Information processing device, information processing method, and program

Abstract

To predict a future demand using a model learned on the basis of machine learning in a further suitable manner.SOLUTION: An information processing device includes: calculation means for calculating a hierarchal index on the basis of sales result data of each of first products as targets and a series of products included in categories, which are hierarchically regulated for each of the first products, to which the first products belong among products as hierarchically classified management objects, for each of the categories; and correction means for correcting the sales result data of the first products, on the basis of the hierarchal index calculated for at least the first products, or the categories of any hierarchy to which the first products belong.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present disclosure relates to an information processing apparatus, an information processing method, and a program.

Background Art

[0002] In conducting product sales, inventory management, sales planning, etc., various techniques have been proposed for predicting the future demand for target products by analyzing past sales performance data, external factors, etc. Also, various techniques for using models learned based on so-called machine learning, such as AI (Artificial Intelligence), in predicting product demand have been studied. Patent Document 1 discloses an example of a technique for predicting product demand by using a model constructed based on machine learning with sales performance data as learning data.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] On the other hand, in product sales performance data, for example, when a loss occurs due to a product being out of stock and unable to be sold, the sales performance data does not include the number of sales that could have been made if there had been inventory. Therefore, even if a model is made to perform demand prediction using product sales performance data as in conventional methods, the original demand may not be reflected in the prediction results. Against this background, there is a demand for the realization of a technique that can more accurately predict future demand by taking into account losses and the like exemplified above and using a model learned based on machine learning such as AI.

[0005] In view of the above problems, an object of the present invention is to enable prediction of future demand in a more suitable manner by using a model trained based on machine learning.

Means for Solving the Problems

[0006] The information processing apparatus according to the present invention includes: a calculation means for calculating a hierarchical index based on sales performance data of each of the target first products and each of a series of products included in the category to which the first product belongs among the products to be managed classified step by step, for each of the first products and each of the hierarchically defined categories; and a correction means for correcting the sales performance data of the first product based on the hierarchical index calculated for the first product or for any category of the hierarchy to which the first product belongs. By using the sales performance data corrected based on such a configuration as an input to the model, it becomes possible to predict future demand taking into account the sales volume of products that would have been sold if no loss had occurred.

Effects of the Invention

[0007] According to the present invention, it becomes possible to realize prediction of future demand in a more suitable manner by using a model trained based on machine learning.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Mode for Carrying Out the Invention

[0009] Hereinafter, preferred embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. In this specification and the drawings, components having substantially the same functional configuration are denoted by the same reference numerals, and redundant description is omitted.

[0010] <Overview> First, an overview of the information processing system 1 according to an embodiment of the present disclosure will be described. The information processing system 1 according to this embodiment manages sales records of products as data, and uses a model (for example, AI) that has been learned based on machine learning using the data of past sales records as learning data to predict the future demand for the target product.

[0011] For example, FIG. 1 shows an example of the system configuration of the information processing system 1 according to this embodiment. As shown in FIG. 1, the information processing system 1 according to this embodiment includes a data management server 110, an analysis server 130, and one or more terminal devices 200. In the example shown in FIG. 1, for simplicity of explanation, the case where there is one terminal device 200 is shown, but the number of terminal devices 200 is not particularly limited, and a plurality of them may be provided. The data management server 110, the analysis server 130, and the terminal device 200 are connected so as to be able to transmit and receive data to and from each other via the network N1. Regarding the network N1, its type is not particularly limited as long as it can connect the data management server 110, the analysis server 130, and the terminal device 200 so that they can mutually transmit and receive data. As a specific example, networks such as the Internet, WAN (Wide Area Network), and LAN (Local Area Network) may be applied as the network N1. Also, as the network N1, a network compliant with wireless communication standards such as LTE or 5G may be applied. Further, the network N1 may be realized by a plurality of networks. In this case, the plurality of networks may include two or more networks of different types from each other, and the type of transmission path or the applied communication method in some of the networks may be different from those of other networks. Also, at least a part of the communication between the data management server 110, the analysis server 130, and the terminal device 200 may be mediated by other communication devices.

[0012] Here, referring to FIG. 2, an overview of each device constituting the information processing system 1 will be described. The data management server 110 schematically shows a server that manages various data. The data managed by the data management server 110 may include sales performance data of each product to be managed such as so-called POS (Point of Sale) data, core data such as sales budgets, internal information such as product information and promotional information such as leaflets, and external data such as weather forecast data. At least a part of these data (particularly, the sales performance data of products) can also be used as input data for a model used by the analysis server 130 described later for predicting future demand, learning data when training the model, and the like. Also, although details will be described later, the data management server 110 may calculate (estimate) various losses (for example, opportunity loss, out-of-stock loss, etc.) based on the sales performance data of products, and correct the sales performance data of products based on the calculation results of the losses.

[0013] The analysis server 130 predicts future demand for at least some of the products to be managed, based on various data (especially sales performance data of products) managed by the data management server 110. The analysis server 130 according to the present embodiment uses a model learned based on machine learning, such as so-called AI, for this prediction of future demand. For example, in the example shown in FIG. 2, the analysis server 130 uses a visitor prediction model that predicts the number of visitors to a store in the future (visitor prediction), and a demand prediction model that predicts future demand based on the result of the visitor prediction. The analysis server 130 inputs data (for example, sales performance data) related to the target product acquired from the data management server 110 into the above model, causes the future demand of the product to be predicted, and outputs the result of the demand prediction to a predetermined output destination (for example, it may be output to the terminal device 200 or the like. Also, the analysis server 130 may use the data (especially the sales performance data of products) managed by the data management server 110 as learning data to perform learning of each of the above-exemplified models based on machine learning. For example, in the example shown in FIG. 2, the analysis server 130 predicts the daily sales volume for the next week. Of course, the case shown in FIG. 2 is merely an example, and the time granularity and period related to the prediction of the sales volume are not particularly limited. Also, the data used as input for predicting future demand is not limited to the example shown in FIG. 2. For example, regarding the type of data, depending on the method of data management and the like, the way of aggregation and the name may be appropriately changed, and other data other than the data described in FIG. 2 may be additionally used. Also, the storage location of the input data and the entity that calculates the prediction result are not limited to only the server. For example, a network service such as a so-called cloud service may be applied, or a local environment of a so-called stand-alone terminal device may be applied.

