JP2022088339A - Information processing device, method, program, and computer-implemented method for providing platform - Google Patents

Abstract

To provide an information processing device, method, and program capable of efficiently determining advertising cost, and a computer-implemented method for providing a platform.SOLUTION: An information processing device 10 for determining information on advertising cost for providing advertisement information on products or services provided at least at one store operated by a business operator by an electronic first advertising medium and a non-electronic second advertising medium includes: a machine learning processing part for performing machine learning with data including the number of customers visiting the store, correlation demographic data having a correlation to the number of visiting customers, advertising cost allocated to the first advertising medium, and advertising cost allocated to the second advertising medium, as teacher data; a prediction model storage part for storing a prediction model generated by the machine learning; and an advertising cost prediction part for predicting first advertising cost to be allocated to the first advertising medium and second advertising cost to be allocated to the second advertising medium when the virtual number of visiting customers is the prescribed number of persons, from the prediction model.SELECTED DRAWING: Figure 1

Description

Information processing equipment that provides information on advertising costs.

As an information processing device that provides information related to advertisements, a device that distributes advertisement information to members living in a predetermined area according to a time zone in which the member will visit the store is known.

Japanese Patent Application No. 2009-19057

The above-mentioned conventional device is for delivering an advertisement to a predetermined member, and does not calculate the advertising cost required for delivering the advertisement.

The present invention has been made in view of the above points, and an object of the present invention is to provide an information processing apparatus capable of efficiently determining advertising costs.

The features of the information processing apparatus according to the present invention are
Information that determines information on advertising costs for providing advertising information on products or services provided at at least one store operated by a business operator by an electronic first advertising medium and a non-electronic second advertising medium. It ’s a processing device,
A teacher is instructed on data including the number of people visiting the store, the correlation demographic data having a correlation with the number of people visiting the store, the advertising cost allocated to the first advertising medium, and the advertising cost allocated to the second advertising medium. A machine learning processing unit that performs machine learning as data,
A predictive model storage unit that stores the predictive model generated by the machine learning,
Advertising cost prediction unit that predicts the first advertising cost to be allocated to the first advertising medium and the second advertising cost to be allocated to the second advertising medium when the number of virtual store visitors reaches a predetermined number from the prediction model. And to prepare.

Advertising costs can be determined efficiently.

It is a block diagram which shows the structure of the information processing apparatus 10 by this embodiment. It is a conceptual schematic diagram which shows a plurality of stores and the area associated with each store. It is a figure which shows the outline of the advertisement provided for attracting customers to a store i about one store i. It is a figure which shows a concrete example of a demographic. It is a figure (a) which shows a specific example of an advertising expense-related parameter, and the figure (b) which shows a specific example of a correction parameter. It is a figure which shows the group of the WEB advertisement expense and the paper medium advertisement expense, and the number of visitors corresponding to the group. It is a block diagram which shows the information processing apparatus 20 by 2nd Embodiment. It is a table showing the correspondence relationship of the number of people who visited the paper medium advertisement forecast corresponding to each amount of the candidate paper medium advertisement expenses for each of a plurality of stores. It is a table which shows the correspondence relation of the estimated number of visits corresponding to each amount of the candidate WEB advertising expense for each of a plurality of stores. It is a flowchart which shows the store visit number prediction model generation process. It is a flowchart which shows the store visit number calculation process.

<<<< Overview of the first embodiment >>>>>
<< First aspect >>
According to the first aspect
Information that determines information on advertising costs for providing advertising information on products or services provided at at least one store operated by a business operator by an electronic first advertising medium and a non-electronic second advertising medium. It ’s a processing device,
A teacher is instructed on data including the number of people visiting the store, the correlation demographic data having a correlation with the number of people visiting the store, the advertising cost allocated to the first advertising medium, and the advertising cost allocated to the second advertising medium. A machine learning processing unit that performs machine learning as data,
A predictive model storage unit that stores the predictive model generated by the machine learning,
Advertising cost prediction unit that predicts the first advertising cost to be allocated to the first advertising medium and the second advertising cost to be allocated to the second advertising medium when the number of virtual store visitors reaches a predetermined number from the prediction model. And, an information processing apparatus including.

The information processing apparatus according to the first aspect determines information regarding advertising costs. The advertising information is provided by an electronic first advertising medium and a non-electronic second advertising medium.

The electronic first advertising medium includes a video surface for electromagnetically displaying advertising information. In addition, advertising information may be provided by communication via a network. The electronic first advertising medium presents information obtained by visually converting electromagnetic information into advertising information. Further, the electronic first advertising medium may display information that can change with time as advertising information.

The non-electronic second advertising medium mainly refers to a medium in which advertising information is visually presented on a paper medium. The second advertising medium is not only paper, but also general paper composed of fibers such as plants, as well as cloth, resin, and the like, so that the advertising information is presented in a fixed manner without changing over time. Including the medium to be used. Further, the second advertising medium includes a medium in which another information can be presented again by thermal processing or the like. The second advertising medium is a medium that does not include a video surface for electromagnetically displaying advertising information.

Advertising information is information about products or services provided at at least one store operated by a business operator.

The information processing device includes a machine learning processing unit, a prediction model storage unit, and an advertising cost prediction unit.

The machine learning processing unit includes the number of people visiting the store, the correlation demographic data having a correlation with the number of people visiting the store, the advertising cost allocated to the first advertising medium, and the advertising cost allocated to the second advertising medium. Machine learning is performed using the data including the above as teacher data. The number of visitors to the store and the correlation demographic data can be used as explanatory variables.

Correlation demographic data is demographic data that correlates with the number of visitors to the store among the demographic data. Demographic data are demographic attributes, specifically, socio-economic characteristics data such as gender, age, area of residence, income, occupation, academic background, and family structure. By using the demographic, parameters that may be related to the store and that are highly correlated with the number of visitors can be extracted and used as teacher data.

The advertising expenses allocated to the first advertising medium and the advertising expenses allocated to the second advertising medium are actual values actually used as advertising expenses in the past. The advertising expenses allocated to the first advertising medium and the advertising expenses allocated to the second advertising medium can be used as teacher data.

The predictive model storage unit is a predictive model generated by machine learning. Any model can be used. For example, random forest, logistic regression, neural network, Markov chain Monte Carlo method, etc. can be appropriately used. Not limited to these, it can be appropriately selected.

The advertising cost prediction unit predicts the first advertising cost to be allocated to the first advertising medium and the second advertising cost to be allocated to the second advertising medium from the prediction model. The first advertising cost to be allocated to the first advertising medium and the second advertising cost to be allocated to the second advertising medium can be set as objective variables. The advertising cost prediction unit determines the first advertising cost and the second advertising cost when the virtual number of visitors to the store reaches a predetermined number from the prediction model. The predetermined number of people is preferably the maximum number of people. It should be noted that not only the maximum number of people but also the number of high-ranking people with a large number of people may be used. In consideration of the overall cost (total of the first advertising cost and the second advertising cost), the number of people may be set to a lower number.

When the first advertising medium and the second advertising medium are used, the allocation between the first advertising cost and the second advertising medium can be efficiently determined based on the past actual value.

<< Second aspect >>
The second aspect is, in the first aspect,
The teacher data further includes at least one of the number, recognition or popularity of the stores.

The number of stores is the number of stores included in a predetermined area and range, and may be density. When the number of stores included in a predetermined area or range is large, it is possible to attract customers even if they are crowded and judge that they are popular. For example, it can be used as an index when deploying in urban areas. If it is low, it can be judged that it is not very popular, and it can be used as an index when expanding to rural areas.

<< Third aspect >>
The third aspect is, in the first aspect,
Further provided is a correlated demographic data extraction unit that extracts demographic data having a positive correlation with the number of visitors to the store as the correlated demographic data among the demographic data based on the type of business of the store.

Since the correlation demographic data is extracted in advance, it is possible to efficiently determine the allocation between the first advertising cost and the second advertising medium. In addition, demographics can be selected according to the type of industry. It can be expanded not only to one industry but also to other industries. For example, it can be used not only in home improvement stores but also in supermarkets and eating and drinking.

<< Fourth aspect >>
The fourth aspect is, in the first aspect,
The machine learning processing unit corrects the teacher data and performs machine learning according to a timing factor or a regional factor in which the store is located.

The teacher data is, in particular, the number of visitors, and the number of visitors can be corrected according to timing factors and regional factors.

<< Fifth aspect >>
The fifth aspect is the fourth aspect in the fourth aspect.
The temporal factor is a factor determined based on a periodic temporal unit including a plurality of days.

Since it repeats periodically, it can be appropriately corrected according to the tendency of change, and an accurate prediction model can be generated. The time unit can be a periodically repeating unit such as a week or a month.

