information on regional characteristics for stores that will perform at or above a specific threshold and store properties for stores that will perform at or above a specific threshold , so it has the effect of being able to recommend combinations of areas and store properties that are suitable for the store type and will improve performance.

FIG. 1 is a schematic diagram of the configuration of the present system. FIG. 1 is a schematic diagram of the process of generating a regional model and extracting recommended regions. FIG. 10 is a schematic diagram illustrating the process of generating a store model and extracting recommended properties. FIG. 10 is a flow diagram showing the processing of a store property recommendation server. FIG. 10 is an explanatory diagram showing an example of displaying recommended areas and recommended stores.

An embodiment of the present invention will be described with reference to the drawings.
[Outline of the embodiment]
A store property recommendation system (this system) according to an embodiment of the present invention generates a trained model by learning regional models and store models that perform well from the performance of multiple stores for a specific store type, extracts regions to recommend using regional characteristic data based on the trained regional model, extracts store properties to recommend using real estate store property data based on the trained store model, and provides information on the extracted regions and store properties to those wishing to open a store, allowing those wishing to open a store to easily find store properties that are likely to perform well in regions that perform well.

[System configuration: Figure 1]
The configuration of this system will be described with reference to Fig. 1. Fig. 1 is a schematic diagram of the configuration of this system.
As shown in Figure 1, this system includes a store property recommendation server 1, a property database (DB) 20 connected to it, a store database (DB) 21, a regional characteristics database (DB) 22, a recommended region database (DB) 23, a recommended property database (DB) 24, a network 3, a real estate brokerage server 4, a property database (DB) 5, and a store opening applicant terminal 6.
The store property recommendation server 1, the real estate agency server 4, and the store opening applicant terminal 6 are connected to a network 3.

[Store property recommendation server 1]
The store property recommendation server 1 comprises a control unit 11, a memory unit 12, and an interface unit 13, to which a property DB 20, a store DB 21, a regional characteristics DB 22, a recommended region DB 23, and a recommended property DB 24 are connected, and which is further connected to a network 3.

The control unit 11 executes a processing program stored in the storage unit 12 to realize the processing described below.
The storage unit 12 stores processing programs as well as data, parameters, etc. required for processing.
In particular, the memory unit 12 stores the pre-learning regional model and store model described below, and trains these models using store information (stored in store DB21) from the store list provided by the store applicant terminal 6, and stores the learned regional model and learned store model that are the learning results.
The interface unit 13 is an interface for connecting to an external device and a network.

[Each DB]
The property DB 20 stores information on properties that are the subject of real estate brokerage provided by the real estate brokerage server 4 .
The store DB 21 stores store data of a store list provided from the store applicant terminal 6, etc. The store data is information about stores that are actually in operation (data on performance such as sales and profits, addresses, business hours, size, rent, etc.). This data on multiple stores is provided by franchise companies (store applicants) and is managed by the store property recommendation server 1.

The regional characteristics DB22 stores official statistical data such as the census (gender, age, occupation, number of people in household, family type, living arrangement, length of residence, etc.), number of households by annual income class, and consumption expenditures (food, eating out, housing, transportation, education and entertainment, etc.).
Furthermore, the regional characteristics DB22 stores information on regional characteristics for each area (e.g., resident profiling data [cluster, factor data]) based on cluster analysis using statistical data for each city, town, village, postal code area, and specific mesh unit on a map.
Here, the regional characteristic data indicating regional characteristics includes information on regional characteristics for each area that has been subjected to cluster analysis and statistical data for each area.

The recommended area DB 23 stores information on areas in which stores are recommended to open through a process described below.
The recommended property DB 24 stores property information of stores that are recommended by the process described below.
The property BD5 stores information about properties that are brokered by real estate companies.

