WO2023140477A1

WO2023140477A1 - Device and method for predicting order quantity of store material

Info

Publication number: WO2023140477A1
Application number: PCT/KR2022/018097
Authority: WO
Inventors: 나준채; 신원용; 유영진
Original assignee: 주식회사 카이로스랩
Priority date: 2022-01-18
Filing date: 2022-11-16
Publication date: 2023-07-27

Abstract

A device for predicting the order quantity of a material required for a store may comprise: a sales data acquisition unit that acquires sales data of each of multiple products of the store; an external data acquisition unit that acquires external data including weather information; a sales data processing unit that processes the sales data to change first sub sales data corresponding to a first product among the multiple products and maintain second sub sales data corresponding to a second product among the multiple products on the basis of the external data; and a prediction unit that predicts future sales quantities of the multiple products on the basis of the processed sales data and predicts the order quantity of the material corresponding to the future sales quantities.

Description

Device and method for predicting order quantity of store materials

The present application relates to a device and method for predicting order quantities of store materials.

A small business owner who operates a store is making an order by figuring out the inventory amount of the product in a visual way before ordering the material. Since this ordering method does not manage inventory by standardizing data based on barcodes of products, it may cause difficulties in inventory management of stores.

If a small business owner can predict product sales volume in a standard way, it may be possible to efficiently manage the amount of inventory of items constituting each product. That is, since the small business owner can hold the optimal amount of stock in the store, it is possible to prevent excessive stock holding and reduce the disposal cost accordingly. In addition, the small business owner can prevent loss of sales opportunity cost due to shortage of products during cooking due to inventory shortage.

Recently, various technologies have been developed in order to predict required materials for products and sales volume of products and use them for management and marketing.

However, the conventional technology has a problem in that the accuracy of sales volume prediction is low without reflecting the industry and menu characteristics reflecting the characteristics of the store.

The background technology of the present application is disclosed in Korean Patent Publication No. 10-2021-0023119.

An object of the present application is to provide an apparatus and method for predicting future sales volume based on quantification of necessary materials corresponding to products to be sold, as to solve the problems of the prior art described above.

The present application is to solve the problems of the prior art described above, and analyzes the influence of external data (weather, traffic conditions, communication connection amount corresponding to the sale date) on the sales volume of each product, and based on this, it is an object to provide a device and method for predicting future sales volume and inventory lead time.

However, the technical problems to be achieved by the embodiments of the present application are not limited to the technical problems as described above, and other technical problems may exist.

As a technical means for achieving the above technical problem, in a device for predicting an order quantity of materials necessary for a store, a sales data acquisition unit acquiring sales data of each of a plurality of products of the store, an external data acquisition unit obtaining external data including weather information, a sales data processing unit processing the sales data to change first sub-sale data corresponding to a first product among the plurality of products based on the external data and maintain second sub-sale data corresponding to a second product among the plurality of products, based on the external data; It may include a prediction unit that predicts future sales volume of the plurality of products based on the processed sales data, and predicts an order quantity of the material corresponding to the future sales volume.

According to an embodiment of the present disclosure, the sales data processing unit may change the first sub-sale data so that a new data item and a value of the new data item are added to the first sub-sale data.

According to an embodiment of the present disclosure, the prediction unit may predict an order quantity of the material corresponding to the future sales volume based on the external data.

According to an embodiment of the present application, a plurality of necessary materials required for any one of the plurality of products and required material information including a quantity of each of the required materials is received from a store device of the store. Further comprising a required material information receiving unit, wherein the prediction unit may predict the order quantity of the material based on the required material information.

According to an embodiment of the present application, any one product includes a first necessary material and a second necessary material, and the quantity of the first necessary material in the store is equal to the quantity of the first necessary material in a second store different from the store, but the quantity of the second necessary material in the store may be different from the quantity of the second necessary material in the second store different from the store.

According to an embodiment of the present application, the recipe information search unit for retrieving recipe information corresponding to any one product from the web (WEB), but the prediction unit based on the recipe information and the required material information It can predict the order quantity of the material.

According to an embodiment of the present disclosure, the sales data is sales data for AI learning, and the prediction unit may predict the future sales volume and the order quantity based on an artificial intelligence algorithm.