<< Sixth aspect >>
The sixth aspect is the first aspect.
The machine learning processing unit has an accuracy verification unit that verifies the accuracy of the prediction model.
The machine learning processing unit generates the prediction model with high accuracy for all the stores of the plurality of stores.

It is possible to create a versatile prediction model for multiple stores, making it easier to make predictions when creating a new store.

<< Seventh aspect >>
The seventh aspect is the sixth aspect in the sixth aspect.
Based on the accuracy of the prediction model of the accuracy verification unit, it is shown whether the correlation demographic data extracted by the correlation demographic data extraction unit is appropriate or inappropriate.

A predictive model can be constructed by selecting the appropriate demographic data from multiple demographic data.

<< Eighth aspect >>
The eighth aspect is, in the first aspect,
The first advertising medium is an advertising medium provided via a network.

<< Ninth aspect >>
The ninth aspect is, in the first aspect,
The second advertising medium is a paper medium.

<< 10th aspect >>
The tenth aspect is the tenth aspect in the first aspect.
The advertising cost prediction unit calculates a virtual number of visitors by giving the first advertising cost allocated to the first advertising medium and the second advertising cost allocated to the second advertising medium to the prediction model.

<<<<Details of this embodiment >>>>>
Hereinafter, embodiments will be described with reference to the drawings.

FIG. 1 is a block diagram showing a configuration of an information processing apparatus 10 according to the present embodiment.

<<< Information processing device 10 >>
The information processing apparatus 10 constructs a predetermined model from demographics, correction parameters, store feature parameters, and advertising cost-related parameters, and allocates WEB advertising costs and paper advertising costs from the constructed model. It is a device to present. The advertisement mainly includes a WEB advertisement provided via a network and a paper medium advertisement provided by a paper medium such as a leaflet.

<< Advertiser >>
The advertiser in the information processing apparatus 10 can be any business operator. In this embodiment, a retail store that sells products related to daily goods and housing equipment, a so-called home center, is taken as an example. The business operator is not limited to this, and a business operator that has a plurality of stores such as supermarkets, department stores, restaurants such as restaurants, etc. is preferable.

When the advertiser is engaged in business in a plurality of industries, it is preferable to process the information processing apparatus 10 for each industry. Conditions and environment may differ depending on the type of industry, and it may not be possible to make an accurate judgment.

FIG. 2 is a conceptual schematic diagram showing a plurality of stores and areas related to each store. In FIG. 2, one business operator has a plurality of stores (store 1, store 2, ..., Store n) for one industry, and the areas that can be related to each store are indicated by a broken line circle. rice field. Residents who may visit store 1 are those who live within the circle of area 1. Residents who may visit store 2 are those who live within the circle of area 2. Residents who may visit the store n are those who live within the circle of the area n.

Those who may visit one store are affected by the population density around one store and the environment such as the means of transportation around it.

FIG. 3 is a diagram showing an outline of an advertisement provided for attracting customers to the store i for one store i. As described above, the advertisements to be provided include WEB advertisements and paper advertisements. That is, the advertisement information about the store i is provided to the residents in the area who may visit the store i by the WEB advertisement and the paper medium advertisement. It is preferable that the range in which the WEB advertisement and the paper medium advertisement are provided is a wide range centered on the location of the store i. In addition, it is preferable that the range of providing WEB advertisements and paper advertisements is broadly defined depending on the type of industry. For example, in the case of stores such as supermarkets that mainly handle fresh food, the range of advertisements is narrowed, and in the case of stores such as home improvement stores that mainly handle daily necessities, the range of advertisements is widened. be able to. The range of providing WEB advertisements and paper advertisements can be appropriately determined according to past achievements and the like.

<< Paper Advertising >>
Paper-based advertisements are advertising media such as leaflets and leaflets that have existed for some time. Some are provided directly by hand, while others are included in newspapers. Furthermore, advertising information may be printed on paper, such as newspapers, magazines, free papers provided free of charge, and free magazines.

The cost of paper advertisements is determined according to the total number of insert advertisements, the number of insert advertisements by content, and the number of insert advertisements by advertisement type.

<< WEB Advertising >>
The WEB advertisement may be any advertisement information provided via an online network or e-mail. The network may be wired or wireless. The network may be any medium through which information can be transmitted. Currently, it is diversifying, and advertisements are delivered to various devices such as PCs, smartphones, tablets, and connected TVs by text information, still images, moving images, and voices.

For example, the WEB advertisement includes a WEB advertisement cost by content, a WEB advertisement cost by advertisement type, a WEB advertisement cost by platform, and the like. In addition, WEB advertisements include search-linked advertisements, display advertisements, and local campaigns.

A search-linked advertisement is an advertisement that is displayed in conjunction with a keyword searched by a user using a search engine. Display advertisements are text-type or banner-type advertisements that are displayed in inventory on websites and apps. A local campaign is an advertisement that is displayed along with a map displayed on a website or application, and is an advertisement that automatically displays the content to residents in a range that may be related to store i. ..

<< Configuration of Information Processing Device 10 >>
The information processing apparatus 10 includes a correlation extraction unit 100, a machine learning processing unit 200, an advertising cost allocation calculation unit 300, and a data storage unit 400.

<< Correlation extraction unit 100 >>
The correlation extraction unit 100 calculates the correlation coefficient of the second parameter with respect to the first parameter, and extracts the second parameter having a correlation with the first parameter. Here, it is preferable to extract parameters having a high positive correlation. In addition, parameters having a negative correlation may be extracted.

The first parameter in this embodiment is the number of visitors to the store. The second parameter includes feature quantities such as demographics, store feature parameters, and advertising cost-related parameters.

<Demo graphic>
Demographics are socio-economic trait data that each person has, such as demographic attributes, specifically gender, age, region of residence, income, occupation, academic background, and family structure. For example, socio-economic characteristic data provided by people living in areas where they may visit stores. It is possible to determine what kind of attribute an individual has easily visits the target store. Demographics may contain parameters that affect the number of visitors.

Demographics that are generally available can be used. In particular, it is preferable that the contents are regularly updated by public institutions such as government agencies. It can be judged by reliable data. In addition to that, information issued by various institutions such as economic research institutes can be used. More detailed information and the latest information can be used.

FIG. 4 is a diagram showing a specific example of the demographic. As demographics, data on population, households, businesses / employees, commerce, consumer spending, income / annual income, savings, automobiles, welfare, areas, land prices, etc. can be used. Specifically, when processing the advertising expenses of store i, the population, households, business establishments / employees, commerce, consumption expenditure, income / annual income, etc. in the area or range that may be related to store i, Use data on savings, automobiles, welfare, area, land prices, etc.

<Store feature parameters>
Store feature parameters include the number of stores, awareness, and popularity. In addition to the total number of stores, the number of stores can be the number of stores deployed in a certain area. If the total number of stores is large, it can be judged that it has permeated the daily life and that it is highly reliable. The ratio of own stores to the total number of stores in the same industry and the total store area in the same industry also affects the degree of recognition. In addition, it is possible to determine whether or not the store is rooted in the area and is used based on the number of stores opened in a certain area and the number of years that have passed since the store was opened. The number of stores may be set to the population of a predetermined area. Awareness and popularity can be obtained from the repeat rate, questionnaires, and the number of reviews posted on the Internet.

In this way, the store characteristic parameter can be used as a measure of evaluation of each store for each store. Store feature parameters may include parameters that affect the number of visitors.

<Advertising cost related parameters>
The advertising cost-related parameter is a parameter related to the cost used for the advertisement related to the target store. Advertising cost-related parameters include the level of advertising cost and the type of advertisement. In the present embodiment, advertisement information about the target store is provided by a WEB advertisement and a paper medium advertisement. The advertising cost-related parameter is a parameter related to the cost required for the WEB advertisement and the paper medium advertisement. Specifically, the advertising cost-related parameter includes a parameter related to the WEB advertising cost and a parameter related to the paper medium advertising cost.

Advertising cost-related parameters such as high and low advertising costs and types of advertisements can affect the number of customers visiting the store.

FIG. 5A is a diagram showing a specific example of advertising cost-related parameters.

For example, there is a relationship between the number of residents who actually visited the store among the residents who displayed the advertisement and the WEB advertising cost. More specifically, when an advertisement related to store i is displayed on a website displayed on a computer used by a resident or an application to be executed, the display of the advertisement triggers the actual display of the advertisement in store i. This is the relationship between the number of residents who visited the store and WEB advertising expenses. It should be noted that the relationship may be a relationship using the number of residents who visited the store or the WEB advertising cost directly, or a relationship using a value processed by a mathematical function. As mentioned above, the number of visitors may be correlated.

In addition, there is a relationship between the demographic element (first element to nth element) and the paper advertising cost. For example, the demographic element can be the number of households, population, or savings in the demographic described above (FIG. 4). As mentioned above, it suffices as long as it has a correlation with the number of visitors.