[0014] The terminal device 200 schematically shows a device that serves as an input / output interface for a user to use the data management server 110 and the analysis server 130. For example, the terminal device 200 may be used as an input / output interface for a user to access the data management server 110 to refer to and update the data managed by the data management server 110. As another example, the terminal device 200 may be used as an input / output interface for receiving various instructions from the user regarding future demand prediction by the analysis server 130 and presenting the results of future demand prediction by the analysis server 130 to the user.

[0015] As described above, with reference to FIGS. 1 and 2, the outline of the information processing system 1 according to an embodiment of the present disclosure has been described. Among the information processing system 1 according to the present embodiment, the data management server 110 in particular will be described in detail regarding its functional configuration and processing separately later.

[0016] <Hardware Configuration> With reference to FIG. 3, an example of the hardware configuration of the information processing apparatus 900 applicable to the data management server 110, the analysis server 130, and the terminal device 200 in the information processing system 1 according to the present embodiment shown in FIG. 1 will be described. As shown in FIG. 3, the information processing apparatus 900 according to the present embodiment includes a CPU (Central Processing Unit) 910, a ROM (Read Only Memory) 920, and a RAM (Random Access Memory) 930. The information processing apparatus 900 also includes an auxiliary storage device 940 and a network I / F 970. The information processing apparatus 900 may also include at least one of an output device 950 and an input device 960. The CPU 910, the ROM 920, the RAM 930, the auxiliary storage device 940, the output device 950, the input device 960, and the network I / F 970 are interconnected via a bus 980.

[0017] The CPU 910 is a central processing unit that controls various operations of the information processing apparatus 900. For example, the CPU 910 may control the operations of the entire information processing apparatus 900. The ROM 920 stores control programs, boot programs, etc. that are executable by the CPU 910. The RAM 930 is the main memory of the CPU 910 and is used as a work area or a temporary storage area for expanding various programs.

[0018] The auxiliary storage device 940 stores various data and various programs. The auxiliary storage device 940 is realized by a storage device such as an HDD (Hard Disk Drive) or a non-volatile memory typified by an SSD (Solid State Drive) that can store various data temporarily or persistently.

[0019] The output device 950 is a device that outputs various information and is used to present various information to the user. For example, the output device 950 is realized by a display device such as a display. In this case, the output device 950 presents information to the user by displaying various display information. Also, as another example, the output device 950 may be realized by an acoustic output device that outputs sounds such as voices or electronic sounds. In this case, the output device 950 presents information to the user by outputting sounds such as voices or electronic sounds. Also, the device applied as the output device 950 may be appropriately changed according to the medium used to present information to the user.

[0020] The input device 960 is used to receive various instructions from the user. In this embodiment, the input device 960 includes input devices such as a mouse, a keyboard, and a touch panel. Also, as another example, the input device 960 may include a sound collection device such as a microphone, and may collect the voice spoken by the user. In this case, various analysis processes such as acoustic analysis and natural language processing are performed on the collected voice, so that the content indicated by this voice is recognized as an instruction from the user. Also, the device applied as the input device 960 may be appropriately changed according to the method of recognizing an instruction from the user. Also, a plurality of types of devices may be applied as the input device 960.

[0021] The network I / F 970 is used for communication with an external device via a network. Note that the device applied as the network I / F 970 may be appropriately changed according to the type of communication path and the applied communication method.

[0022] The CPU 910 expands the program stored in the ROM 220 or the auxiliary storage device 940 to the RAM 930 and executes this program, thereby realizing the functional configuration shown in FIG. 4 and the processes described later with reference to FIGS. 6 and 7.

[0023] <Functional configuration> Referring to FIG. 4, an example of the functional configuration of the information processing system 1 according to this embodiment will be described, paying particular attention to the configurations of the data management server 110 and the analysis server 130.

[0024] First, the configuration of the data management server 110 will be described. The data management server 110 includes a storage unit 111, an index calculation unit 112, a loss calculation unit 113, a correction unit 114, and a data output unit 115.

[0025] The storage unit 111 is a storage area used by the data management server 110 to manage various data, and stores various data to be managed. The storage unit 111 stores, for example, data used for future demand prediction such as the sales performance data and product information illustrated in FIG. 2.

[0026] In the information processing system 1 according to this embodiment, a series of products to be managed are classified and managed step by step for each category defined in a hierarchical manner. For example, FIG. 5 is a diagram showing an example of a product classification method, and schematically shows a situation where products to be managed are classified step by step for each category defined in a hierarchical manner. Specifically, FIG. 5 shows an example of the case where the product to be managed, "3-piece beef croquettes", is classified. In the example shown in FIG. 5, "3-piece beef croquettes" is more broadly classified into the category of "hot side dishes" (first layer), and when classified in more detail, it is sequentially classified in the order of the category of "fried foods" (second layer), the category of "croquettes and fries" (third layer), and the category of "meat croquettes" (fourth layer). As described above, in the information processing system 1 according to the present embodiment, the products, which are the minimum units of inventory management (SKU: Stock Keeping Unit), are classified and managed step by step such that the higher the hierarchy, the more general the category corresponding to the classification. In addition, in the present disclosure, for the sake of convenience, when classifying and managing products step by step, the layer corresponding to the topmost category is also referred to as the "first layer", and the larger the numerical value indicating the layer, the lower the layer. That is, in the example shown in FIG. 5, the category of "simmered vegetables" is defined as the category of the "first layer", and the category of "deep-fried foods", which is a further detailed classification of the category of simmered vegetables, is defined as the category of the "second layer". Similarly, the category of "croquettes & fried foods", which is a further detailed classification of the category of deep-fried foods, is defined as the category of the "third layer". In addition, the category of "meat croquettes", which is a further detailed classification of the category of croquettes & fried foods, is defined as the category of the "fourth layer". In addition, in the present embodiment, for the sake of convenience, the product itself as an SKU may also be treated as a category for easier explanation. That is, when the target category is the bottommost category, the category indicates the product itself as an SKU. On the other hand, in the present embodiment, for the sake of convenience, a series of products belonging to the category corresponding to some layers may also be treated as the products to be analyzed. As a specific example, in the example shown in FIG. 5, when a series of products corresponding to the category of "meat croquettes" are taken as the analysis target, the product "meat croquettes" is indicated without particularly distinguishing the series of products belonging to the category of "meat croquettes". In addition, in the information processing system 1 according to the present embodiment, when managing the sales performance of products, in addition to the sales performance of the product itself, the sales performance for each category is also managed (that is, sales performance data is retained). Note that, as described by way of example with respect to FIG. 5, the type of product to be managed by the information processing system 1 according to the present embodiment is not limited. That is, not limited to products classified as prepared foods, any product that can be classified step by step as illustrated in FIG. 5 can be applied as a product to be managed by the information processing system 1 according to the present embodiment. Further, there are no particular limitations on how the categories for each level are defined or the number of levels when products are classified and managed step by step for each category defined in a hierarchical manner.