According to an embodiment of the present disclosure, the prediction unit may predict the future sales volume based on an artificial intelligence algorithm and predict the order quantity based on the artificial intelligence algorithm.

According to an embodiment of the present application, in a method for predicting the order quantity of materials required for a store, obtaining sales data of each of a plurality of products of the store;

The method may include obtaining external data including weather information, processing the sales data to change first sub-sales data corresponding to a first product among the plurality of products based on the external data and maintaining second sub-sales data corresponding to a second product among the plurality of products based on the external data, predicting future sales volume of the plurality of products based on the processed sales data, and predicting an order quantity of the material corresponding to the future sales volume.

The above-described problem solving means are merely exemplary and should not be construed as intended to limit the present disclosure. In addition to the exemplary embodiments described above, additional embodiments may exist in the drawings and detailed description of the invention.

According to the above-mentioned problem solving means of the present application, the device for predicting the order quantity of the store material has the effect of managing the inventory status of the required material corresponding to a plurality of products based on future sales volume.

According to the above-described problem solving means of the present application, the device for predicting the order quantity of the store material has the effect of reducing the cost by confirming in advance the timing of ordering the necessary material.

However, the effects obtainable herein are not limited to the effects described above, and other effects may exist.

1 relates to a device for predicting order quantities of store materials according to an embodiment of the present disclosure.

2 is a configuration diagram schematically illustrating a data pre-processing process of a sales data processing unit according to an embodiment of the present disclosure.

3 is a configuration diagram schematically illustrating a process of pre-processing data when a corresponding menu does not exist in a public data set according to an embodiment of the present disclosure.

4 is an example in which the device for predicting the order quantity of store materials according to an embodiment of the present disclosure refers to a sub-sales table.

5 is a diagram showing a result of searching for recipe information according to an embodiment of the present application by way of example.

6 is a flowchart of a method for predicting the order quantity of store materials according to an embodiment of the present disclosure.

Hereinafter, embodiments of the present application will be described in detail so that those skilled in the art can easily practice with reference to the accompanying drawings. However, the present disclosure may be implemented in many different forms and is not limited to the embodiments described herein. And in order to clearly describe the present application in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the present specification, when a part is said to be “connected” to another part, this includes not only the case of being “directly connected” but also the case of being “electrically connected” with another element interposed therebetween.

Throughout the present specification, when a member is said to be located “on”, “upper”, “upper”, “below”, “below”, or “lower” another member, this includes not only the case where a member is in contact with another member, but also the case where another member exists between the two members.

Throughout the present specification, when a part "includes" a certain component, it means that it may further include other components without excluding other components unless otherwise stated.

Throughout the specification of the present application, the store device may be a user terminal.

1 relates to a device 100 for predicting order quantities of store materials, according to an embodiment of the present disclosure.

1, the device for predicting the order quantity of store materials 100 may include a sales data acquisition unit 110, an external data acquisition unit 120, a necessary material information reception unit 130, a sales data processing unit 140, and a prediction unit 150.

According to one embodiment of the present application, the mutual components of the store material order quantity prediction device 100 may communicate through a data management server. The data management server refers to a connection structure capable of exchanging information between nodes such as terminals and servers. Examples of such data management servers include a 3rd Generation Partnership Project (3GPP) network, a Long Term Evolution (LTE) network, a 5G network, a World Interoperability for Microwave Access (WIMAX) network, the Internet, a Local Area Network (LAN), a Wireless Local Area Network (LAN), a Wide Area Network (WAN), and a Personal Area Network (PAN). Network), wifi network, Bluetooth network, satellite broadcasting network, analog broadcasting network, DMB (Digital Multimedia Broadcasting) network, etc. are included, but are not limited thereto.

The data management server may extract information on necessary materials from the web using web scraping. Here, web scraping may be a program designed to extract only necessary data from among data appearing on a web page about a recipe.

In addition, the data management server may transmit information contained in the sales data acquisition unit 110 and the external data acquisition unit 120 on the web. The data management server may be an interface that provides information to utilize the web. For example, the data management server may utilize API (application, programming, interface).