The above-mentioned demographics, store feature parameters, and various feature quantities included in advertising cost-related parameters may be used directly to calculate the correlation coefficient, but various functions, such as mathematical functions such as logarithms, may be calculated. The correlation coefficient may be calculated using the function value of.

<Correction parameter>
The correction parameter is a parameter for correcting various feature amounts included in the above-mentioned demographic, store feature parameter, and advertising cost-related parameter. Even if the parameters are judged to have low correlation, there is a possibility that the correlation can be enhanced by correcting them.

In this embodiment, the correction parameter includes a geographical factor parameter, a timing factor parameter, and a special factor parameter.

Geographic factor parameters are parameters based on the location of the store. For example, the feature amount is corrected according to the store located in the urban area, the store located in the rural area, the distance from the nearest station or the main road, the size of the parking lot, and the like. For example, when a business operator (advertiser) has multiple stores and there is a difference in the feature amount among the multiple stores, the feature amount is aligned according to the geographical factor parameter, so that the correlation can be obtained for the multiple stores. There is a possibility of increasing.

The time factor parameter corrects the feature amount according to the time zone (morning, afternoon, midnight), day of the week, holiday, and other time factors. In addition, seasonal factors (year-end, New Year holidays, Obon, Golden Week, etc.) can also include temporal factor parameters. By aligning the features in a specific period with the features in another period, there is a possibility of enhancing the correlation.

Special factor parameters are parameters due to sudden events or specific events that are neither geographical nor temporal. For example, there are disasters such as earthquakes and typhoons, and the effects of sudden economic changes. By aligning the features in the area or period where a sudden event or a specific event occurred with the features in other periods, there is a possibility of enhancing the correlation.

By correcting various feature quantities according to the tendency of these correction parameters, there is a possibility of enhancing the correlation.

FIG. 5B is a diagram showing a specific example of the correction parameter. Specifically, it can be correction information by Golden Week, correction information by Obon, correction information by the end of the year, correction information by typhoon, and the like.

<< Machine Learning Processing Unit 200 >>
The machine learning processing unit 200 has a machine learning unit 210, an accuracy verification unit 220, and a prediction model storage unit 230.

<Machine learning unit 210>
The machine learning unit 210 uses the values of various parameters extracted by the correlation extraction unit 100 as teacher data to generate a prediction model. Predictive model generation involves selecting the model to use and machine learning for the selected model. Predictive models are generated by machine learning. Machine learning uses the above-mentioned demographics, store feature parameters, advertising cost-related parameters, and other feature quantities.

<Predictive model>
As a concept of the prediction model M, for example, M (f, g, h) (here, f (WEB advertising cost, correction parameter), g (paper advertisement cost, correction parameter), and h (demographic). , Correction parameter)) can be generated using the relationship. f is a function related to WEB advertising expenses, and is corrected according to geographical factors, timing factors, and special factors by correction parameters. g is a function related to paper advertising expenses, and is corrected according to geographical factors, timing factors, and special factors by correction parameters. h is a function related to demographics, and is corrected according to geographical factors, timing factors, and special factors by correction parameters.

Further, as another concept of the prediction model M, M (f, g, h) (here, f (WEB advertising cost), g (paper medium advertising cost), and h (demographic, correction parameter)). ) Can be generated using the relationship. f is a function related to WEB advertising expenses. g is a function related to paper advertising expenses. h is a function related to demographics, and is corrected according to geographical factors, timing factors, and special factors by correction parameters. In this model, only demographics are corrected. A simple model can be provided.

Further, as another concept of the prediction model M, M (f, g, h, i) (here, f (WEB advertising cost), g (paper advertising cost), h (demographic), It can be generated using the relationship i (correction parameter)). f is a function related to WEB advertising expenses. g is a function related to paper advertising expenses. h is a function related to demographics. i is a function related to the correction parameter. By making the correction parameter a separate function, it is possible to more accurately correct the WEB advertising cost, the paper medium advertising cost, and the demographic.

The prediction model is not limited to these, and other concepts may be used. It may be used as appropriate according to the type of business.

<Accuracy verification unit 220>
The accuracy verification unit 220 verifies the prediction model generated by the machine learning unit 210. For verification, demographics not used in machine learning, store feature parameters, advertising cost-related parameters, etc. are used. For verification, for example, a method such as RMSE (Root Mean Squared Error) (root mean squared difference) or MAE (Mean Absolute Error) average absolute error can be used. Specifically, the verification uses predicted values predicted by a prediction model generated by machine learning, and actual measured values such as actual demographics, store feature parameters, and advertising cost-related parameters.

If the error is large as a result of the verification, return to the machine learning unit 210, select another model, perform machine learning again to generate a prediction model, and the accuracy verification unit 220 again generates the prediction. Validate the model.

<Predictive model storage unit 230>
The predictive model storage unit 230 stores various predictive model data required for the predictive model, such as the type of the predictive model and the coefficient of the predictive model determined by the accuracy verification unit 220, which is generated by the machine learning unit 210.

By reading out the predictive model data stored in the predictive model storage unit 230, it is possible to execute the process using the confirmed predictive model.

<< Advertising cost allocation calculation unit 300 >>
The advertising cost allocation calculation unit 300 calculates the number of visitors according to the WEB advertising cost and the paper medium advertising cost by using the prediction model generated and confirmed by the machine learning processing unit 200, and presents it to the user.

For example, the number of visitors is calculated by giving a predetermined WEB advertising cost and paper media advertising cost to the prediction model. Specifically, a combination of at least one WEB advertising expense and at least one paper-based advertising expense is determined in advance, and a set of the WEB advertising expense and the paper-based advertising expense is given to the prediction model to WEB. Calculate the number of visitors corresponding to the combination of advertising expenses and paper advertising expenses. The combination of the WEB advertising cost and the paper medium advertising cost that maximizes the number of visitors is presented to the user as the optimum distribution of the WEB advertising cost and the paper medium advertising cost. The optimum distribution of WEB advertising expenses and paper advertising expenses is displayed on a display or the like (not shown) connected to the information processing apparatus 10.

In addition to the allocation of WEB advertising expenses and paper advertising expenses that maximize the number of visitors, it is also possible to present some of the top allocations with a large number of people. Distribution priorities may be set according to various conditions such as the total amount of WEB advertising expenses and paper advertising expenses, the convenience of purchasing products, and various conditions such as store renovation.

FIG. 6 is a diagram showing a set of WEB advertising expenses and paper advertising expenses and the number of visitors corresponding to the set. The result of calculating the number of visitors corresponding to the set of the WEB advertising cost and the paper medium advertising cost by appropriately determining the set of the WEB advertising cost and the paper medium advertising cost and using the generated prediction model is shown. In the example shown in FIG. 6, when the WEB advertising cost is A and the paper medium advertising cost is B, the expected number of visitors to the store is the largest.

<< Data storage unit 400 >>
The data storage unit 400 stores data of various parameters used in the correlation extraction unit 100 and the machine learning processing unit 200. For example, the data storage unit 400 may include demographics, correction parameters (geographical factor parameters, timing factor parameters, special factor parameters), store feature parameters, advertising cost-related parameters (WEB advertising costs, paper advertising costs), and the like. Stores various information such as numerical values.

Various types of information stored in the data storage unit 400 are appropriately collected and updated with the latest information. The timing of the update may be daily, weekly, monthly, or when something unexpected happens or changes occur. Even if the prediction model is fixed at a certain point in time, it may be necessary to change the prediction model due to various events that occur after that.

When various information stored in the data storage unit 400 is changed, the machine learning unit 210 is returned to, machine learning is performed again to generate a prediction model, and the accuracy verification unit 220 again uses the generated prediction model. Verify.

By doing so, it is possible to generate and use an optimal prediction model according to changes in the environment and circumstances.

<<<<<< Second Embodiment >>>>>
In the second embodiment, a prediction model is generated by distinguishing between the number of people who visit the store without providing the advertisement and the number of people who visit the store when the advertisement is provided. The number of people who will visit the store is calculated from the generated prediction model.

Further, in the second embodiment, the allocation of the cost required for the paper advertisement and the cost required for the WEB advertisement is determined from the number of visitors calculated from the prediction model. The paper advertisement and the WEB advertisement are the same as those in the first embodiment.