[0027] The data output unit 115 outputs the data stored in the storage unit 111 to an external device (for example, the analysis server 130). As a specific example, the data output unit 115 may output, as input data to a model that the analysis server 130 uses for future demand prediction, at least data related to products that are targets of demand prediction among a series of data (for example, sales performance data, product information, etc.) stored in the storage unit 111, to the analysis server 130. As another example, the data output unit 115 may output the data (for example, sales performance data, product information, etc.) stored in the storage unit 111 to the analysis server 130 as learning data for the analysis server 130 to perform learning of a model used for future demand prediction.

[0028] The index calculation unit 112, the loss calculation unit 113, and the correction unit 114 calculate chance loss and stockout loss, the details of which will be described later, and correct the sales performance data based on the calculation results of these losses. In the information processing system 1 according to the present embodiment, by applying such correction to the sales performance data, for example, it becomes possible to predict demand taking into account the quantity of products that could have been sold if there had been no out-of-stock of products, or the quantity of products that could have been decreased due to oversupply. Note that hereinafter, in order to distinguish the demand for the number of sales estimated based on the sales performance data to which correction has been applied based on the calculation results of the above losses from the demand for the number of sales estimated based on the sales performance data to which no correction has been applied, it may be referred to as the true number of sales demand. Hereinafter, each of the index calculation unit 112, the loss calculation unit 113, and the correction unit 114 will be described in detail.

[0029] The index calculation unit 112 calculates an index corresponding to each category corresponding to each hierarchy to which the product to be analyzed (hereinafter also referred to as the target product) belongs, according to the sales performance of the products belonging to the category (hereinafter also referred to as the hierarchy-specific index). The hierarchy-specific index to be calculated at least includes a visitor index and a purchase index, the details of which will be described separately later.

[0030] In the present disclosure, the visitor index is an index indicating an indicator of the number of customers who are presumed to have visited for the purpose of purchasing products belonging to the target category. It is calculated based on the ratio between the number of product purchasers for each store and the number of purchasers of the category to which the target product belongs. Specifically, the visitor index is calculated based on the relational expression shown as (Equation 1) below.

[0031]

Equation

[0032] In the above (Equation 1), the number of purchasers of the target category corresponds to the number of purchasers in the category to which the target product belongs. Also, the total number of purchasers corresponds to the number of purchasers for a series of products recorded in the sales performance data (such as POS data) for the target store. For example, if the number of purchasers of a series of products in a certain supermarket is 1000 people and the number of purchasers of products belonging to the fresh food category is 100 people, then the visitor index for products belonging to the fresh food category is 100 people / 1000 people = 0.1. Note that the visitor index calculated based on (Equation 1) corresponds to an example of the "second index". Also, the target product corresponds to an example of the "first product".

[0033] In the present disclosure, the purchase index is an index indicating an indicator of how many products (SKUs) are sold per a predetermined number of people (here, 1000 people for convenience in this embodiment) for a series of products for each store regarding the target SKU (product). Specifically, the purchase index is calculated based on the relational expression shown as (Equation 2) below.

[0034]

Number

[0035] In the above (Formula 2), the number of sales of the target SKU corresponds to the number of units of the target SKU (the product which is the minimum unit of inventory management) sold. For example, assuming that the target SKU belongs to the category of "delicatessen", if the number of purchasers of delicatessen is 500 and the number of sales of the target SKU is 10, then the purchase index for this SKU is 1000 / 500*10 = 20. In addition, when the customer arrival index by time period is used, it is preferable that the customer arrival index is calculated based on the information of the number of purchasers in a certain period such as a specified one hour or morning / afternoon. Also, the purchase index calculated based on (Formula 2) corresponds to an example of the "first index".

[0036] The loss calculation unit 113 calculates various losses that are presumed to have occurred regarding the sales of the product to be analyzed among the past sales records of each product managed as sales performance data. The losses to be calculated at least include the chance loss and the out-of-stock loss, the details of which will be described separately later.

[0037] The chance loss in the present disclosure is a loss corresponding to the number of products that could have been sold if the target product had been available for purchase, but could not be sold due to factors such as out-of-stock. The chance loss is calculated based on information such as the characteristics of the target product (for example, the peak time of the number of sales of the target product), the customer arrival index, the purchase index, and the final sales time of the target product. Regarding the opportunity loss, the calculation method differs between products for which sufficient sales performance data has been accumulated for analysis like existing products, and products for which insufficient sales performance data has been accumulated to the extent that analysis is possible like new products (and thus, products for which no sales performance data has been accumulated). Therefore, the calculation method of opportunity loss will be explained separately for the case of existing products (when sufficient sales performance data has been accumulated and the data volume is equal to or greater than the threshold) and the case of new products (when insufficient sales performance data has been accumulated and the data volume is less than the threshold).

[0038] First, an example of the calculation method of opportunity loss in this embodiment will be described by focusing on the case where an existing product is the target product. The opportunity loss when an existing product is the target product is calculated based on, for example, the relational expression shown as (Equation 3) below.

[0039]

Equation

[0040] As a specific example, when a certain product A runs out of stock at 15:00 before the afternoon peak time, the number of product A that could have been sold by the afternoon peak time is calculated as the opportunity loss based on the above (Equation 3). In this embodiment, the opportunity loss is calculated for the period from the final sales time of the target product to the peak time of the sales volume. However, regarding the period for which the opportunity loss is calculated, as long as it is a time period after the final sales time, it may be appropriately changed according to the characteristics of the product. As a specific example, the opportunity loss may be calculated for the period from the final sales time of the target product (target SKU) to the closing time. In this case, for example, when a certain product A runs out of stock before closing, the number of product A that could have been sold by the closing time is calculated as the opportunity loss. Also, the chance loss calculated based on (Equation 3) corresponds to an example of the "first loss". That is, the chance loss calculated based on (Equation 3) is calculated based on a hierarchical index (purchase index) calculated based on sales performance data of a series of products including at least the first product (target product). Note that (Equation 3) is a relational expression for enabling the calculation of the chance loss even when no special additional data is acquired for calculating the chance loss, and the method is not particularly limited as long as information corresponding to the chance loss can be acquired. As a specific example, if additional data such as the time not placed on the store shelf can be used, the chance loss may be acquired using the data. Also, when the purchase index by time zone is used, it is preferable that the purchase index is calculated based on information on the number of purchasers and the number of sales when divided into a certain period such as a specified one hour or morning / afternoon.