According to one embodiment of the present application, the sales data acquisition unit 110 and the external data acquisition unit 120 may communicate with the user terminal. The user terminal is, for example, a smart phone, a smart pad, a tablet PC, and the like, and PCS (Personal Communication System), GSM (Global System for Mobile communication), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple) Access) and Wibro (Wireless Broadband Internet) terminals.

According to an embodiment of the present application, the sales data acquisition unit 110 may obtain sales data of each of a plurality of products of a store. In the description of an embodiment of the present disclosure, the sales data may include a sales amount corresponding to each of a plurality of products of a store, a list of a plurality of products sold in the store, and a date of sale of the plurality of products.

In this regard, the sales data acquisition unit 110 may acquire sales data through a sales management device in charge of managing point-of-sale information. Here, the sales management device may be a device for managing a payment system by a small merchant. In addition, the sales management device may be a point of sale (POS) that manages a payment system, or a kiosk capable of unmanned ordering in a store. In addition, the sales management device may manage the payment system through the user terminal. Here, the payment system may be an electronic financial service system in which a user directly trades products with a store and makes payment through a user terminal. However, it is not limited thereto.

[Table 1] POS sales data

The sales data in [Table 1] is exemplary, and sales data of other types of sales management devices may be applied.

The sales data acquisition unit 110 may acquire sales data from the sales management device at a preset cycle. For example, the sales data acquisition unit 110 may acquire sales data information of each product on a daily, weekly, and monthly basis through a module installed in the sales management device.

[Table 2] Sales data with sales management device

The sales volume by date of each menu in [Table 2] is exemplary, and a menu of a different type from the weekly and monthly sales volume may be applied.

The sales data acquisition unit 110 may update the sales data by combining the acquired sales data with the sales data before acquisition. The updated sales data has an effect of increasing the accuracy of predicting sales volume for a plurality of products. Here, the sales data acquisition unit 110 may acquire sales data of the sales management device in association with the user terminal.

In addition, sales data may be preprocessed through a data preprocessing module to predict inventory and sales volume.

[Table 3] Preprocessed sales data combined with previous sales data

Meanwhile, the sales data obtainer 110 may obtain similar sales data from similar stores in a store size and a radius within a preset distance when there is no sales data prior to acquisition from a store.

In addition, the sales data acquisition unit 110 may perform data pre-processing that integrates the sales data before acquisition through the data pre-processing module and the acquired similar sales data. Due to this, the store material order quantity prediction device 100 has an effect of introducing future sales volume prediction using similar sales data to small business owners who are considering opening a store.

The sales data acquisition unit 110 may receive locations of similar stores within a preset distance from a map library. Here, the map library may be a program related to a map service providing location and longitude data. The map service may provide a national land information base map, a digital geography map, a road name address, and the like. As an example, the map library may be Google Map.

The external data acquisition unit 120 may acquire external data including weather information. In the description of an embodiment of the present disclosure, the weather information may include temperature corresponding to a date, humidity, whether it is sunny, precipitation, wind direction and strength, and the like. In addition, the weather information may include temperature, humidity, whether it is sunny, precipitation, wind direction and strength, etc. corresponding to time within a date.

[Table 4] External Data

According to one embodiment of the present application, the external data obtaining unit 120 may acquire external data from the data management server. The weather server may collect a plurality of external data by using a data management server for a plurality of weather portal sites. In other words, the external data acquisition unit 120 may acquire a plurality of external data from the weather server. According to an embodiment of the present application, the plurality of weather portal sites may be weather data open portals, air pollution information, and the like. The plurality of external data may include weather information and air pollution information, but is not limited thereto.

Also, the external data acquisition unit 120 may obtain external data including information on traffic volume. In this case, the external data acquisition unit 120 may include public transportation information including information on getting on and off within a certain distance of the store by using location and longitude data. In this case, the boarding and alighting information may include the number of people boarding and alighting in correspondence with time.

Also, the external data obtaining unit 120 may obtain external data including the amount of communication connections. Here, the amount of communication connection may be an amount of connection per unit time of accessing a sales management device of a store through a user terminal. For example, the external data acquisition unit 120 may obtain information about the amount of product order requests from the user terminal and the amount of product order requests corresponding to unit time.