<< 11th aspect >>
According to the eleventh aspect
Information that determines information on advertising costs for providing advertising information on products or services provided at at least one store operated by a business operator by an electronic first advertising medium and a non-electronic second advertising medium. It ’s a processing device,
Past information including the number of past visitors to the store, the past advertising expenses required to provide the first advertising medium, and the past advertising expenses required to provide the second advertising medium are used as teacher data. ,
An advertisement-visiting number model for calculating the estimated number of ad-visiting number of customers who will visit a store when an advertisement is provided using at least one of the first advertising medium and the second advertising medium (for example, described later). Advertising effect prediction model (paper advertisement store visitor number prediction model, WEB advertisement visitor number prediction model), etc.
An advertisement-free store visitor model for calculating the number of unpredicted number of advertisements that will visit a store when an advertisement using both the first advertisement medium and the second advertisement medium is not provided (for example, a base described later). A machine learning processing unit (for example, a machine learning processing unit described later) that generates a line prediction model by machine learning, and a machine learning processing unit (for example, a machine learning processing unit described later).
Based on the predicted number of visitors with advertisement and the number of unpredicted visits to the store, the first advertising cost required to provide the first advertising medium and the second advertising cost required to provide the second advertising medium are calculated. An information processing device including an advertising cost determination processing unit for determining (for example, an advertising cost-estimated number of customers visiting a store generation unit, which will be described later) is provided.

The information processing device generates two types of prediction models, an advertisement-visited store number model and an advertisement-less-visited number model, by machine learning.

Analyzing the visitors who come to the store, it can be mainly classified into the following four patterns.
(1) A person who has been stimulated by an advertisement and has selected the store by the advertisement.
(2) Those who were originally using the store regardless of the presence or absence of advertisements, which stimulated demand by advertisements.
(3) Those who originally had demand regardless of the presence or absence of advertisements and selected the store by advertisement.
(4) Those who originally had demand regardless of the presence or absence of advertisements and originally used the store regardless of the presence or absence of advertisements.

The above-mentioned visitors (1) to (3) are those who are generated by the effect of the advertisement regarding the demand or the store. On the other hand, the visitor in (4) is a person who is not affected by the effect of the advertisement. The eleventh aspect is an information processing device that calculates the total number of visitors to the store by calculating the number of unpredicted advertisement visitors, which is the number of visitors in (4).

The advertisement-free number of visitors model is a model for calculating the number of visitors to a store when advertisements using both the first advertisement medium and the second advertisement medium are not provided. The advertisement-free number-of-store model is a predictive model for calculating the number of people who will visit a store without actively providing advertisements such as the first advertising medium and the second advertising medium. Aggressive advertising refers to advertising that is intended to be communicated to society in general. It does not include information contained in passive advertisements other than the first advertising medium and the second advertising medium, such as word-of-mouth and private transmission. Negative advertisements are advertisements that are not intended to be conveyed to society in general, and refer to information that accidentally exerts an advertising function.

The advertisement-visited store number model is a predictive model for calculating the number of people who will visit the store by providing advertisements using the first advertising medium and the second advertising medium. That is, it is a model for calculating the number of people who may come to the store by active advertisement, that is, an advertisement issued with the intention of communicating to society in general.

The first advertising cost and the second advertising cost are determined based on the predicted number of visitors with advertisement and the number of unpredicted visitors with advertisement. In general, if the amount of advertising expenses is large, the number of people visiting the store will be large, and if the amount of advertising expenses is small, the number of people visiting the store will be small. In this model of the number of people visiting the store with advertisements, even if the amount of advertising expenses is increased to some extent or more, the number of people visiting the store is saturated, and a model that is in line with reality is adopted. For example, a model using a logistic regression formula using a sigmoid function can be used as an advertisement visitor number model. Depending on the amount of advertising expenses, the number and frequency of advertising will be determined, and the ease with which advertisements will reach prospective customers will change. By using the advertising store visitor model, it is possible to generate a correspondence between the predicted number of store visits and advertising costs.

The predicted number of visitors can be determined by the number of predicted visitors with advertisements and the number of unpredicted visitors with advertisements. For example, the sum of the predicted number of visitors with advertisement and the number of unpredicted visitors with advertisement can be the predicted number of visitors. The first advertising cost and the second advertising cost such that the predicted number of store visits is the number of store visits desired by the business operator can be determined by referring to the correspondence between the predicted number of store visits and the advertising cost.

<< 12th aspect >>
The twelfth aspect is the eleventh aspect in the eleventh aspect.
The advertising expense determination processing unit
From the above-mentioned model of the number of people visiting the store with advertisements
The estimated number of visitors to the store with the first advertisement corresponding to each of the plurality of first candidate advertisement expenses that are candidates for the first advertisement expense, and
The estimated number of visitors to the store with the second advertisement corresponding to each of the plurality of second candidate advertisement expenses, which are candidates for the second advertisement expense, is calculated.
The first correspondence relationship in which the estimated number of visitors with the first advertisement is associated with each of the plurality of first candidate advertising expenses, and
A second correspondence relationship is generated in which the estimated number of visitors with the second advertisement is associated with each of the plurality of second candidate advertisement expenses.

A plurality of first candidate advertising expenses are predetermined as candidates for the first advertising expenses. At least one of the plurality of first candidate advertising expenses is selected and becomes the first advertising expense. By using the model of the number of people visiting the store with advertisements and the model of the number of people visiting stores without advertisements, it is possible to calculate the estimated number of people visiting the store with advertisements corresponding to each of the plurality of first candidate advertising expenses. It is possible to generate a first correspondence relationship in which the calculated number of people who visit the store with the first advertisement is associated with each of the plurality of first candidate advertisement costs.

Similarly, a plurality of second candidate advertising expenses are predetermined as candidates for the second advertising expenses. At least one of the plurality of second candidate advertising expenses is selected as the second candidate advertising expense. By using the model of the number of people visiting the store with advertisements and the model of the number of people visiting stores without advertisements, it is possible to calculate the estimated number of people visiting the store with advertisements corresponding to each of the plurality of second candidate advertising expenses. It is possible to generate a second correspondence relationship in which the calculated number of people who visit the store with the second advertisement is associated with each of the plurality of second candidate advertisement costs.

With reference to the generated first correspondence relationship, the first advertising cost can be determined based on the number of visitors desired by the business operator. With reference to the generated second correspondence relationship, the second advertising cost can be determined based on the number of visitors desired by the business operator.

<< 13th aspect >>
The thirteenth aspect is the twelfth aspect.
The advertising expense determination processing unit
For each of the multiple stores
The first correspondence and the second correspondence are generated.

Since the first correspondence relationship and the second correspondence relationship are generated for each of the plurality of stores, the first advertising cost and the second advertising cost based on the desired number of visitors can be determined for each store.

The overall first and second advertising costs can be determined for some stores so that the total number of visitors to some stores is the desired number of visitors. For example, it is possible to easily set a budget for advertising expenses for some stores included in a given area.

<< 14th aspect >>
The fourteenth aspect is the thirteenth aspect.
The advertising expense determination processing unit
Based on the desired number of visitors to at least one store, the first expected number of customers with advertisements and the second expected number of customers with advertisements are determined with reference to the first correspondence relationship and the second correspondence relationship.
Based on the determined number of people visiting the store with the first advertisement, one of the plurality of first candidate advertising expenses is selected as the first advertising expense with reference to the first correspondence relationship.
Based on the determined number of people visiting the store with the second advertisement, the second correspondence relationship is referred to, and one of the plurality of second candidate advertising expenses is selected and used as the second advertising expense. Includes selection process.

The first advertising cost and the second advertising cost can be determined so that the desired expected number of visitors to the store is realized.

<< Fifteenth aspect >>
The fifteenth aspect is the eleventh aspect.
The machine learning processing unit
It has an information selection process for selecting information to be used as the teacher data according to the type of business of the store.

The types of information in stores are diverse, and the types of information that contribute to the number of visitors differ depending on the type of store. By making it possible to select the information used as teacher data according to the type of business of the store, it is possible to generate a highly accurate model of the number of people visiting the store with advertisements and the model of the number of people visiting the stores without advertisements.

<< 16th aspect >>
The sixteenth aspect is the twelfth aspect.
The advertising expense determination processing unit
It has a predicted number of visitors adjusting process for adjusting the first predicted number of visitors with advertisements and the second predicted number of visitors with advertisements based on the characteristic information of the store.

It may be difficult to include information indicating the characteristics of the store in the teacher data. By adjusting the number of visitors calculated by the model of the number of visitors with advertisements and the model of the number of visitors without advertisements based on the characteristics of the store, it is possible to obtain the number of visitors according to the characteristics of the actual store. You can make a more appropriate budget.