[0041] Next, an example of the method for calculating the chance loss in the present embodiment will be described by focusing on the case where a new product is the target product. The chance loss when a new product is the target product is calculated based on the relational expression shown below as (Equation 4).

[0042]

Number

[0043] Note that the category to which the target SKU in (Equation 4) belongs is not necessarily limited to the category corresponding to the layer immediately above the target SKU, and any category corresponding to a layer higher than the target SKU can be applied as long as it is a category corresponding to a higher layer than the target SKU. For example, as in the example shown in FIG. 5, when "3-piece deep-fried wagyu croquettes" is the target product, as the customer index and the purchase index, an index calculated for any of the meat croquettes corresponding to the fourth layer category, the croquettes / fried foods corresponding to the third layer category, and the fried foods corresponding to the second layer category may be applied. Also, when calculating the opportunity loss by time zone, it is advisable to calculate using the prepared visitor index or purchase index by time zone. In calculating the opportunity loss, various indexes with a shorter span than the period to be calculated may be accumulated for calculation. For example, when calculating the opportunity loss by day, the result of calculating and summing up the opportunity losses by time zone may be treated as the opportunity loss by day. It is advisable to take an approach of determining the value to be actually used after calculating the index and loss with the smallest granularity from the accuracy of the existing data. Also, the opportunity loss calculated based on (Equation 4) corresponds to an example of the "second loss". Also, among a series of products belonging to the category corresponding to any hierarchy to which the target product (target SKU) is assigned, other products other than the target product correspond to an example of the "second product". That is, the opportunity loss calculated based on (Equation 4) is calculated based on the hierarchical indexes (visitor index and purchase index) calculated based on the sales performance data of a series of products including at least one of the first product (target product) and the second product.

[0044] The stockout loss in the present disclosure is a loss corresponding to the number of products that are presumed to have been able to be decreased originally, based on the number of products sold at a price discounted as a result of factors such as oversupply of products. The stockout loss is calculated as the number of products for which selling itself is substantially in the red (in other words, the number of products that may be decreased) based on information such as the stockout rate (in other words, the price discount rate) and the cost rate. The stockout loss is calculated based on the relational expression shown as (Equation 5) below.

[0045]

Number

[0046] That is, assuming that a certain product A was sold 10 pieces at 50% off, the lost sales would be 10×(1 - 0.5) = 5. Therefore, for example, even if the number of product A is reduced by 4, which is less than the lost sales of 5, there is a high possibility that there will be no particular problem in inventory management. Rather, it is expected that the effect of further reducing the deficit will be achieved. When calculating the lost sales by time period, it is advisable to calculate the lost sales based on the sales volume divided into a certain period such as a specified 1 hour or morning / afternoon. Also, when calculating the lost sales, the lost sales with a shorter span than the period to be calculated may be accumulated for the calculation. Also, the lost sales calculated based on (Formula 5) corresponds to an example of the "third loss".

[0047] As described above, the loss calculation unit 113 calculates the opportunity loss and lost sales exemplified above for at least the products (for example, SKUs) that are the target of future demand prediction among a series of products managed by the data management server 110.

[0048] The correction unit 114 corrects the sales performance data of the product that is the target of the loss calculation among the sales performance data stored in the storage unit 111 based on the loss (for example, opportunity loss or lost sales) calculated by the loss calculation unit 113. As a specific example, the correction unit 114 may correct the sales performance data by adding the opportunity loss calculated for the target product to the sales performance data of the target product. By making such a correction, it becomes possible to reflect, for example, the number of products that could have been sold if inventory had been secured in the sales performance data of the target product. Also, the correction unit 114 may correct the sales performance data by subtracting the lost sales calculated for the target product from the sales performance data of the target product. By making such a correction, it becomes possible to reflect, for example, the number of products that would presumably not have to be sold at a discounted price due to leftovers, etc. in the sales performance data of the target product.

[0049] In addition, when correcting the sales performance data based on the various losses described above, the correction unit 114 may determine whether or not to correct the sales performance data for the target product according to the product characteristics of the target product. As a specific example, for products with a certain sales volume, in demand forecasting and inventory management of the products, the same quantity is handled as a matter of habit, and a situation where such handling is correct can also be assumed. Therefore, the correction unit 114 may exclude products with a certain sales volume from the target of correcting the sales performance data. In addition, for products with a large variation in sales volume (for example, products with a large dispersion in daily sales volume), situations where purchases are not made due to operational errors, or situations where the order of supply or cooking order, etc. cannot be observed due to a shortage of manpower, etc., resulting in missed sales opportunities, can be presumed. Therefore, the correction unit 114 may include products with a large variation in sales volume in the target of correcting the sales performance data. In addition, the correction unit 114 may determine whether or not to correct the sales performance data for the target product according to the trend tendency of the target product in a predetermined period unit. As a specific example, when looking at the monthly trend tendency, the correction unit 114 may include products with a large difference in the tendency of the occupancy rate of sales performance in indicators such as daily, hierarchical, and SKU-by-SKU (for example, products with large changes) in the target of correcting the sales performance data. In this case, for example, products with a large spike in occupancy rate or products with a large decrease in occupancy rate in some periods may be included in the target of correcting the sales performance data.

[0050] In contrast, for products affected by external factors such as order restrictions, the correction unit 114 may exclude them from the target of correcting the sales performance data. In addition, for products sold based on prior reservations, since the number of products assumed from the beginning will be sold at the expected price, there is a high possibility that there is no situation where opportunity loss or shortage loss may occur. Therefore, the correction unit 114 may exclude products sold based on prior reservations from the target of correcting the sales performance data. Further, the correction unit 114 may exclude the product from the target for correcting the sales performance data according to the peak time of the sales quantity of the product. As a specific example, for a product whose peak time of sales quantity is in the time period immediately after opening, it is assumed that the store wants to sell out quickly. In such a case, even if the sales quantity of the product does not increase in the subsequent time period, it does not necessarily correspond to a missed opportunity. Therefore, when such a situation can be assumed, the correction unit 114 may exclude the corresponding product from the target for correcting the sales performance data. Also, for products for which the inventory can be stably and sufficiently secured, such as products that can be stored for a long time (for example, frozen products), the possibility of a missed opportunity due to inventory shortage or the like is low. Therefore, the correction unit 114 may exclude the products for which the inventory can be stably and sufficiently secured as described above from the target for correcting the sales performance data. Also, for products that are renewed and sold from evening to night without a plan, they may be excluded from the correction target. As a specific example, for products that are remade from the products arranged at the storefront immediately after opening and then sold again, only the data of the products before renewal may be sufficient at the planning stage, and the products after renewal may be excluded from the correction target.