Also, the external data acquisition unit 120 may obtain external data including at least one of weather, traffic, and communication connection. In other words, the external data acquisition unit 120 may acquire a plurality of external data including weather, traffic, and communication connection. Each of the plurality of external data may be information corresponding to weather, traffic, and communication connection.

In other words, the external data acquisition unit 120 may add and collect external data affected by weather so that the prediction unit 150, which will be described later, increases the accuracy of forecasting future sales volume.

The necessary material information receiving unit 130 may receive necessary material information including a plurality of necessary materials required for any one product among a plurality of products and a quantity of each of the necessary materials from a store device of a store. Here, the product may be a food corresponding to a list of food prepared by the restaurant. Required material information may be components of food required to prepare food. In addition, the necessary material information may include a basic unit so as to measure a certain amount with respect to the physical quantity of each of the necessary materials. The required material information receiving unit 130 may include required material information corresponding to the basic unit of the quantity of each required material. The basic unit may include g (gram) and Kg (kilo gram), which are one of units representing weight. In addition, the store may be a restaurant that manufactures food and sells it to consumers. It is not limited to this.

The required ingredient information receiving unit 130 may convert units used for cooking into basic units. The required ingredient information receiving unit 130 may convert each cooking unit of a plurality of required ingredients into a basic unit through the data management server. For example, the cooking units of onions (3/5 pieces) and sesame leaves (4 sheets) can be converted into basic units of onions (150g) and sesame leaves (8g). Since the conversion method of the required material information receiving unit 130 is obvious to those skilled in the art, a detailed description thereof will be omitted.

To assist in an embodiment of the present disclosure, the first sub sales data and the second sub sales data may be any one of a plurality of sub sales data. Here, the sub sales data may include at least one or more external data among sales data and a plurality of external data.

According to an embodiment of the present application, the sales data processing unit 140 may change and maintain the configuration of sub sales data based on the correlation coefficient. The sales data processing unit 140 may add or delete at least one external data among a plurality of external data in the configuration of the sub sales data.

Specifically, the sales data processing unit 140 may change first sub-sale data corresponding to a first product among a plurality of products based on external data and process the sales data to maintain second sub-sale data corresponding to a second product among a plurality of products. The sales data processing unit 140 may change the first sub sales data when a correlation coefficient of any one of the plurality of external data is less than a reference value through a linear regression model. The sales data processing unit 140 may maintain second sub sales data when a correlation coefficient of each of a plurality of external data is greater than or equal to a reference value.

Also, the sales data processing unit 140 may configure a sub sales table by combining sales data and external data having a correlation coefficient greater than or equal to a reference value among a plurality of external data. In addition, the sub sales table may be configured to correspond to external data having a product and a correlation coefficient greater than or equal to a reference value.

[Table 5] Data set combining sales data and external data

For example, referring to [Table 5], a sub sales table data set may be configured corresponding to temperature, which is external data of an iced Americano product. Here, the sub sales table may be managed by communicating with the data management server. In addition, the sub sales table data set may be referenced and used in the prediction unit 150 to be described later.

Also, the sales data processing unit 140 may derive a plurality of correlation coefficients corresponding to each of the plurality of external data and the sales data. The sales data processing unit 140 may change and maintain a plurality of external data based on a plurality of correlation coefficients.

The correlation coefficient may be derived by linear regression analysis. Here, the linear regression analysis is a regression analysis technique that identifies the relationship between the dependent variable and the independent variable and predicts the dependent variable through modeling of the dependent variable Y and one or more independent variables X. Since it is obvious to those skilled in the art, a more detailed description will be omitted.

The sales data may be sales data for AI learning. AI that learns based on time series, which will be described later, can learn based on AI learning sales data. Here, the artificial intelligence learning sales data may set different weights of data items based on the correlation coefficient.

2 is a configuration diagram schematically illustrating a data pre-processing process of the sales data processing unit 140 according to an embodiment of the present disclosure.

Referring to FIG. 2 , the sales data processing unit 140 may pre-process a plurality of external data and sales data.