The network 3 is assumed to be a network such as the Internet.
The real estate intermediation server 4 searches for properties that match the conditions etc. from the store property recommendation server 1 or the store opening applicant terminal 6 and returns the results.
Furthermore, the real estate intermediary server 4 provides (uploads) part or all of the property information in the property DB 5 to the shop property recommendation server 1. The uploaded property information is accumulated in the property DB 20.

The store applicant terminal 6 stores store data of a store list for actual stores that the applicant operates or supervises, and uploads the store data to the store property recommendation server 1. The uploaded store data of the store list is accumulated in the store DB 21.
Here, the person wishing to open a store is someone who already operates multiple stores or supervises or manages multiple stores and wishes to open another store (for example, a franchise management company, etc.).

The store opening applicant terminal 6 also accesses the real estate brokerage server 4 to search for properties, and also accesses the store property recommendation server 1 to receive recommendations for store opening areas (recommended areas) and properties (recommended properties).

Furthermore, if the store data in the store DB 21 is provided by a specific franchise management company and the use of that store data is limited to that franchise management company, then the person wishing to open a store will be an affiliate of that management company.
However, if the provider of the store data does not limit the use of the store data to those related to the operating company, but allows anyone to use it, then anyone may wish to open a store.

[Processing on the store property recommendation server]
Next, an outline of the processing performed by the store property recommendation server 1 in this system will be described.
The store property recommendation server 1 has a control unit 11 that reads the processing program from the memory unit 12, thereby realizing a store data upload means, a property data upload means, a regional model generation means, a store model generation means, a recommended region extraction means, a recommended property extraction means, and a store opening report generation means.

[Store data upload method]
The store data uploading means inputs store data of the store list uploaded from the store applicant terminal 6 via the network 3 and stores it in the store DB 21.
[Property data upload method]
The property data uploading means inputs property data stored in the property DB 5 from the real estate intermediary server 4 via the network 3 and stores it in the property DB 20 .

[Regional model generation means]
The regional model generation means extracts regional characteristic data relating to a region with favorable business performance (a region with business performance equal to or greater than a specific threshold) from the regional characteristic DB 22 based on the store data stored in the store DB 21, and trains a pre-prepared model. The model is a model for machine learning in artificial intelligence (AI).
The trained model then becomes a "regional model" that calculates the probability of whether a region is performing well for each business type.

[Store model generation means]
The store model generation means extracts store-related data for stores with favorable business performance (stores with business performance equal to or greater than a specific threshold) from the store data stored in the store DB 21, and trains a pre-prepared model. The model is a machine learning model for artificial intelligence (AI).
The trained model then becomes a "store model" that calculates the probability of whether a store is performing well for each business type.

[Method for extracting recommended areas]
The recommended area extraction means uses the regional model generated by the regional model generation means to input regional characteristic data for a specific area stored in the regional characteristic DB22 into the regional model, calculates the probability that a store of a specific business type will perform well if it is opened in the specific area, and outputs areas where the probability is higher than a specific threshold to the recommended area DB23 as recommended areas.

[Method for extracting recommended properties]
The recommended property extraction means uses the store model generated by the store model generation means to input property data of a specific property stored in the property DB20 into the store model, calculates the probability that a store of a specific business type will perform well if opened in the specific property, and outputs properties with a probability higher than a specific threshold to the recommended property DB24 as recommended properties.

[Store opening report generation means]
The store opening report generation means ranks the combinations of area and store by weighting the probability of the extracted area and the probability of the extracted store using a coefficient from the multiple areas extracted by the recommended area extraction means and the multiple stores extracted by the recommended property extraction means. The store property recommendation server 1 provides this ranked information to the store opening candidate terminal 6 as final recommendation information (store opening report).

In addition, the store opening report generating means generates a report (sales report) that can be used for sales for real estate agents based on the extracted area and store information.
The sales report shows, in numerical terms, whether the area of the property being brokered is suitable for a particular business type (whether good business performance can be expected) and whether the store in the property is suitable for a particular business type. This makes it easier for real estate agents to recommend properties for specific business types based on objective data, allowing them to broker properties more efficiently.