[0051] Further, when correcting the sales performance data based on the various losses described above, the correction unit 114 may adjust the correction value based on predetermined conditions. As a specific example, from a statistical perspective, assuming that the variation in the sales quantity follows a normal distribution, 95% of the data will be included in the so-called ±2SD range (the range corresponding to twice the value of the standard deviation). The correction unit 114 may utilize such characteristics to adjust the correction value applied to the correction of the sales performance data. Specifically, the correction unit 114 may sum up the correction values based on the various losses for each SKU, and then determine whether the corrected sales performance data falls within the ±2SD range of the past spousal sales volume, minimum sales volume, etc. Moreover, when it is determined that the corrected sales performance data does not fall within the ±2SD range of the past maximum sales volume, minimum sales volume, etc., the correction unit 114 may adjust the correction value so that the corrected sales performance data falls within the ±2SD range.

[0052] Also, when correcting the sales performance data based on the above losses, the correction unit 114 may store the pre-correction sales performance data and the post-correction sales performance data separately in the storage unit 111 without overwriting the pre-correction sales performance data. By applying such control, it becomes possible to use the pre-correction sales performance data and the post-correction sales performance data appropriately according to the situation at that time.

[0053] In the above description, an example has been described in which future demand prediction is mainly performed for SKUs as target products, and opportunity loss and stockout loss are calculated, and sales performance data is corrected based on these losses. On the other hand, the target set as the target product is not necessarily limited to SKUs. For example, a situation can be envisioned where future demand prediction is performed after a series of products belonging to a desired hierarchical category are set as target products. As a specific example, in the case of the example shown in FIG. 5, "3-piece deep-fried beef croquettes" corresponds to an SKU, but a situation where future demand prediction is performed for a series of products classified as meat croquettes corresponding to the fourth-level category may apply. In such a case, after setting a series of products belonging to "meat croquettes" as target products, opportunity loss and stockout loss can be calculated for the series of products belonging to the meat croquettes, and sales performance data can be corrected based on these losses. Also, for the category in which "meat croquettes" is classified in this case, the categories "fried foods", "hot side dishes", and "croquettes & fries" corresponding to higher levels than the meat croquettes are applicable.

[0054] Next, the configuration of the analysis server 130 will be described. The analysis server 130 includes a storage unit 131, a prediction processing unit 132, and a learning processing unit 133.

[0055] The storage unit 131 is a storage area used by the analysis server 130 for managing various data, and stores various data to be managed. The storage unit 111 may store data of models that are learned based on machine learning and used for various predictions (e.g., future demand prediction, etc.), such as the customer arrival prediction model and the demand prediction model described with reference to FIG. 2.

[0056] The prediction processing unit 132 performs various predictions using the data managed by the data management server 110. As a specific example, the prediction processing unit 132 may acquire data related to the target product, such as sales performance data of the target product, from the data management server 110, and perform a future demand prediction of the target product based on the data. In this case, for example, the prediction processing unit 132 may recognize the target product to be the subject of future demand prediction based on an instruction from a user via the terminal device 200, and acquire the sales performance data of the product from the data management server 110. Then, the prediction processing unit 132 may input the data acquired from the data management server 110 into the above-described customer visit prediction model or demand prediction model to perform a future demand prediction of the target product. At this time, the prediction processing unit 132 may perform a future demand prediction of the target product using, as an input, the sales performance data corrected by the above-described index calculation unit 112, loss calculation unit 113, and correction unit 114. By applying such control, for example, it becomes possible to predict the future demand (i.e., the true sales volume demand) of the target product on the premise of data taking into account the above-described opportunity loss and stockout loss.

[0057] Further, the prediction processing unit 132 may output information corresponding to the result of the future demand prediction of the target product to a predetermined output destination. As a specific example, the prediction processing unit 132 may transmit information corresponding to the result of the future demand prediction of the target product to the terminal device 200. Thereby, it becomes possible to present information corresponding to the result of the future demand prediction of the target product to the user via the terminal device 200. Also, as another example, the prediction processing unit 132 may output information corresponding to the result of the future demand prediction of the target product to an external device that performs other analysis (for example, a server that formulates a sales plan, etc.). By applying such control, it is also possible to realize a mechanism for formulating a sales plan that achieves a sales budget, takes into account various constraints, and maximizes gross profit, using the result of the future demand prediction of the target product taking into account the true sales volume demand as an additional input. As another example, the prediction processing unit 132 may feedback information according to the result of future demand prediction of the target product to the data management server 110. By applying such a configuration, the data management server 110 can take into account the result of future demand prediction by the prediction processing unit 132 when correcting the sales performance data. Note that an example of the process related to the correction of the sales performance data taking into account the result of future demand prediction will be described in detail separately as a modification example.

[0058] The learning processing unit 133 performs learning of a model that the prediction processing unit 132 uses for various predictions, using the data managed by the data management server 110 as learning data based on so-called machine learning. As a specific example, the learning processing unit 133 may perform learning of a customer visit prediction model or a demand prediction model by using various data (for example, product sales performance data, product information, sales budget, etc.) managed by the data management server 110 as learning data. At this time, the learning processing unit 133 may apply the sales performance data corrected by the above-described index calculation unit 112, loss calculation unit 113, and correction unit 114 as learning data. By applying such control, it becomes possible to cause the customer visit prediction model and the demand prediction model to learn so that, for example, prediction of future demand (that is, true sales volume demand) taking into account the above-described opportunity loss and shortage loss is performed.

[0059] Note that the configuration shown in FIG. 3 is merely an example, and the functional configuration of the information processing system 1 is not necessarily limited to the example shown in FIG. 1 as long as the functions of the above-described respective components are realized. For example, at least one of the data management server 110 and the analysis server 130 may be realized by cooperation of a plurality of devices. As a specific example, some of the functions of a series of components of the data management server 110 may be realized by another external device different from the data management server 110. As another example, at least a part of the processing load of a series of components of the components of the data management server 110 may be distributed to a plurality of devices. The same applies to the analysis server 130. As another example, the data management server 110 and the analysis server 130 may be integrally configured. As another example, at least one of the data management server 110 and the analysis server 130 may be realized as a so-called network service typified by cloud services.