Specifically, the sales data processing unit 140 acquires sales data from POS, KIOSK, PC, WEB, etc., combines and preprocesses a preprocessed sales data set and a preprocessed external data set through a data preprocessing module, and generates prediction data used to predict future sales.

For example, data predicting future sales of iced americano can be generated through a preprocessing process that acquires iced americano sales data from POS, KIOSK, PC, and WEB, and combines the preprocessed sales data set with an external data set corresponding to iced americano from a public data set table.

Referring to FIG. 3 , if an external data combination corresponding to a corresponding menu does not exist in the public data set, the data processing unit 140 may combine each external data and select a combination having the smallest error to update the public data set.

For example, when watermelon juice sales data is obtained and there is no public data set corresponding to the preprocessed sales data set, watermelon juice and each external data (temperature, humidity, day of the week, amount of sunlight, bus, subway, etc.) are combined, the combination with the smallest error is selected, the public data set is updated, and predictive data for predicting future sales of watermelon juice can be generated.

In addition, when there is no existing sales data, the sales data processing unit 140 may pre-process sales data for each menu by comparing region and industry data and applying similar sales data.

Specifically, sales data and requested external data may be received and each external data may be preprocessed through a data preprocessing module.

For example, for external data, external data sources such as national spatial information portal, national statistics portal, statistical information sharing service, weather data open portal, highway public data portal, etc. can be linked with API to request and receive external data including weather, traffic volume, and communication connection volume, and each data can be preprocessed through a data preprocessing module.

Specifically, the sales data processing unit 140 may generate an input data set for each menu by combining actual sales data including menu, date, and sales volume with open external data including traffic, weather, communication, temperature, humidity, and the like.

According to one embodiment of the present application, the prediction unit 150 may predict future sales of a plurality of products based on the processed sales data. The prediction unit 150 may predict future sales volume of a plurality of products based on a plurality of external data having a correlation coefficient greater than or equal to a reference value. The prediction unit 150 may extract a correlation coefficient for each of a plurality of external data by applying a linear regression analysis capable of deriving the correlation coefficient.

Specifically, the prediction unit 150 may predict the order quantity of materials corresponding to future sales volume. The prediction unit 150 may predict future sales of a plurality of products by returning to sales data based on external data and sales data. The predictor 150 may predict future sales volume by returning to previous sales data based on external data and sales data acquired at a preset period.

In addition, the prediction unit 150 may predict the order quantity of materials corresponding to the future sales volume based on the external data. In addition, the prediction unit 150 may select a plurality of sub-sale data having a similarity of a predetermined value or more based on a plurality of external data having a correlation coefficient equal to or greater than a reference value. Here, the plurality of sub sales data may include a plurality of external data having an ideal value of the reference value.

Specifically, the prediction unit 150 may predict future sales volume by using the sub-sale data having the largest similarity among the selected plurality of sub-sale data. In other words, the prediction unit 150 may predict future sales volume by using the processed sales data having the largest similarity among the plurality of processed sales data. Here, the degree of similarity may be based on any one of components of sub sales data and processed sales data.

According to one embodiment of the present application, the prediction unit 150 sets the sales data (one of a plurality of products) of a product to be predicted as a dependent variable Y and any one of a plurality of external data as an independent variable X, and performs linear regression analysis based on Equation 1 to derive a correlation coefficient.

[Equation 1]

Y = b*x+ epsilon

Here, b is a parameter vector calculated using the least squares method, and ipsilon may represent an error term.

The correlation coefficient may include a Root Mean Square Error (RMSE) of any one of a plurality of external data and sales data.

For example, the prediction unit 150 may predict future sales volume of a product by determining based on the order of external data having a large correlation coefficient. In other words, the predictor 150 may determine the sales data based on a plurality of data in order of a root mean square deviation in a small order.