<< 17th aspect >>
According to the seventeenth aspect
A method of determining information on advertising costs for providing advertising information on products or services provided in at least one store operated by a business operator by an electronic first advertising medium and a non-electronic second advertising medium. And
Past information including the number of past visitors to the store, the past advertising expenses required to provide the first advertising medium, and the past advertising expenses required to provide the second advertising medium are used as teacher data. ,
An advertisement-visiting number model for calculating the estimated number of ad-visiting number of customers who will visit a store when an advertisement is provided using at least one of the first advertising medium and the second advertising medium (for example, described later). Advertising effect prediction model (paper advertisement store visitor number prediction model, WEB advertisement visitor number prediction model), etc.
An advertisement-free store visitor model for calculating the number of unpredicted number of advertisements that will visit a store when an advertisement using both the first advertisement medium and the second advertisement medium is not provided (for example, a base described later). A machine learning step (for example, processing by a machine learning processing unit described later) that generates a line prediction model by machine learning, and
Based on the predicted number of visitors with advertisement and the number of unpredicted visits to the store, the first advertising cost required to provide the first advertising medium and the second advertising cost required to provide the second advertising medium are calculated. A method is provided that includes an advertising cost determination step to be determined (for example, an advertising cost described later-a process by a generation unit corresponding to an estimated number of visitors).

<< 18th aspect >>
According to the eighteenth aspect
A program that determines information on advertising costs for providing advertising information on products or services provided at at least one store operated by a business operator by an electronic first advertising medium and a non-electronic second advertising medium. And
Past information including the number of past visitors to the store, the past advertising expenses required to provide the first advertising medium, and the past advertising expenses required to provide the second advertising medium are used as teacher data. ,
An advertisement-visiting number model for calculating the estimated number of ad-visiting number of customers who will visit a store when an advertisement is provided using at least one of the first advertising medium and the second advertising medium (for example, described later). Advertising effect prediction model (paper advertisement store visitor number prediction model, WEB advertisement visitor number prediction model), etc.
An advertisement-free store visitor model for calculating the number of unpredicted number of advertisements that will visit a store when an advertisement using both the first advertisement medium and the second advertisement medium is not provided (for example, a base described later). A machine learning step (for example, processing by a machine learning processing unit described later) that generates a line prediction model by machine learning, and
Based on the predicted number of visitors with advertisement and the number of unpredicted visits to the store, the first advertising cost required to provide the first advertising medium and the second advertising cost required to provide the second advertising medium are calculated. A program including an advertising cost determination step to be determined (for example, an advertising cost described later-processing by a generation unit corresponding to the estimated number of visitors to the store) is provided.

<< 19th aspect >>
According to the nineteenth aspect
A platform for determining advertising costs for providing advertising information on products or services provided in at least one store operated by a business operator by an electronic first advertising medium and a non-electronic second advertising medium. Is a computer implementation method that provides
Past information including the number of past visitors to the store, the past advertising expenses required to provide the first advertising medium, and the past advertising expenses required to provide the second advertising medium are used as teacher data. ,
An advertisement-visiting number model for calculating the estimated number of ad-visiting number of customers who will visit a store when an advertisement is provided using at least one of the first advertising medium and the second advertising medium (for example, described later). Advertising effect prediction model (paper advertisement store visitor number prediction model, WEB advertisement visitor number prediction model), etc.
An advertisement-free store visitor model for calculating the number of unpredicted number of advertisements that will visit a store when an advertisement using both the first advertisement medium and the second advertisement medium is not provided (for example, a base described later). A machine learning processing unit (for example, a machine learning processing unit described later) that generates a line prediction model by machine learning, and a machine learning processing unit (for example, a machine learning processing unit described later).
Based on the predicted number of visitors with advertisement and the number of unpredicted visits to the store, the first advertising cost required to provide the first advertising medium and the second advertising cost required to provide the second advertising medium are calculated. A computer implementation method is provided that provides a platform including an advertising cost determination processing unit (for example, an advertising cost-estimated number of customers visiting a store, which will be described later) to be determined.

For example, a platform can be constructed on at least one information processing device such as a server configured to be able to communicate via a network. The platform may be constructed by distributing it to a plurality of information processing devices. The program of the eighteenth aspect described above can be run on the platform. The computer mounting method of the nineteenth aspect is a method of mounting the program of the eighteenth aspect on a computer in order to provide such a platform.

<<< Definition of terms in the second embodiment >>>
<< Types of number of visitors >>

<desired number of visitors>
The desired number of visitors means the number of visitors desired by the business operator (store).

<Number of people visiting the store>
The number of people who have visited the store in the past means the number of people who have visited the store so far, mainly for machine learning.

<Predicted number of visitors>
The estimated number of visitors means the estimated number of visitors calculated by executing the prediction model. The sum of the estimated number of visitors to the base and the estimated number of visitors to the advertising effect, which will be described later, is the estimated number of visitors.

<Base forecast number of visitors>
The base predicted number of visitors means the number of visitors calculated by executing the baseline prediction model.

<Advertising effect forecast number of visitors>
The number of people who visit the store for predicting the advertising effect means the number of people who visit the store for predicting the advertising effect calculated by executing the advertising effect prediction model. The sum of the expected number of visitors to the paper-based advertisement and the estimated number of visitors to the WEB advertisement, which will be described later, is the estimated number of visitors to the advertising effect. The advertising effect prediction model has a paper-based advertisement store visitor number prediction model and a WEB advertisement store visitor number prediction model.

<Forecast number of visitors to paper advertisements>
The number of people visiting a store for which a paper advertisement is predicted is the number of people visiting the store calculated by executing the model for predicting the number of people visiting a paper advertisement. The number of people who will visit the store by providing the paper advertisement is the number of people who are expected to come to the store by providing the paper advertisement.

<Predicted number of visitors to WEB advertisements>
The WEB advertisement forecast number of visitors is the number of visitors calculated by executing the WEB advertisement visit number prediction model. The number of people who visit the store is predicted to be the number of people who will come to the store by providing the WEB advertisement.

<< Advertising expenses >>
<Collecting advertising expenses>
Collected advertising costs mean past advertising costs collected so far, mainly for machine learning.

<Candidate advertising expenses>
Candidate advertising expenses are advertising expenses that will be required to provide advertisements. Candidate advertising costs have multiple candidate amounts to select a specific amount.

<<< Information processing device 20 >>
FIG. 7 is a block diagram showing the information processing apparatus 20 according to the second embodiment.

The information processing device 20 includes a processor (CPU (central processing unit), etc.), ROM (read-only memory), RAM (random access memory), HDD (hard disk drive), SSD (solid state drive), and I / F (communication interface). It is equipped with a device) and an input operation device (keyboard, mouse, touch panel, etc.) (not shown). The information processing device 20 can be various processing devices such as a personal computer, a tablet computer, and a portable terminal device.

The information processing device 20 is a device capable of executing various processes such as various arithmetic processes, various data processes, and various communication processes with other devices (servers, terminal devices, etc.).

As shown in FIG. 7, the information processing apparatus 20 has a machine learning processing unit, a data storage unit, an advertising cost-estimated number of store visitor generation unit, and a store-related feature adjustment unit. The data storage unit stores various values and information of learning parameters, advertising expenses-estimated number of customers visiting the store, and store-related feature parameters.

<< Machine Learning Processing Department >>
The machine learning processing unit has a machine learning parameter selection unit, a baseline prediction model generation, an advertising effect prediction model generation, a prediction model accuracy verification unit, and a prediction model storage unit. Use appropriate machine learning. For machine learning, for example, random forest, logistic regression, neural network, Markov chain Monte Carlo method, etc. can be appropriately used. Not limited to these, machine learning can appropriately utilize an appropriate algorithm.

<< Machine learning parameter selection unit >>
The machine learning parameter selection unit selects the parameters used for machine learning.

<Parameter 1 used for machine learning>
The parameters used for machine learning are mainly the number of people who come to the store, the type of store, the sales floor area of the store, the number of competitors within 5 km, the reach rate of paper advertisements, the number of weeks (seasonal), and the issuance of paper advertisements. There are the number of copies, the amount of WEB advertisements, etc.

The number of people visiting the store is the number of people visiting each store in the past that has been collected so far. This is the total number of people who have visited each of the multiple stores so far.

The types of stores are classified according to the type of business. The types of stores include, for example, supermarkets, drug stores, discount stores, home improvement stores, and apparel. The types of stores are not limited to these, and can be various types.

The sales floor area of a store is the area occupied by the sales floor of the store. In other words, the sales floor area of a store is the total floor area of the sales floor actually used to sell a product. The sales floor area of the store is the area occupied by the sales floor, and the area of the parking lot or the like is excluded.

The number of competitors within 5 km is the number of competing stores located within a radius of 5 km around the store. The radius is not limited to 5 km, and a length that affects the store may be appropriately selected. The radius should be appropriate so that the relationship between advertising costs and the number of visitors can be analyzed accurately.