[0060] As described above, with reference to FIG. 3, an example of the functional configuration of the information processing system 1 according to the present embodiment has been described, particularly focusing on the configurations of the data management server 110 and the analysis server 130.

[0061] <Processing> An example of the processing of the information processing system 1 according to the present embodiment will be described, particularly focusing on the processing of the data management server 110 and the analysis server 130. For example, FIG. 6 is a flowchart showing an example of the flow of a series of processes of the information processing system 1. FIG. 6 shows an example of the processing when the analysis server 130 performs future demand prediction on at least a part of a series of products managed by the data management server 110 as target products.

[0062] In S110, the data management server 110 executes various preprocessings on a series of data to be managed. As a specific example, the data management server 110 collects data that can be used for future demand prediction, such as sales performance data and product information for each of a series of products to be managed, and combines various collected data as necessary. In addition, the data management server 110 performs preprocessings such as so-called cleaning on the collected series of data (particularly, preprocessings for executing subsequent processes shown as S120 to S130).

[0063] In S120, the data management server 110 executes processing related to estimating the true sales volume demand for at least target products among a series of products to be managed, which are targets for future demand prediction. Specifically, the data management server 110 estimates losses (such as opportunity losses and stockout losses) that would occur based on the sales performance data of the products to be managed, and corrects the sales performance data of the target products based on the estimation results of the losses so that the true sales volume demand is reflected. Then, the data management server 110 outputs data to be input to the model for product demand prediction, including at least the corrected sales performance data (i.e., the sales performance data reflecting the estimation results of the true sales volume demand), to the analysis server 130.

[0064] Here, referring to FIG. 7, a specific example of the processing shown as S120 in FIG. 6 will be described in detail. In S121, the data management server 110 calculates a hierarchical index corresponding to each category corresponding to each hierarchy to which the target product belongs according to the sales performance of the products belonging to the category, according to the target product to be analyzed. The hierarchical index calculated in S121 includes at least the customer arrival index described above with reference to (Equation 1) and the purchase index described above with reference to (Equation 2).

[0065] In S122, when calculating the correction value of the sales performance data of the target product, the data management server 110, as a preprocessing step, acquires information used for analyzing the sales performance of the target product. As a specific example, the data management server 110 identifies the peak time of the sales volume of the target product. At this time, if information indicating the peak time of the sales volume of the target product is included in the product information of the target product, the information included in the product information may be used. As another example, the data management server 110 may identify the peak time of sales volume of target products by classifying a series of products to be managed for each peak time of sales volume using a model trained based on so-called machine learning. In this case, for example, information such as GTIN code, average purchase time, average purchase time (first purchase by day), average purchase time (last purchase by day), average purchase time (by day) average purchase quantity, average purchase quantity by time period (every hour), etc. may be used as inputs to the model. Furthermore, the model may be constructed so that it classifies each of the series of products to be managed for each peak time zone of sales volume (for example, morning, afternoon, evening, late night, etc.).

[0066] In S123, the data management server 110 calculates various losses based on the hierarchical index calculated in S121 and the result of the preprocessing in S122. The various losses calculated in S123 include the chance loss described above with reference to (Equation 3) and (Equation 4), and the shortage loss described above with reference to (Equation 5). Note that the loss calculated in S123 corresponds to the correction value applied to the sales performance data. In S124, the data management server 110 calculates the correction value to be applied to the sales performance data by summing up the correction values calculated as various losses in S122. Also, at this time, the data management server 110 may adjust the correction value to be applied to the sales performance data so that the corrected sales performance data satisfies a predetermined condition. In S125, the data management server 110 corrects the sales performance data of the target product based on the correction values calculated in S123 and S124. Furthermore, the data management server 110 outputs, to the analysis server 130, data to be input in future demand prediction that includes at least the corrected sales performance data (that is, the sales performance data reflecting the estimated result of the true sales volume demand).

[0067] As described above, with reference to FIG. 7, an example of the process shown as S120 in FIG. 6 has been described in detail.

[0068] Here, refer to FIG. 6 again. In S130, the analysis server 130 uses the input data related to the future demand prediction (for example, sales performance data of target products, product information, etc.) obtained from the data management server 110 as the input to the customer arrival prediction model, so that the customer arrival prediction model predicts the number of customers visiting the target store during the future target period (for example, next week, etc.). In S140, the analysis server 130 inputs the prediction result of the customer arrival prediction model in S130 and the input data related to the future demand prediction obtained from the data management server 110 into the demand prediction model, so that the demand prediction model predicts the demand for the target product during the future target period (for example, next week, etc.). Moreover, the analysis server 130 outputs information corresponding to the result of the future demand prediction of the target product to a predetermined output destination. As a specific example, the analysis server 130 may present information corresponding to the result of the demand prediction to the user via the terminal device 200. As another example, the analysis server 130 may output information corresponding to the result of the demand prediction to an external device that performs other analyses (for example, a server that formulates a sales plan, etc.).

[0069] As described above, with reference to FIGS. 6 and 7, an example of the processing of the information processing system 1 according to the embodiment has been described, particularly focusing on the processing of the data management server 110 and the analysis server 130.

[0070] <Modification Example> A modification example of the information processing system 1 according to the present embodiment will be described below. As described above, the analysis server 130 may feedback information corresponding to the result of the future demand prediction of the target product to the data management server 110. By applying such a mechanism, the data management server 110 can also execute processes related to the estimation of the true sales volume demand (for example, processes related to the correction of the sales performance data of the target product) by taking into account the feedback from the analysis server 130.

