JP2016133816A - Commodity demand prediction system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a commodity demand prediction system for predicting commodities which will become well-selling commodities, based on past purchase information of plural commodities, adjusting a number of purchasing them, suppressing a number of a dead stock, and contributing to increase of sales.SOLUTION: A commodity demand prediction system is configured to: store purchase history information in which information for specifying customers who purchased plural commodities which are sold at a store, and selling information of the commodities of the customers are associated; make a ranking of a sales of the plural commodities in a past first predetermined period, based on the purchase history information; extract the commodities whose sales rank is higher rank, as past well-selling commodity information; identify customers who purchased the commodities related to the past well-selling commodity information in a second predetermined period, as purchasers of past well-selling commodities; extract latest purchase commodity information which is information related to commodities which plural past purchasers of well-selling commodities purchased in a latest third predetermined period, from the purchase history information; and make a ranking of a number of purchasers of past well-selling commodities who purchased commodities in the third predetermined period, for predicting the commodities whose purchaser number rank is higher rank, as future well-selling commodities.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a product demand prediction system that predicts a product that is a best seller from past purchase information of a plurality of products sold in a store.

  In apparel sales and general merchandise sales, there is a technique for predicting future best-selling products from the number of products sold and purchasing them in advance, thereby increasing the number of sales and improving sales. Conventionally, this prediction method relies mainly on the intuition and experience of skilled buyers. For this reason, it is not easy to appropriately grasp the best-selling products and to appropriately purchase products that are expected to increase in the near future.

  Buyers use information on past best-selling products (for example, products with the highest number of sales (within the top 10) in a certain period in the past), temperature, weather, trend information, etc. as information for forecasting top-selling products in the near future Often done. However, it is difficult to predict the best-selling products and their purchase quantities directly from such information, and it is necessary to rely on the intuition and experience of skilled buyers to determine the final products and their quantities. The situation continues.

  It takes a long time to train skilled buyers. Moreover, in the current situation where intuition and experience are the main points in making judgments, even experienced buyers can avoid the loss of product sales opportunities due to excess inventory or missing items. There wasn't.

JP 2006-139343 A

  As described above, in view of the difficulty in predicting a hot selling product, the present invention predicts a product that will be a future selling item from past purchase information of a plurality of products, and a product that is predicted to become a selling item. The purpose is to provide a product demand forecasting system that adjusts the number of purchases at an early stage from the start of sales, suppresses defective product inventory, and contributes to improved sales.

  In order to solve the above-mentioned problems, the inventor pays attention to products that have been sold from the past purchase information of a plurality of products and customers who have purchased the product, and the analysis of the product purchase trends by the customer predicts products that will be sold in the future. It was found useful and reached the present invention.

That is, the invention of claim 1
Purchasing history information storage means for storing, for a plurality of products sold in a store, information for identifying a customer who has purchased the product and sales history information in which sales information of the product by the customer purchasing the product is associated. When,
Based on the purchase history information, the number of sales of the plurality of products sold at the store in the first predetermined period in the past is ranked, and the products from the top to the predetermined ranking Specific period selling product information extraction means for extracting as a specified period selling product information that is past selling product information,
The customer who purchased the product related to the selling product for the specified period in the second predetermined period in the past is recognized as the selling product purchaser for the specified period which is the past selling product purchaser, and sold for the specified period. Specific period selling goods purchaser information, which is information relating to the line item purchaser, is generated, and the specified period selling goods purchaser information is stored in the specified period selling goods purchaser information DB. Generating means;
The purchase information on the latest purchase product information, which is information related to the product purchased by the plurality of sale products for the specified period sold in the specific period sale product purchaser information DB, in the latest third predetermined period. Latest purchase product information extraction means to extract from history information,
Based on the latest purchase product information, among the products sold in the latest third predetermined period included in the latest purchase product information, a purchase act is performed during the latest third predetermined period. Ranking the number of the purchasers of the selling products for the specified period that were made, and the number of selling products for the specified period selling products was sold in the third predetermined period from the top to the predetermined rank This is a product demand forecasting system comprising a sales product prediction information generating means for generating sales product prediction information using a product as a future sales product.