The paper advertisement reach rate is the ratio of the paper advertisements delivered to prospective customers among all the distributed paper advertisements. Paper advertisements include, for example, insert leaflets attached to newspapers. When the paper advertisement is a newspaper insert leaflet, the newspaper subscription rate can be set to the paper advertisement reach rate. The newspaper subscription rate is, for example, the number of home-delivered newspaper subscriptions to households within 5 km from the store. In other words, it is the number (percentage) that can deliver the insert leaflet attached to the newspaper. In this way, by using the subscription rate of newspapers, the reach rate of paper advertisements can be concretely quantified to ensure objectivity. For this reason, it becomes easy to accurately analyze and judge the advertising effect of the leaflet attached to the newspaper and distributed.

Paper advertisements are not limited to leaflets attached to newspapers. For example, there are various forms of paper media and various paper media advertisements such as distribution of leaflets by posting and distribution of free paper in a specific area. Even with such a paper medium, it suffices as long as the advertising reach rate can be concretely quantified and the advertising effectiveness can be judged. In any case, if it is a paper medium that can objectively determine the number and ratio of paper advertisements delivered to prospective customers, it can be targeted for the paper advertisement reach rate regardless of the type of paper media. can.

The number of weeks (seasonal) is a serial number assigned in order from 1 for each week (every 7 days) throughout the year from the beginning of the year to the end of the year (all year). Every time the year changes, it is numbered again from 1. Depending on the number of weeks, it is possible to analyze and identify the seasonality of the number of visitors in detail.

The circulation of paper advertisements is the number of paper advertisements issued so far. The cost required to provide the paper advertisement can be calculated from the number of copies of the paper advertisement issued.

The WEB advertisement placement amount is the number of WEB advertisements displayed so far on a display such as a terminal device via a network. The cost required to provide the WEB advertisement can be calculated from the amount of the WEB advertisement placed.

<Parameter 2 used for machine learning>
The parameters used for machine learning are not limited to those described above, but other parameters can also be used. The parameters used for machine learning may be any parameters that can collect values and information and have the possibility of predicting the number of visitors. For example, the purchase possibility parameter indicating the possibility that the product handled in the store will be purchased, the service availability parameter indicating the possibility that the service provided by the store will be used, and the movement regarding how to access the store. Means parameters and the like can be used.

Specifically, the purchaseability parameters can be set as follows according to the business type and the like. In the case of supermarkets, grocery purchasing power can be a purchaseability parameter. In the case of drug stores, drug purchasing power can be a purchaseability parameter. In the case of discount stores, the value of consumer spending can be a purchaseability parameter. In the case of a home improvement store, the purchasing power of pet food, the purchasing power of gardening products / same goods, and the purchasing power of furniture / household goods can be used as purchaseability parameters. In the case of apparel, the purchasing power of clothing and the purchasing power of footwear can be used as purchaseability parameters.

The service availability parameters can be set as follows according to the business type and the like.
In the case of supermarkets, the number of family members and the type of household such as a single-person household can be used as service availability parameters. In the case of a drug store, the service availability parameter can be the population certified as requiring long-term care 1-5. In the case of discount stores, the area of low-income areas and man-made can be used as service availability parameters. In the case of a home improvement store, the living area, the ratio of homeowners, the number of workers in the primary industry, etc. can be used as service availability parameters. For apparel, the number of high-paying taxpayers can be used as a service availability parameter.

The mode of transportation parameters are generally railroad, private car, bus, taxi, bicycle, walking, etc. for various business formats.

The machine learning parameter selection unit appropriately selects the various parameters described above according to the type of industry and the like. That is, the machine learning parameter selection unit can select parameters for generating a prediction model according to the type of industry and the like. By selecting appropriate parameters according to the type of industry, it is possible to generate a highly accurate prediction model.

<< Baseline prediction model generator >>
The baseline prediction model generation unit generates a baseline prediction model by machine learning. The baseline prediction model is a prediction model for calculating the base prediction number of visitors. The baseline prediction model generation unit is a prediction model for calculating the number of people who will visit the store if the advertisement is not provided.

As described above, there are two types of advertisements, paper advertisements and WEB advertisements. The baseline prediction model generation unit is a prediction model for calculating the number of people who will visit the store if neither the paper advertisement nor the WEB advertisement is provided. It may be used as a model for calculating the number of people who will visit the store if only one of the paper advertisement and the WEB advertisement is not provided.

With the baseline prediction model, it is possible to calculate the base prediction number of visitors for each of a plurality of stores. Furthermore, the baseline prediction model can be used to calculate the base prediction number of visitors corresponding to the number of weeks for each of a plurality of stores.

<< Advertising effect prediction model generation unit >>
The advertising effectiveness prediction model generation unit generates an advertising effectiveness prediction model by machine learning. The advertising effect prediction model is a prediction model for calculating the number of visitors to the advertising effect prediction store. The advertising effect prediction model has a paper-based advertisement store visitor number prediction model and a WEB advertisement forecast store visitor number prediction model.

<Model for predicting the number of visitors to paper advertisements>
The number of people visiting a paper-based advertisement is calculated by the model for predicting the number of people visiting a paper-based advertisement. The paper-based advertisement visitor number prediction model is a prediction model for calculating the number of people who will visit the store triggered by the paper-based advertisement when the paper-based advertisement is provided. Using the paper-based advertisement visitor number prediction model, calculate the estimated number of paper-based advertisement visits corresponding to each of the candidate (candidate paper-based advertisement costs) amounts of the costs (paper-based advertisement costs) required to provide the paper-based advertisements. be able to.

The approximate tendency of the number of people visiting a paper-based advertisement based on the model for predicting the number of people visiting a paper-based advertisement is as follows. For example, when the amount of paper-based advertising expenses is reduced, the number of people who visit the paper-based advertising forecast decreases, and when the amount of paper-based advertising expenses is increased, the number of people who visit the paper-based advertising forecast increases. Even if the amount of paper advertising expenses exceeds a certain level, the expected number of paper advertising visitors will be saturated.

Further, the paper-based advertisement visitor number prediction model can be used to calculate the paper-based advertisement visitor number for each of a plurality of stores. Specifically, the paper medium advertisement visitor number prediction model can be used to calculate the expected number of paper medium advertisement visits corresponding to each of the amount of candidate paper medium advertisement expenses for each of a plurality of stores. Furthermore, it is also possible to calculate the estimated number of people visiting a paper advertisement store corresponding to the number of weeks for each of a plurality of stores by using the model for predicting the number of people visiting a paper advertisement store.

<Web advertising visitor number prediction model>
The number of WEB advertisement visits is calculated by the WEB advertisement visit number prediction model. The WEB advertisement visitor number prediction model is a prediction model for calculating the number of people who will visit a store triggered by the WEB advertisement when the WEB advertisement is provided. With the WEB advertisement visitor number prediction model, it is possible to calculate the WEB advertisement forecast visitor number corresponding to each of the candidate (candidate WEB advertisement cost) amount of the cost (WEB advertisement cost) required to provide the WEB advertisement.

The approximate tendency of the number of WEB advertisement visits by the WEB advertisement visit number prediction model is as follows. For example, when the amount of WEB advertising expenses is reduced, the number of people who visit the WEB advertising forecast is small, and when the amount of WEB advertising expenses is increased, the number of people who visit the WEB advertising forecast is large. Even if the amount of WEB advertising expenses exceeds a certain level, the expected number of WEB advertising visitors will be saturated.

Further, the WEB advertisement visit number prediction model can be used to calculate the WEB advertisement visit number prediction number for each of a plurality of stores. Specifically, the WEB advertisement visit number prediction model can be used to calculate the WEB advertisement forecast number of visitors corresponding to each of the candidate WEB advertisement expenses for each of a plurality of stores. Further, the WEB advertisement visit number prediction model can be used to calculate the WEB advertisement visit number corresponding to the number of weeks for each of a plurality of stores.

<< Predictive model accuracy verification unit >>
The predictive model accuracy verification unit verifies the accuracy of the base forecast store visits calculated by the baseline prediction model generation unit, the paper media advertisement forecast visits and the WEB advertisement forecast visits calculated by the advertisement effect prediction model generation unit. .. For verification, it is preferable to use data not used in machine learning. A part of the data used for machine learning and the data used for verification may be duplicated.

<< Predictive model storage >>
The prediction model storage unit stores the prediction model generated by the baseline prediction model generation unit and the prediction model generated by the advertising effect prediction model generation unit. The predictive model storage unit stores various predictive model data necessary for the predictive model, such as the type of the predictive model and the coefficients and constants of the predictive model.

<< Advertising Expenses-Estimated Number of Visitors to the Store >>
Advertising cost-estimated number of customers visiting the store The generation unit generates a correspondence relationship between the amount of candidate (candidate advertising cost) of the cost required to provide the advertisement and the estimated number of people visiting the store corresponding to each amount of the candidate advertising cost. .. Candidate advertising costs are virtual advertising costs for use in predictive models. Candidate advertising costs range from low to high. The range of the amount of candidate advertising expenses may be the range necessary for predicting the number of visitors. Assuming that the advertisement was provided using each amount of the candidate advertisement cost, the number of visitors who will come to the store triggered by the provision of the advertisement is calculated for each amount of the candidate advertisement cost using the prediction model. ..