For better understanding, it is assumed that, for example, the maximum temperature and the average temperature are any one of a plurality of external data having a correlation coefficient greater than or equal to a reference value. Assume that the correlation coefficient of the external data with the highest temperature is 0.8 and the correlation coefficient with the lowest temperature is 0.6. Also, let's assume that the maximum temperature obtained periodically for the given input is 20 degrees and the average temperature is 10 degrees. At this time, the prediction unit 150 may select a plurality of sub-sale data having a predetermined or higher similarity at a maximum temperature of 20 degrees as a first step. As a second step, the prediction unit 150 may select a plurality of sub-sales data having a predetermined or higher similarity at an average temperature of 10 degrees. Finally, the prediction unit 150 may predict future sales volume using sub-sale data having the greatest similarity. Here, the sub sales data may be processed sales data. As a result, the store material order quantity prediction device 100 may predict future sales volume for a plurality of products based on the future sales volume.

4 is an example in which the order quantity prediction device 100 of store materials according to an embodiment of the present disclosure refers to a sub sales table.

The present application will explain the procedures and references for configuring the sub sales table in the store material order quantity prediction device 100 in FIG. 4 . In addition, in the present application, the order quantity prediction device 100 of the store material in FIG. 4 will describe the prediction of the future sales volume with reference to the sub-sales table.

Referring to FIG. 4 , the device for predicting the order quantity of store materials 100 may refer to the sub sales table. In this case, the list of products of the sub sales table may be products manufactured by the store. The store material order quantity prediction device 100 may generate a sub-sales table based on the sales data obtained from the sales data acquisition unit 110 and the external data obtained from the external data acquisition unit 120 . Here, the sub sales table may be created based on the sub sales data.

Specifically, the sales data acquisition unit 110 may receive a list of products from a store through a data management server. The sales data acquisition unit 110 may obtain a list of received products and organize them in a sub sales table.

The external data acquisition unit 120 may receive external data through a data management server. The received external data may be configured in a sub sales table in correspondence with a product list based on a correlation coefficient.

The sales data processor 140 may generate a plurality of sub sales data based on the external data and the sales data. In this case, the sales data processing unit 140 may configure external data having a correlation coefficient greater than or equal to a reference value among a plurality of external data based on the sales data as sub sales data.

The sales data processing unit 140 may change the first sub-sale data so that a new data item and a value of the new data item are added to the first sub-sale data. The sales data processing unit 140 may add and change external data into the first sub sales data based on the correlation coefficient. The sales data processing unit 140 may correspond sales data and external data to dates.

The sales data processing unit 140 may create a sub sales table based on the added and changed first sub sales data. The sales data processing unit 140 may extract data items from sub sales data. The sales data processing unit 140 may include external data in the sub sales table by corresponding data items. For this reason, the sales data processing unit 140 may check whether or not an unregistered product of the store is received based on the sub sales table when an unregistered product unregistered in the store is received.

When an unregistered product that is not registered in the sub sales table is received, the sales data processing unit 140 requests the recipe information search unit 20 to extract the recipe information by web scraping the unregistered product from the web. In this case, the recipe information may be information about necessary ingredients needed to cook the product.

In this regard, FIG. 5 is a diagram showing a result of searching for recipe information according to an embodiment of the present disclosure by way of example.

Referring to FIG. 5 , the recipe information search unit 20 may retrieve recipe information corresponding to any one product from the web (WEB). The recipe information search unit 20 may select recipe information from a plurality of webs based on an adoption rate generated using at least one of frequency, recommendation, inquiry, and site reliability. Here, the selected recipe information may be a necessary material corresponding to an unregistered product. Specifically, the adoption rate may be calculated by the following [Equation 1].

[Equation 1]

In this regard, the adoption rate can be calculated by inputting frequency weight to α, recommendation weight to β, hit weight to γ, and site reliability to δ in [Equation 1].

In addition, the recipe information search unit 20 may select recipe information having the highest calculated adoption rate score among the retrieved recipes.

The prediction unit 150 may predict the ordered quantity of ingredients based on recipe information and required material information. The predictor 150 may determine the similarity with previous sales data based on the obtained sales data to derive the order amount of the product. The prediction unit 150 may predict the order quantity of the material based on the order quantity of the product, the recipe information, and the material information in the required material information.

Any one of the plurality of products may include a first required material and a second required material. In other words, a plurality of products in the sub sales table may include different recipe information. The order quantities of the materials may be predicted based on the order quantities of products corresponding to the first required material and the second required material, respectively.