[Region model generation and recommended region extraction: Figure 2]
Next, an outline of the process of generating a region model and extracting a recommended region will be described with reference to Fig. 2. Fig. 2 is a schematic diagram of the process of generating a region model and extracting a recommended region.
As shown in FIG. 2, to generate a regional model, the regional characteristics DB 22 is referenced from the store DB 21 to learn about regions where stores of a specific business type are performing well, and the regional model 31 learns related regional data as training data.
The regional model 31 is adjusted and trained so that regional data of stores with good business performance has a high score (high probability), and a trained regional model 32 is generated.

Then, candidate area data is read from the area characteristics DB 22 into the trained area model 32, and scores (probabilities) are output. Area data with high scores (high probabilities) are determined to correspond to areas with good business performance, and are output to the recommended area DB 23 together with the numerical values obtained as recommended areas.
Therefore, the recommended area DB 23 stores a plurality of recommended areas and their numerical values.

Specifically, the regional characteristics DB22 stores regional characteristic data from statistical data for each area, or the numerical values of regional characteristic data from the results of cluster analysis based on that statistical data, as regional data, which are then trained into the regional model. When extracting recommended regions, the regional data is also input into the trained regional model, and a score is output. To obtain a high score, it is desirable to select areas with similar (comparable) population structures, or areas with the same or similar clusters, and input them into the trained regional model.

[Store model generation and recommended property extraction: Figure 3]
Next, an outline of the process of generating a store model and extracting recommended properties will be described with reference to Fig. 3. Fig. 3 is a schematic diagram of the process of generating a store model and extracting recommended properties.
As shown in FIG. 3, to generate a store model, store data relating to stores in a specific business category that are performing well is taken from the store DB 21 and used as training data to make the store model 33 learn the data.
The store model 33 is adjusted and trained so that store data of stores with good business performance will have a high score (high probability), and a trained store model 34 is generated.

Then, candidate property data from the property DB 20 is read into the trained store model 34, and scores (probabilities) are output. Here, property data with high scores (high probabilities) is determined to correspond to stores with good business performance, and is output to the recommended property DB 24 together with the numerical values obtained as recommended properties.
Therefore, the recommended property DB 24 stores a plurality of recommended properties and their numerical values.

Furthermore, when loading candidate property data from the property DB 20 into the trained store model 34, it is also possible to load only property data located in areas extracted (narrowed down) as recommended areas into the trained store model 34.
In this way, obtaining scores for candidate property data from areas narrowed down as recommended areas is efficient in processing and makes it easier to understand recommended areas and properties for those wishing to open a store.

[Processing of the store property recommendation server: Figure 4]
Next, the processing of the shop property recommendation server 1 will be described with reference to Fig. 4. Fig. 4 is a flow diagram showing the processing of the shop property recommendation server.
As shown in Figure 4, the store property recommendation server 1 executes a processing program, and the store data upload means receives store data from the store list from the store applicant terminal 6 and performs a store data upload process to store the data in the store DB 21 (S1).

Furthermore, the property data uploading means receives property data from the property DB 5 from the real estate intermediary server 4 and performs a property data upload process to store the data in the property DB 20 (S2).
The order of the store data upload process and the property data upload process does not matter, and the property data upload process may be performed first.

The regional model generation means refers to the regional characteristics DB 22 for the regions of stores in the store DB 21 that are performing well, causes the regional model 31 to learn the corresponding regional data, and generates a trained regional model 32 (S3).
The store model generating means also performs processing to make the store model 33 learn the data of stores in the store DB 21 that are performing well, and generate a trained store model 34 (S4).
The order of the area model generation process and the store model generation process does not matter, and the store model generation process may be performed first. Also, the property data upload process may be performed after the area model generation process and the store model generation process.