The invention of claim 2
The selling product forecast information is based on the specified period selling product information for each predetermined product group among the plurality of products sold in the store, and the specified period selling product information for each predetermined product group. Of a plurality of products sold at the store, which are generated based on the specific period selling goods purchaser information for each predetermined product group generated and the most recently purchased product information for each predetermined product group. The commodity demand prediction system according to claim 1, wherein the commodity commodity prediction information is for each predetermined commodity group.

The invention of claim 3
The point-of-sales product prediction information is specified at the specific store generated based on the point-of-sale item information at the specific store in the plurality of stores and the point-of-sale product information at the specific store. The selling product prediction information at a specific store among the plurality of stores, which is generated based on the period selling product purchaser information and the latest purchase product information at the specific store. 3. A commodity demand prediction system according to 1 or 2.

The invention of claim 4
The specified period selling product purchaser information generating means performs a process of generating the specified period selling product purchaser information in each of the plurality of consecutive first predetermined periods in the past. When storing the selling product purchaser information for the specific period generated in one predetermined period in the selling product purchaser information DB, a plurality of the first selling items stored in the selling product purchaser information DB are stored. 4. The process of deleting the specified period selling product purchaser information of the oldest period in the specified period selling product purchaser information in one predetermined period. A commodity demand prediction system according to claim 1.

  Such a commodity demand prediction method executed by the system of the present invention includes the following steps.

  For a plurality of products sold in a store, purchase history information that associates information for identifying a customer who has purchased the product with sales information of the product by the customer purchasing the product is stored in the purchase history information storage unit. Process.

  Based on the purchase history information, the number of sales of the plurality of products sold at the store in the first predetermined period in the past is ranked, and the products from the top to the predetermined ranking The process of extracting as a sale item product information for a specific period, which is past sale item information.

  The customer who purchased the product related to the selling product for the specified period in the second predetermined period in the past is recognized as the selling product purchaser for the specified period which is the past selling product purchaser, and sold for the specified period. A step of generating specific period selling goods purchaser information that is information relating to a line item purchaser and storing the specified period selling goods purchaser information in the specified period selling goods purchaser information DB.

  The purchase information on the latest purchase product information, which is information related to the product purchased by the plurality of sale products for the specified period sold in the specific period sale product purchaser information DB, in the latest third predetermined period. A process of extracting from history information.

  Based on the latest purchase product information, among the products sold in the latest third predetermined period included in the latest purchase product information, a purchase act is performed during the latest third predetermined period. Ranking the number of the purchasers of the selling products for the specified period that were made, and the number of selling products for the specified period selling products was sold in the third predetermined period from the top to the predetermined rank A process of generating sales forecast information as a future selling product.

  The customer who purchased the product related to the selling product information for the specified period in the second predetermined period in the past is identified as the selling product purchaser for the specified period that is the past selling product purchaser, and the specified In the step of generating specific-period selling product purchaser information, which is information related to the periodic selling product purchaser, and storing the specified-period selling product purchaser information in the specified-period selling product purchaser information DB, the past In the plurality of first predetermined periods that are consecutive, the processing for generating the specific period selling product purchaser information is performed, and the specific period selling generated in the first predetermined period most recently. When storing merchandise purchaser information in the selling merchandise purchaser information DB, the merchandise purchasers for the specific period in the plurality of first predetermined periods stored in the selling merchandise purchaser information DB Affection In, it can be made with a process for deleting the specific period sales muscle purchaser information of the oldest time.

  According to the present invention, a product that is expected to be sold in the future is predicted from past purchase information of a plurality of products, and the number of purchased products that are predicted to be sold is adjusted at an early stage from the start of sales. It is possible to provide a product demand forecasting system that reduces inventory and contributes to sales improvement.