The quantity of the first necessary material in the store is the same as the quantity of the first necessary material in the second store different from the store, but the quantity of the second necessary material in the store may be different from the quantity of the second necessary material in the second store different from the store. In addition, the order quantity of the material of the store and the second store different from the store may be different from the future order quantity corresponding to each.

The prediction unit 150 may predict future sales volume and order quantity based on an artificial intelligence algorithm. The predictor 150 may predict future sales volume based on an artificial intelligence algorithm learned based on sales data for artificial intelligence learning. In this case, the sales data for learning may be a combination of external data and sales data based on a correlation coefficient.

The artificial intelligence algorithm may be an artificial intelligence model based on long short-term memory (LSTM) that is effective for learning sequence data. In addition, according to an embodiment of the present application, the artificial intelligence model may be a bidirectional long short-term memory (LSTM) model in which accuracy is increased by specifically learning learning data in forward and backward directions. In addition, this bidirectional LSTM-based artificial intelligence model may be configured with an architecture including a batch normalization layer and a BiLSTM layer.

The LSTM (Long-Short Term Memory) algorithm is one of the artificial recurrent neural network (RNN) architectures used in the field of deep learning. Unlike feed-forward neural networks, feedback connections exist. Therefore, according to the LSTM algorithm, learning and processing can be performed not only for a single data point but also for an entire data sequence.

These LSTM algorithms are suitable for classifying, processing, and predicting predictions based on time series data, and LSTM has the advantage of resolving the vanishing gradient problem that can occur in training through traditional RNNs.

However, the type of artificial intelligence model for predicting future sales volume disclosed herein is not limited to the above-described LSTM algorithm, and according to an embodiment of the present invention, in addition to the LSTM algorithm, Attention algorithm, Transformer algorithm, BERT (Bidirectional Encoder Representations from Transformers) algorithm, etc. may include time-series data analysis-based artificial intelligence algorithms that predict future sales volume through analysis of sequence data reflecting time-series changes in sales data.

That is, in the present application, algorithm models based on various time-series data analysis previously known or developed in the future may be applied.

The predictor 150 may predict inventory required time of the store and a second store different from the store based on future sales volume and required materials. The prediction unit 150 may compare future sales volume corresponding to the order amount of the store and the second store different from the store with the inventory amount of products remaining in the store to predict the inventory required time. In addition, the predictor 150 may predict required inventory times for each of the first required material and the second required material.

Meanwhile, when an unregistered product is received, the prediction unit 150 may add the unregistered product to the sub sales table based on an artificial intelligence learning algorithm. The prediction unit 150 may refer to the sub sales table by making a request to the sales data processing unit 140 . In this case, the sales data processing unit 140 may generate a plurality of artificial intelligence learning algorithms based on each of a plurality of combined external data combined based on a plurality of external data.

The prediction unit 150 may select combination external data that forms the minimum gradient of the loss function when a plurality of artificial intelligence learning algorithms corresponding to each of the plurality of combination external data are trained. Here, the artificial intelligence learning algorithm may refer to a model that finds a relationship of one or more dependent variables using the most appropriate straight line. This is created as a model that minimizes the mean squared error (MSE) of the difference between the actual value and the predicted value, and in general, the performance of the model can be measured by the ratio of the variance ratio (R-square) of the predicted variable.

The prediction unit 150 may register the selected combined external data and the non-registered product in the sub sales table. Accordingly, the prediction unit 150 may predict future sales volume by referring to the sub sales table even when unregistered products are received.

The prediction unit 150 may predict future sales of unregistered products based on similar sales data. In addition, the prediction unit 150 may predict the order quantity of products by applying an artificial intelligence learning algorithm based on similar sales data. Due to this, the prediction unit 150 can confirm in advance when to order the necessary materials, and thus has an effect of reducing costs.

The prediction unit 150 may request the ordering party to supply the required material based on the inventory required time. At this time, the prediction unit 150 may request to supply a preset number of necessary materials.

6 is a flow chart of a method for predicting the order quantity of store materials according to an embodiment of the present disclosure.

Hereinafter, based on the details described above, the operation flow of the present application will be briefly reviewed.

In step S11, the order quantity prediction device 100 of the store materials may obtain sales data of each of a plurality of products of the store.