The recommended area extraction means reads the area data in the area characteristics DB 22 using the trained area model 32, extracts recommended areas, and performs processing to output the recommended areas to the recommended area DB 23 (S5).
Furthermore, the recommended property extraction means reads property data from the property DB 20 using the trained store model 34, extracts recommended properties, and performs processing to output the properties to the recommended property DB 24 (S6). Here, only the property data of the extracted recommended area may be read into the trained store model 34.
The order of the recommended area extraction process and the recommended property extraction process does not matter, and the recommended property extraction process may be performed first. Also, the property data upload process may be performed after the recommended area extraction process.

Next, the store opening report generating means performs processing to generate a store opening report from the recommended area stored in the recommended area DB 23 and the recommended property stored in the recommended property DB 24 (S7).
As mentioned above, the store opening report is a ranking of regions and properties, and shows the likelihood of successful business performance for each region and property, ranked by probability value.
The store opening report generating means may generate a sales report for real estate agents.

[Application example]
In this system, the recommended area extraction means extracts recommended areas, and the recommended property extraction means extracts recommended properties, but if a person wishing to open a store selects a wide analysis area including multiple small areas (postal code area or mesh area) and also selects the store size, small areas in the wide area that are likely to perform well may be displayed in different colors or shades of color, and stores that are likely to perform well may be displayed using different sized icons.

[Example of recommended areas and recommended stores: Figure 5]
A display example of recommended areas and recommended stores in the application example will be described with reference to Fig. 5. Fig. 5 is an explanatory diagram showing a display example of recommended areas and recommended stores.
Specifically, small areas with a high probability of performing well are displayed in a darker color, and small areas with a low probability are displayed in a lighter color. Also, the colors may be, for example, red for areas with a high probability, orange for small areas with a medium probability, and yellow for small areas with a low probability.
In addition, the icon display for stores is differentiated by changing the shape of the icon depending on the type of business, with the icons of stores with a high probability of performing well being displayed larger and the icons of stores with a low probability of performing well being displayed smaller.

In the example of FIG. 5, the recommended area and the recommended store are displayed on the same map, but they may be displayed on separate maps.
In addition, parameters that can be set by those who wish to open a store include, for the region, specification of a specific wide area (an area consisting of multiple small areas), and for the store, whether it is in front of a station or not, the store's type of business (for example, convenience store, etc.), store size (business area), trade area range, etc. For store size, it is possible to exclude stores larger than a certain business area and narrow down the search to small stores.

Regarding the trade area range, which is a parameter related to the store set by the prospective store owner, the prospective store owner can set the trade area size in detail to a radius of 0.3 km, 0.5 km, 1.0 km, 1.5 km, etc., but it is also possible to have the trade area range selected automatically using AI.
The automatic selection (automatic determination) of the trade area range will now be described in detail.
The store model generation means learns the conditions and sales amounts of multiple existing stores in all trade area sizes (for example, 0.3 to 1.5 km radius) and generates a store model (AI model) of a store with high sales (ideal store), and calculates the accuracy of the ideal store model.
Then, the accuracy is calculated for all trade area sizes, and the trade area size with the highest accuracy is recommended as the trade area range.

For automatic determination of trade area size, an AI model such as "Elastic Net" is used. Elastic Net is a method that combines regularized ridge regression and LASSO regression. Like LASSO regression, Elastic Net can generate a reduced-dimensional model by generating zero-value coefficients.

[Effects of this embodiment]
According to this system, the store property recommendation server 1 learns data on stores that are performing well and data on the areas where those stores are located from the performance of multiple stores for a specific store type, to generate a learned area model 32 and a learned store model 34, extracts areas to recommend using area characteristic data based on the learned area model 32 and stores them in a recommended area DB 23, extracts properties to recommend using real estate store property data based on the learned store model 34 and stores them in a recommended property DB 24, and provides information on the extracted areas and extracted properties to a store applicant terminal 7 as a store opening report, thereby having the effect of enabling store applicants to easily find store properties that are likely to perform well in areas where they are performing well.