Candidate advertising expenses include candidate paper media advertising expenses and candidate WEB advertising expenses. Advertising costs include paper advertising costs and WEB advertising costs.

Specifically, assuming that the paper medium advertisement is provided using each amount of the candidate paper medium advertisement cost, the estimated number of paper medium advertisement visits that will come to the store triggered by the provision of the paper medium advertisement is calculated on paper. Media advertising Calculated for each amount of candidate paper media advertising expenses using the store visitor prediction model. For example, as shown in FIG. 8, the advertising cost-estimated number of visitors to the store generation unit responds to the candidate paper media advertising cost for each of a plurality of stores (A store, B point, ..., J store) for the paper medium advertisement. Calculate the estimated number of people visiting the store for paper advertisements corresponding to each amount of the above, and generate the correspondence relationship.

Similarly, assuming that the WEB advertisement is provided using each amount of the candidate WEB advertisement expenses, the number of WEB advertisement visits that will come to the store triggered by the provision of the WEB advertisement is determined by the WEB advertisement visit number prediction model. , Calculated for each amount of candidate WEB advertising expenses. For example, as shown in FIG. 9, the advertising cost-estimated number of customers visiting the store generation unit has each of the candidate WEB advertising costs for each of a plurality of stores (store A, point B, ..., store J) for the WEB advertisement. WEB advertisement forecast number of people visiting the store corresponding to the amount of the above is calculated, and the correspondence relationship is generated and stored.

<< Store-related feature adjustment department >>
The store-related feature adjustment unit adjusts the number of store visitors calculated by the paper-based advertisement visitor number prediction model and the WEB advertisement visitor number prediction model by using the adjustment number of store visits based on the store-related characteristics.

Store-related features include, for example, tenant composition. In recent years, supermarkets may be rented to various tenants as well as their own supermarkets. For example, a drug store, a 100-yen shop, a cleaning shop, or the like may be operating as a tenant. Customers who come to the store may stop by not only supermarkets but also these tenants. Therefore, the number of visitors includes the influence of these tenants.

It is clear that the presence of tenants affects the number of customers visiting a store, but it is often difficult to clearly separate the number of customers visiting a supermarket from the number of customers visiting a store. For this reason, it becomes difficult to include the number of visitors to the tenant in the teacher data for machine learning. Therefore, the number of store visits caused by the existence of tenants is added to the predicted number of store visits calculated by the prediction model as the number of store visits for adjustment.

For example, when a cleaning shop is included in the tenant, the number of people visiting the store for adjustment may be adjusted according to the experience and estimation so far, such as "Distributing 1,000 leaflets will increase the number of visitors by 100". Can be specified in. By appropriately determining the number of visitors for adjustment, the number of visitors can be determined so as to be close to the actual number of visitors. By adding various adjustment visits, the expected number of visitors can be brought closer to reality.

Store-related features include not only tenant composition, but also store types and store environment information. Store types include supermarkets, drug stores, discount stores, home improvement stores, and apparel. The store environment information includes purchase possibility parameters that indicate the possibility that the products handled in the store will be purchased, service availability parameters that indicate the possibility that the services provided by the store will be used, and access to the store. Transportation means parameters related to the method can be used. If these parameters can be included in the teacher data for machine learning, they are included in the teacher data for machine learning. When it is difficult to include these parameters in the teacher data for machine learning, it is processed as the number of visits for adjustment.

<<< Store visit number prediction model generation process >>>
FIG. 10 is a flowchart showing a store visit number prediction model generation process.

First, the processor selects parameters according to the type of industry / store (step S1111). This process is executed by the machine learning parameter selection unit described above.

Next, the processor selects a parameter that does not include past advertising costs (step S1115). In this process, among the parameters selected in the process of step S1111, a parameter that does not include the advertising cost is selected.

Next, the processor machine-learns the parameter data selected in step S1115 as teacher data to generate a baseline prediction model (step S1117). This process is executed by the baseline prediction model generation unit described above.

Next, the processor selects a parameter that includes past advertising costs (step S1119). In this process, among the parameters selected in the process of step S1111, the parameter including the advertising cost is selected.

Next, the processor machine-learns the parameter data selected in step S1119 as teacher data to generate an advertising effect prediction model (step S1121). This process is executed by the above-mentioned advertising effect prediction model generation unit.

Next, the processor verifies the accuracy of the generated prediction model (step S1123). This process is executed by the predictive model accuracy verification unit described above.

Next, the processor stores the generated prediction model (step S1125). This process is executed by the predictive model storage unit described above.

<<< Number of visitors calculation process >>>
FIG. 11 is a flowchart showing the process of calculating the number of visitors to the store.

First, the processor calculates the base predicted number of visitors using the baseline prediction model (step S1211).

Next, the processor calculates the expected number of visitors to the paper advertisement store and the predicted number of visitors to the WEB advertisement using the advertisement effect prediction model (step S1213).

Next, the processor generates a correspondence relationship between the paper advertisement cost and the expected number of visitors to the paper medium advertisement (step S1215). This process is executed by the above-mentioned advertising cost-estimated number of customers visiting the store generation unit.

Next, the processor generates a correspondence relationship between the WEB advertising cost and the expected number of WEB advertising visits (step S1217).

Next, the processor determines whether or not to use the store visitor adjustment parameter (step S1219).

Next, the processor determines the number of customers visiting the store for adjustment (step S1221). This process is executed by the store-related feature adjustment unit described above.

Next, the processor determines the display condition (step S1223). Display conditions can be determined according to the store you want to display and the amount of advertising expenses.

Next, the processor displays the number of visitors according to the display conditions on a display such as a terminal device (step S1225), and ends this subroutine.

<<<<< Utilization of predicted number of visitors >>>>>
By executing the above-mentioned store visitor number calculation process, it is possible to utilize the forecast of the number of visitor number and the amount of advertising cost as follows.

<< First utilization of the store visitor prediction model >>
In one store, a store visitor prediction model can be used to determine the distribution of advertising costs so that the business operator has the desired number of visits.

The sum of the base predicted number of store visits calculated by the baseline prediction model, the number of paper advertisement forecast visits calculated by the paper advertisement store number prediction model, and the WEB advertisement forecast number of store visits calculated by the WEB advertisement visit number prediction model is Select a combination of candidate paper advertising expenses and candidate WEB advertising expenses, which is the desired number of visitors to one store. Specifically, the amount of the paper medium advertising cost and the amount of the WEB advertising cost are determined from the correspondence relationship of the candidate paper medium advertising cost shown in FIG. 8 and the correspondence relationship of the candidate WEB advertising cost shown in FIG. ..

For example, a combination in which the desired number of store visits at store A is the sum of the base forecast number of store visits, the paper media advertisement forecast number of visitors, and the WEB advertisement forecast number of store visits, and the paper media advertisement forecast number of store visits is shown in FIG. The estimated number of visitors is selected from FIG. The figure is omitted for the estimated number of visitors to the store. From the selected combination, the amount of paper advertising expenses and the amount of WEB advertising expenses can be determined. In this way, it is possible to determine the allocation between the paper-based advertising cost and the WEB advertising cost that can realize the desired number of visitors to the store A.

When there are a plurality of combinations of the paper medium advertising cost and the WEB advertising cost, the combination in which the sum of the paper medium advertising cost and the WEB advertising cost is the cheapest is selected. Depending on the selected combination, it is possible to determine the allocation between the paper advertising cost and the WEB advertising cost that can realize the desired number of visitors. In this way, by using the store visit number prediction model, it is possible to determine the optimum allocation between the paper advertising cost and the WEB advertising cost.

<< Second utilization of the store visitor prediction model >>
It can be used to determine the optimal allocation of advertising costs when there is an advertising budget limit for multiple specific stores.

Select the combination of the amount of paper advertising expenses and the amount of WEB advertising expenses, which have the largest total number of visitors in all of the multiple stores. Similar to the first utilization, the amount of paper advertising expenses and the amount of WEB advertising expenses are determined from the correspondence relationship of paper advertising expenses shown in FIG. 8 and the correspondence relationship of WEB advertising expenses shown in FIG. do.

When there are a plurality of combinations, the combination in which the sum of the amount of paper advertising expenses and the amount of WEB advertising expenses is the cheapest is selected. Depending on the combination selected, the allocation of paper advertising expenses and WEB advertising expenses can be determined within the budget. In this way, by using the store visit number prediction model, it is possible to determine the optimum allocation between the paper advertising cost and the WEB advertising cost.

<< Third utilization of the store visitor prediction model >>
It can be used to determine the optimal allocation of advertising costs when there is a budget limit for all stores.