In step S12, the device 100 for predicting order quantities of store materials may obtain external data including weather information.

In step S13, the store material order quantity prediction device 100 changes the first sub-sale data corresponding to the first product among the plurality of products based on the external data and the second product among the plurality of products. The sales data may be processed to maintain the second sub-sale data corresponding to the sales data.

In step S14, the store material order quantity prediction device 100 may predict the future sales volume of a plurality of products based on the processed sales data, and predict the order quantity of materials corresponding to the future sales volume.

In the foregoing description, steps S11 to S14 may be further divided into additional steps or combined into fewer steps, depending on an embodiment of the present invention. Also, some steps may be omitted if necessary, and the order of steps may be changed.

The method for predicting the order quantity of store materials according to an embodiment of the present application may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. Program instructions recorded on the medium may be those specially designed and configured for the present invention or those known and usable to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program instructions such as ROM, RAM, and flash memory. Examples of program instructions include high-level language codes that can be executed by a computer using an interpreter, as well as machine language codes such as those produced by a compiler. The hardware devices described above may be configured to act as one or more software modules to perform the operations of the present invention, and vice versa.

In addition, the above-described method for predicting the order quantity of store materials may be implemented in the form of a computer program or application executed by a computer stored in a recording medium.

The above description of the present application is for illustrative purposes, and those skilled in the art will understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present application. Therefore, the embodiments described above should be understood as illustrative in all respects and not limiting. For example, each component described as a single type may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined form.

The scope of the present application is indicated by the following claims rather than the detailed description above, and all changes or modifications derived from the meaning and scope of the claims and equivalent concepts thereof should be construed as being included in the scope of the present application.

Claims

In a device for predicting the order quantity of materials required for a store,

a sales data acquisition unit acquiring sales data of each of a plurality of products of the store;

an external data obtaining unit acquiring external data including weather information;

a sales data processing unit that processes the sales data to change first sub-sale data corresponding to a first product among the plurality of products based on the external data and maintain second sub-sale data corresponding to a second product among the plurality of products; and

An order quantity prediction device for store materials including a predictor for predicting future sales volume of the plurality of products based on the processed sales data and predicting an order quantity of the material corresponding to the future sales volume.
According to claim 1,

The sales data processing unit,

and changing the first sub-sale data so that a new data item and a value of the new data item are added in the first sub-sale data.
According to claim 1,

The prediction unit,

The device for predicting the order quantity of the store material, which predicts the order quantity of the material corresponding to the future sales volume based on the external data.
According to claim 1,

Further comprising a necessary material information receiving unit for receiving necessary material information including a plurality of necessary materials required for any one product of the plurality of products and a quantity of each of the necessary materials from a store device of the store,

Wherein the prediction unit predicts the order quantity of the material based on the required material information, the store material order quantity prediction device.
According to claim 4,

Any one of the above products includes a first necessary material and a second necessary material,

The quantity of the first required material in the store is the same as the quantity of the first required material in the second store different from the store, but the quantity of the second required material in the store is different from the store. The second required material quantity in the second store is different from the quantity of the required material in the store.
According to claim 4,

Further comprising a recipe information search unit for retrieving recipe information corresponding to any one product from the web (WEB),

Wherein the prediction unit predicts the order quantity of the material based on the recipe information and the required material information, the store material order quantity prediction device.
According to claim 1,

The sales data is sales data for artificial intelligence learning,

Wherein the prediction unit predicts the future sales volume and predicts the order quantity based on an artificial intelligence algorithm.
According to claim 7,

wherein the prediction unit predicts the future sales volume based on an artificial intelligence algorithm, and predicts the order quantity based on the artificial intelligence algorithm.
In the method of predicting the order quantity of materials required for the store,

acquiring sales data of each of a plurality of products of the store;

obtaining external data including weather information;

Processing the sales data to change first sub-sale data corresponding to a first product among the plurality of products based on the external data and maintain second sub-sale data corresponding to a second product among the plurality of products; and

Predicting a future sales volume of the plurality of products based on the processed sales data, and predicting an order quantity of the material corresponding to the future sales volume.
A computer-readable recording medium on which a program for executing the method of claim 9 is recorded on a computer.