[0071] For example, when the error rate (in other words, the degree of deviation) between the demand prediction result of the target product fed back from the analysis server 130 and the true sales volume demand estimated for the target product exceeds the threshold value, the data management server 110 may control the method for estimating the true sales volume demand of the target product. As a specific example, for completely new products that are not renewals, events may occur where the number of samples of the sales performance of products belonging to the same category is small and the value as statistical summary data is low, or the sub-classification of the category is incorrect. Therefore, for example, when the above error rate exceeds the threshold value, the data management server 110 may change the category level referred to when calculating the hierarchical index (for example, the customer arrival index or the purchase index) to a higher level. Specifically, in the example shown in FIG. 5, it is assumed that initially, the calculation of various losses and the correction of sales performance data based on the loss were performed using the hierarchical index targeting croquettes·fries corresponding to the third-level category. Under such a premise, for example, it is assumed that the error rate between the demand prediction result of the target product by the analysis server 130 and the true sales volume demand of the target product by the data management server 110 exceeds the threshold value. In this case, the data management server 110 may, for example, change the hierarchical index referred to when calculating various losses and correcting the sales performance data based on the loss to the hierarchical index targeting fried foods corresponding to the second-level category, which is a higher level than the croquettes·fries corresponding to the third-level category. Also, the threshold value related to the determination of the above error rate can be appropriately set according to the use case of the information processing system 1 according to this embodiment. As a specific example, the above threshold value may be set based on the average of the error rates between the future demand prediction result of each of a series of products to be managed and the estimated result of the true sales volume demand for the product. In addition, regarding the control of the category hierarchy referred to when calculating the hierarchical index based on the above error rate, it may be periodically executed at predetermined intervals. As a specific example, an error rate is determined monthly between the result of predicting the future demand for a product and the estimated result of the true demand for the number of sales of the product, and control is performed on the hierarchy of the category referred to when calculating the hierarchical index according to the result of the determination. Also, even when the hierarchy of the category referred to when calculating the hierarchical index is changed to a higher hierarchy, if an improvement in the above error rate is expected, control may be applied to return to a lower hierarchy.

[0072] <Conclusion> As described above, in the information processing system 1 according to the present embodiment, the data management server 110 classifies and manages a series of products to be managed step by step for each category defined in a hierarchical manner. Further, the data management server 110 calculates a hierarchical index corresponding to each hierarchy according to the sales performance of the products belonging to the category, based on the sales performance data of each of a series of products including at least either the target first product or another second product having a common category hierarchy with the first product. Then, the data management server 110 corrects the sales performance data of the first product based on the hierarchical index calculated for at least one of the hierarchies to which the first product or the second product belongs. With the above configuration, it becomes possible to estimate various losses that are difficult to interpret simply from the data alone, such as the number of products that could have been sold if there had been no stock shortage, and the number of sales that would not have had to be sold at a fire-sale price if there had been no leftovers. Thus, according to the information processing system 1 according to the present embodiment, it is possible to further improve the quality of the data accumulated in the past, such as the sales performance data used for analysis such as demand prediction, and transform the data into data that enables more accurate analysis in line with the actual situation. Thereby, for example, it becomes possible to predict the future demand (true demand for the number of sales) of the target product in line with the actual situation, taking into account the losses that would have occurred in the past as described above.

[0073] Note that the above-described embodiments are merely examples and do not necessarily limit the configuration or processing of the present invention. Various modifications and changes may be made without departing from the technical idea of the present invention. In addition, the present invention includes a program for realizing the functions of the above-described embodiments and a computer-readable recording medium storing the program.

[0074] Also, the following configurations also belong to the technical scope of the present disclosure. (1) An information processing apparatus comprising: a calculation means for calculating a hierarchical index based on sales performance data of the first product and each of a series of products included in the category to which the first product belongs among the products to be managed classified step by step, for each of the target first products and for each of the hierarchically defined categories; and a correction means for correcting the sales performance data of the first product based on the hierarchical index calculated for the first product or for any category of the hierarchy to which the first product belongs. (2) The information processing apparatus according to (1), wherein the hierarchical index includes a first index corresponding to a ratio of the number of purchasers of the first product belonging to the category to the number of purchasers of a series of products belonging to the category for which the hierarchical index is calculated. (3) The information processing apparatus according to (2), wherein the hierarchical index includes a second index corresponding to a ratio of the number of purchasers of a series of products belonging to the category for which the hierarchical index is calculated to the number of purchasers of a series of products to be managed. (4) The information processing apparatus according to (2), wherein the correction means corrects the sales performance data of the first product by adding a correction value corresponding to a first loss calculated based on the number of purchasers of the first product in a predetermined time period indicated by the sales performance data of the first product and the first index calculated for any category of the hierarchy to which the first product belongs, to the sales performance of the first product. (5) If the data volume of the sales performance data of the first product is equal to or greater than a threshold value, the correction means corrects the sales performance data of the first product based on a correction value corresponding to the first loss, for the information processing apparatus according to (4). (6) The correction means adds a correction value corresponding to a second loss calculated based on the number of visitors indicated by the sales performance data of a series of products belonging to any category to which the first product belongs, the first index calculated for the category, and the second index, to the sales performance of the first product, thereby correcting the sales performance data of the first product, for the information processing apparatus according to (3). (7) If the data volume of the sales performance data of the first product is less than a threshold value, the correction means corrects the sales performance data of the first product based on a correction value corresponding to the second loss, for the information processing apparatus according to (6). (8) The correction means subtracts a correction value corresponding to a third loss calculated based on the sell - through rate related to the discount of the price of the first product and the number of sales of the first product at a price corresponding to the sell - through rate, from the sales performance of the first product, thereby correcting the sales performance data of the first product, for the information processing apparatus according to any one of (1) to (7). (9) The correction means determines whether or not to target the first product for correction of the sales performance data according to the product characteristics of the first product, for the information processing apparatus according to any one of (1) to (8). (10) The correction means targets the first product for correction of the sales performance data when the first product meets at least one of the following conditions: when the number of sales is constant, when the variation in the number of sales exceeds a threshold value, when the sales peak time is after noon, and when the change in the trend of the occupancy rate of the sales performance under a predetermined condition exceeds a threshold value, for the information processing apparatus according to (9). (11) If the first product meets at least one of the following conditions: when there were restrictions on previous orders, when it is a reserved product, when the sales peak time is in the morning, and when it is a product that can be stored long-term, the correction means excludes the first product from the targets for correcting the sales performance data, the information processing apparatus according to (9) or (10). (12) If the corrected sales performance data of the first product is not within the range of twice the standard deviation of the maximum or minimum past sales volume of the first product, the correction means controls the correction value for correcting the sales performance data so that the sales performance data is within the range, the information processing apparatus according to any one of (1) to (11). (13) Prediction means for predicting the demand for a target product using past sales performance data as input and outputting the result of the prediction, and inputting the sales performance data of the first product corrected by the correction means to the prediction model, so that the prediction model predicts the demand for the first product. When the error rate between the prediction result of the demand for the first product by the prediction means and the corrected sales performance data of the first product exceeds a threshold value, the hierarchical index applied to the correction of the sales performance data of the first product hereafter is changed to the hierarchical index corresponding to a category in a higher hierarchy than when the sales performance data was corrected for the prediction of the demand for the first product, the information processing apparatus according to any one of (1) to (12). (14) When the error rate between the prediction result of the demand for the first product by the prediction means and the corrected sales performance data of the first product improves from a state where it exceeds the threshold value to a state where it is below the threshold value, the hierarchical index applied to the correction of the sales performance data of the first product hereafter is changed from the hierarchical index corresponding to the category in the higher hierarchy that was previously changed to at least the hierarchical index corresponding to a category in a hierarchy lower than the higher hierarchy, the information processing apparatus according to (13). (15) An information processing method executed by an information processing apparatus, comprising: a calculation step of calculating a hierarchical index based on sales performance data of each target first product and each hierarchically defined category, for each of the first product and each series of products included in the category to which the first product belongs among the products to be managed classified step by step; and a correction step of correcting the sales performance data of the first product based on the hierarchical index calculated for the first product or any category of the hierarchy to which the first product belongs. (16) A program for causing a computer to execute: a calculation step of calculating a hierarchical index based on sales performance data of each target first product and each hierarchically defined category, for each of the first product and each series of products included in the category to which the first product belongs among the products to be managed classified step by step; and a correction step of correcting the sales performance data of the first product based on the hierarchical index calculated for the first product or any category of the hierarchy to which the first product belongs.