The present invention is suitable for a store property recommendation system and program that can recommend areas that are suitable for a store's business type and can improve business performance, and further recommend store properties in those areas that will achieve good business performance.

1...Store property recommendation server, 3...Network, 4...Real estate brokerage server, 5...Property database (DB), 6...Store prospective owner terminal, 11...Control unit, 12...Memory unit, 13...Interface unit, 20...Property database (DB), 21...Store database (DB), 22...Regional characteristics database (DB), 23...Recommended area database (DB), 24...Recommended property database (DB), 31...Area model, 32...Trained area model, 33...Store model, 34...Trained store model

Claims (6)

A store property recommendation system including a store property recommendation server that recommends store properties,
The store property recommendation server, for a specific business type, uses a trained regional model that has trained data on the regional characteristics of a region where a store with performance above a specific threshold is located from a plurality of store data including the region and performance, to determine whether the regional characteristics of a candidate region are regional characteristics suitable for a store with performance above a specific threshold, and uses the determination result to recommend a region suitable for a store of the specific business type ;
Using a trained store model that has been trained on store data of stores that have performance equal to or greater than the specific threshold based on store data of stores that have performance equal to or greater than the specific threshold, a determination is made as to whether or not a store property that is a real estate brokerage target is a store property for a store that will have performance equal to or greater than the specific threshold, and a store property for a store of the specific business type is recommended using the determination result;
A store property recommendation system characterized by generating a store opening report including a ranking of combinations of the recommended areas and the recommended store properties for a specific business type based on information on regional characteristics for stores that will perform above the specific threshold and store properties for stores that will perform above the specific threshold. The store property recommendation system of claim 1, characterized in that when the store property recommendation server determines whether the regional characteristics of a candidate area are regional characteristics suitable for a store that will achieve performance above a specific threshold, it generates a trained regional model by training a training regional model with data on the regional characteristics of stores that have performance above the specific threshold, and inputs the data on the regional characteristics of the candidate area into the trained regional model to determine whether the regional characteristics of the candidate area are regional characteristics suitable for a store that will achieve performance above the specific threshold. The store property recommendation system described in claim 1, characterized in that when the store property recommendation server determines whether a store property that is a real estate brokerage target is a store property intended for a store that will achieve performance above a specific threshold, it generates a trained store model by training a learning store model with store data of stores that have performance above the specific threshold, and inputs the data of the store property that is a real estate brokerage target into the trained store model to determine whether the store property is a store property intended for a store that will achieve performance above a specific threshold. A store property recommendation system as described in any one of claims 1 to 3, characterized in that when a wide area consisting of multiple recommended area units is specified by a person wishing to open a store, the store property recommendation server recommends areas within the wide area suitable for stores of a specific type. A store property recommendation system as described in any one of claims 1 to 4, characterized in that when the size of the store to be recommended is specified by a person wishing to open a store, the store property recommendation server recommends store properties for a specific store within the range of the size of the store . A computer program executed on a store property recommendation server that recommends store properties,
The store property recommendation server uses a trained regional model that has trained data on regional characteristics of stores with performance equal to or above a specific threshold based on regional characteristics of the region where the store has performance equal to or above a specific threshold from a plurality of store data including the region and performance, for a specific business type, to determine whether the regional characteristics of a candidate region are regional characteristics suitable for stores with performance equal to or above the specific threshold, and uses the determination result to recommend a region suitable for stores of the specific business type;
Using a trained store model that has been trained based on store data of stores that have performance equal to or greater than the specific threshold, a determination is made as to whether or not a store property that is a real estate brokerage target is a store property for a store that will have performance equal to or greater than the specific threshold, and a store property for a store of the specific business type is recommended using the determination result;
A program characterized by causing the program to function to generate a store opening report including a ranking of the combination of the recommended area and the recommended store property for a specific business type based on information on regional characteristics for stores that will perform above the specific threshold and store properties for stores that will perform above the specific threshold .