Select the combination of stores so that the total number of visitors is the largest. When there are a plurality of selected combinations, the amount of paper advertising expenses and the amount of WEB advertising expenses, which have the largest number of stores, are selected. The distribution of advertising expenses is determined by the amount of paper advertising expenses and the amount of WEB advertising expenses at each store at that time. In this way, by using the store visit number prediction model, it is possible to determine the optimum allocation between the paper advertising cost and the WEB advertising cost.

<<<<< Scope of embodiment >>>>>
As mentioned above, the first embodiment and the second embodiment have been described, but the descriptions and drawings that form part of this disclosure should not be understood as limiting. Various embodiments not described here are included.

10 Information processing device 20 Information processing device 100 Correlation extraction unit 200 Machine learning processing unit 300 Advertising cost allocation calculation unit 400 Data storage unit

Claims (19)

Information that determines information on advertising costs for providing advertising information on products or services provided at at least one store operated by a business operator by an electronic first advertising medium and a non-electronic second advertising medium. It ’s a processing device,
A teacher is instructed on data including the number of people visiting the store, the correlation demographic data having a correlation with the number of people visiting the store, the advertising cost allocated to the first advertising medium, and the advertising cost allocated to the second advertising medium. A machine learning processing unit that performs machine learning as data,
A predictive model storage unit that stores the predictive model generated by the machine learning,
Advertising cost prediction unit that predicts the first advertising cost to be allocated to the first advertising medium and the second advertising cost to be allocated to the second advertising medium when the number of virtual store visitors reaches a predetermined number from the prediction model. And, an information processing device equipped with. The information processing apparatus according to claim 1, wherein the teacher data further includes at least one of the number, recognition, or popularity of the stores. A billing unit further comprising a correlation demographic data extraction unit that extracts demographic data having a positive correlation with the number of visitors to the store as the correlation demographic data among the demographic data based on the type of business of the store. Item 1. The information processing apparatus according to Item 1. The information processing device according to claim 1, wherein the machine learning processing unit corrects the teacher data and performs machine learning according to a timing factor or a regional factor in which the store is located. The information processing apparatus according to claim 4, wherein the timing factor is a factor determined based on a periodic time unit including a plurality of days. The machine learning processing unit has an accuracy verification unit that verifies the accuracy of the prediction model.
The information processing device according to claim 1, wherein the machine learning processing unit generates the prediction model with high accuracy for all stores of a plurality of stores. The information processing apparatus according to claim 6, indicating whether the correlation demographic data extracted by the correlation demographic data extraction unit is appropriate or inappropriate based on the accuracy of the prediction model of the accuracy verification unit. The information processing device according to claim 1, wherein the first advertising medium is an advertising medium provided via a network. The information processing device according to claim 1, wherein the second advertising medium is a paper medium. The advertising cost prediction unit calculates a virtual number of visitors by giving the first advertising cost allocated to the first advertising medium and the second advertising cost allocated to the second advertising medium to the prediction model. The information processing apparatus according to claim 1. Information that determines information on advertising costs for providing advertising information on products or services provided at at least one store operated by a business operator by an electronic first advertising medium and a non-electronic second advertising medium. It ’s a processing device,
Past information including the number of past visitors to the store, the past advertising expenses required to provide the first advertising medium, and the past advertising expenses required to provide the second advertising medium are used as teacher data. ,
An advertisement-based store visitor number model for calculating the estimated number of advertised-existing store visits that would come to a store when an advertisement is provided using at least one of the first advertising medium and the second advertising medium, and an advertising-based store visitor number model.
By machine learning, an advertisement-free number-of-stores model for calculating the number of unpredicted number of advertisements that will visit a store when an advertisement using both the first advertisement medium and the second advertisement medium is not provided. The machine learning processing unit to generate and
Based on the predicted number of visitors with advertisement and the number of unpredicted visits to the store, the first advertising cost required to provide the first advertising medium and the second advertising cost required to provide the second advertising medium are calculated. An information processing device including an advertising cost determination processing unit for determining. The advertising expense determination processing unit
From the above-mentioned model of the number of people visiting the store with advertisements
The estimated number of visitors to the store with the first advertisement corresponding to each of the plurality of first candidate advertisement expenses that are candidates for the first advertisement expense, and
The estimated number of visitors to the store with the second advertisement corresponding to each of the plurality of second candidate advertisement expenses, which are candidates for the second advertisement expense, is calculated.
The first correspondence relationship in which the estimated number of visitors with the first advertisement is associated with each of the plurality of first candidate advertising expenses, and
The information processing apparatus according to claim 11, wherein a second correspondence relationship in which a second advertising presence predicted number of visitors is associated with each of the plurality of second candidate advertising expenses is generated. The advertising expense determination processing unit
For each of the multiple stores
The information processing apparatus according to claim 12, which generates the first correspondence relationship and the second correspondence relationship. The advertising expense determination processing unit
Based on the desired number of visitors to at least one store, the first expected number of customers with advertisements and the second expected number of customers with advertisements are determined with reference to the first correspondence relationship and the second correspondence relationship.
Based on the determined number of people visiting the store with the first advertisement, one of the plurality of first candidate advertising expenses is selected as the first advertising expense with reference to the first correspondence relationship.
Based on the determined number of people visiting the store with the second advertisement, the second correspondence relationship is referred to, and one of the plurality of second candidate advertising expenses is selected and used as the second advertising expense. The information processing apparatus according to claim 13, which includes a selection process. The machine learning processing unit
The information processing apparatus according to claim 11, further comprising an information selection process for selecting information to be used as the teacher data according to the type of business of the store. The advertising expense determination processing unit
The information processing apparatus according to claim 12, further comprising a predicted number of visitors adjusting process for adjusting the predicted number of visitors with the first advertisement and the predicted number of visitors with the second advertisement based on the characteristic information of the store. A method of determining information on advertising costs for providing advertising information on products or services provided in at least one store operated by a business operator by an electronic first advertising medium and a non-electronic second advertising medium. And
Past information including the number of past visitors to the store, the past advertising expenses required to provide the first advertising medium, and the past advertising expenses required to provide the second advertising medium are used as teacher data. ,
An advertisement-based store visitor number model for calculating the estimated number of advertised-existing store visits that would come to a store when an advertisement is provided using at least one of the first advertising medium and the second advertising medium, and an advertising-based store visitor number model.
By machine learning, an advertisement-free number-of-stores model for calculating the number of unpredicted number of advertisements that will come to the store if the advertisement using both the first advertisement medium and the second advertisement medium is not provided. Machine learning steps to generate and
Based on the predicted number of visitors with advertisement and the number of unpredicted visits to the store, the first advertising cost required to provide the first advertising medium and the second advertising cost required to provide the second advertising medium are calculated. How to include the advertising cost determination step to determine. A program that determines information on advertising costs for providing advertising information on products or services provided at at least one store operated by a business operator by an electronic first advertising medium and a non-electronic second advertising medium. And
Past information including the number of past visitors to the store, the past advertising expenses required to provide the first advertising medium, and the past advertising expenses required to provide the second advertising medium are used as teacher data. ,
An advertisement-based store visitor number model for calculating the estimated number of advertised-existing store visits that would come to a store when an advertisement is provided using at least one of the first advertising medium and the second advertising medium, and an advertising-based store visitor number model.
By machine learning, an advertisement-free number-of-stores model for calculating the number of unpredicted number of advertisements that will come to the store if the advertisement using both the first advertisement medium and the second advertisement medium is not provided. Machine learning steps to generate and
Based on the predicted number of visitors with advertisement and the number of unpredicted visits to the store, the first advertising cost required to provide the first advertising medium and the second advertising cost required to provide the second advertising medium are calculated. A program that includes advertising cost determination steps to determine. A platform for determining advertising costs for providing advertising information on products or services provided in at least one store operated by a business operator by an electronic first advertising medium and a non-electronic second advertising medium. Is a computer implementation method that provides
Past information including the number of past visitors to the store, the past advertising expenses required to provide the first advertising medium, and the past advertising expenses required to provide the second advertising medium are used as teacher data. ,
An advertisement-based store visitor number model for calculating the estimated number of advertised-existing store visits that would come to a store when an advertisement is provided using at least one of the first advertising medium and the second advertising medium, and an advertising-based store visitor number model.
By machine learning, an advertisement-free number-of-stores model for calculating the number of unpredicted number of advertisements that will come to the store if the advertisement using both the first advertisement medium and the second advertisement medium is not provided. The machine learning processing unit to generate and
Based on the predicted number of visitors with advertisement and the number of unpredicted visits to the store, the first advertising cost required to provide the first advertising medium and the second advertising cost required to provide the second advertising medium are calculated. A computer implementation method that provides a platform with an advertising cost determination processing unit to determine.

Images