Explanation of Signs

[0075] 1 Information processing system 110 Data management server 111 Storage unit 112 Index calculation unit 113 Loss calculation unit 114 Correction unit 115 Data output unit 130 Analysis server 131 Storage unit 132 Prediction processing unit 133 Learning processing unit 200 Terminal device

Claims

1. Calculation means for calculating a hierarchical index based on the sales performance data of the first product and each of a series of products included in the category to which the first product belongs among the products to be managed classified step by step, for each of the first products to be targeted and for each of the hierarchically defined categories; Correction means for correcting the sales performance data of the first product based on the hierarchical index calculated for the first product or for any category of the hierarchy to which the first product belongs; An information processing apparatus comprising the above.

2. The hierarchical index includes a first index corresponding to the ratio of the number of purchasers of the first product belonging to the category to the number of purchasers of a series of products belonging to the category for which the hierarchical index is calculated, for the information processing apparatus according to claim 1.

3. The hierarchical index includes a second index corresponding to the ratio of the number of purchasers of a series of products belonging to the category for which the hierarchical index is calculated to the number of purchasers of a series of products to be managed, for the information processing apparatus according to claim 2.

4. The correction means corrects the sales performance data of the first product by adding a correction value corresponding to a first loss calculated based on the number of purchasers of the first product in a predetermined time period indicated by the sales performance data of the first product and the first index calculated for any category of the hierarchy to which the first product belongs, to the sales performance of the first product, for the information processing apparatus according to claim 2.

5. The correction means corrects the sales performance data of the first product based on the correction value corresponding to the first loss when the data volume of the sales performance data of the first product is equal to or greater than a threshold value, for the information processing apparatus according to claim 4.

6. The correction means corrects the sales performance data of the first product by adding a correction value corresponding to a second loss calculated based on the number of visitors indicated by the sales performance data of a series of products belonging to any category of the hierarchy to which the first product belongs, and the first index and the second index calculated for the category, to the sales performance of the first product, for the information processing apparatus according to claim 3.

7. The correction means corrects the sales performance data of the first product based on a correction value corresponding to the second loss when the data volume of the sales performance data of the first product is less than a threshold value. The information processing apparatus according to claim 6.

8. The correction means subtracts, from the sales performance of the first product, a correction value corresponding to a third loss calculated based on a markdown rate related to a discount on the price of the first product and the number of sales of the first product at a price corresponding to the markdown rate, thereby correcting the sales performance data of the first product. The information processing apparatus according to claim 1.

9. The correction means determines whether or not to target the first product for correction of the sales performance data according to the product characteristics of the first product. The information processing apparatus according to claim 1.

10. The correction means, when the first product has a constant number of sales, when the variation in the number of sales exceeds a threshold value, when the sales peak time is after noon, and when the change in the trend of the occupancy rate of sales performance under predetermined conditions exceeds a threshold value, targets the first product for correction of the sales performance data when at least one of the above applies. The information processing apparatus according to claim 9.

11. The correction means, when the first product has been subject to ordering restrictions in the past, is a reserved product, when the sales peak time is in the morning, and when it is a product that can be stored for a long time, excludes the first product from the target for correction of the sales performance data when at least one of the above applies. The information processing apparatus according to claim 9.

12. When the corrected sales performance data of the first product is not included in the range of twice the standard deviation of the past maximum or minimum sales volume of the first product, the correction means controls the correction value related to the correction of the sales performance data so that the sales performance data is included in the range. The information processing apparatus according to claim 1.

13. Predicting means for predicting the demand for a target product using past sales performance data as an input and outputting the result of the prediction, and inputting the sales performance data of the first product corrected by the correction means to the prediction model, thereby causing the prediction model to predict the demand for the first product. When the error rate between the prediction result of the demand for the first product by the prediction means and the corrected sales performance data of the first product exceeds a threshold value, the hierarchical index applied to the correction of the sales performance data of the first product hereafter is changed to the hierarchical index corresponding to a category in a higher hierarchy than when the sales performance data used for predicting the demand for the first product was corrected. The information processing apparatus according to claim 1.

14. When the error rate between the prediction result of the demand for the first product by the prediction means and the corrected sales performance data of the first product improves from a state exceeding the threshold value to a state equal to or lower than the threshold value, the hierarchical index applied to the correction of the sales performance data of the first product hereafter is changed from the hierarchical index corresponding to the category in the higher hierarchy that has been changed previously to at least the hierarchical index corresponding to a category in a hierarchy lower than the higher hierarchy. The information processing apparatus according to claim 13.

15. An information processing method executed by an information processing apparatus, a calculation step of calculating a hierarchical index based on the sales performance data of the first product and each of a series of products included in the category to which the first product belongs among the products to be managed classified step by step, for each first product as an object and for each category defined hierarchically; a correction step of correcting the sales performance data of the first product based on the hierarchical index calculated for the first product or for any category in the hierarchy to which the first product belongs; An information processing method including the above.

16. Causing a computer to calculate a hierarchical index based on the sales performance data of the first product and each of a series of products included in the category to which the first product belongs among the products to be managed classified step by step, for each first product as an object and for each category defined hierarchically; correct the sales performance data of the first product based on the hierarchical index calculated for the first product or for any category in the hierarchy to which the first product belongs; A program for executing the above.

Images