It is a figure showing an example of the system configuration | structure of this invention. It is a figure showing an example of composition of a server apparatus. It is a figure showing an example of the past selling merchandise information. It is a figure explaining the authorization of the past selling goods purchaser, and its authorization period. An example of a specific period selling product table is shown among the specific period selling product purchaser information. An example of a specific period selling product purchaser table among the selling items for specific period selling products is shown. It is a figure explaining the specific period selling goods purchaser accumulation | storage information utilized for the prediction process of future selling goods. It is a conceptual diagram explaining the sales peak of the 1st thru | or 4th predetermined period used by this invention, and clothing. It is a figure showing the example of the list of goods predicted as future hot selling goods by this invention. FIG. 10 is a diagram illustrating the number of sales for five weeks from the start of sales and the sales ranking by item name for the product illustrated in FIG. 9.

  Hereinafter, an example of an embodiment of the present invention will be described with reference to the accompanying drawings. In this embodiment, the “product” for predicting the selling is used as the clothing item, but the selling product can be predicted for other various products. In addition, in the following, an embodiment of a product demand prediction system based on product sales information at department store stores is described. However, product demand prediction is not limited to department store stores and product demand prediction at stores that sell various products. Can be a system.

1. Configuration of the present invention The commodity demand forecasting system of the present invention includes a purchase history information storage means, a specified period selling product information extracting means, a specified period selling product purchaser information generating means, a recently purchased product information extracting means, a selling product, which will be described later. It consists of a computer equipped with prediction information generation means and the like.

  In the embodiment shown in FIG. 1, such a computer is used as the server device 40. The server device 40 is connected to the client terminal 20 and the store side terminal 30 used by a department store or the like so as to be able to communicate with each other via a communication network 50 such as an Internet communication network or a wireless communication network defined by a wireless communication standard. It is embodiment comprised.

  The client terminal 20 includes a personal computer and a tablet terminal used in each department such as a department store.

  The store side terminal 30 includes various terminals such as a POS terminal used in a POS system in a store such as a department store, a personal computer used in an office in a store such as a department store, and a tablet terminal.

  Conventionally known business systems such as a POS system and a customer system for grasping and managing sales result information such as product name, price, quantity, date and time of a product that has been sold and information related to a product purchaser (customer) are the client terminals. 20 and the store side terminal 30.

  The server device 40 includes information transmission / reception means 41 that performs mutual information communication with the client terminal 20 and the store side terminal 30 via a communication network 50 such as a wireless LAN or an Internet communication network.

  The server device 40 includes a purchase history information storage means 42a, a specified period selling goods information extraction means 44a, a specified period selling goods purchaser information database 45 (hereinafter, the database is referred to as "DB" in the present specification and drawings), Period selling goods purchaser information generating means 44b, latest purchased goods information extracting means 46a, selling goods forecast information generating means 46b are provided.

  In the illustrated embodiment, the purchase history information management unit 42 manages purchase history information tabulation processing for a plurality of products sold in a store. The purchase history information storage means 42a in the purchase history information management unit 42 specifies information for identifying a customer who has purchased the product for a plurality of products sold in the store, and sales of the product by the customer purchasing the product. The purchase history information associated with the information is stored.

  In the illustrated embodiment, the specified period selling item product information extraction means 44a in the selling item information management unit 44 for the specified period sells at the store during the first predetermined period in the past from the purchase history information. The number of sales of the plurality of products being ranked is ranked, and the products whose sales numbers are from the top to a predetermined rank are extracted as sales information for a specific period, which is past sales information.

  Similarly, in the illustrated embodiment, the specified period selling product purchaser information generation unit 44b in the specified period selling product purchaser information management unit 44 sets the product related to the specified period selling product information in the past. The customer purchased in the predetermined period of time is identified as a purchaser of a certain period selling product that is a past selling product, and the selling item purchaser information for a specified period that is information related to the selling product for the specified period Is stored in the specific period selling goods purchaser information DB 45.

  The specified period selling product purchaser information DB 45 stores the specified period selling product purchaser information.

  In the illustrated embodiment, the most recently purchased product information extraction means 46a in the selling product prediction unit 46 is the latest selling product purchaser for a specified period stored in the specified period selling product purchaser information DB 45. The latest purchase product information, which is information related to the product purchased during the third predetermined period, is extracted from the purchase history information.

  Similarly, in the illustrated embodiment, the selling product prediction information generating unit 46b in the selling product prediction unit 46 is based on the most recently purchased product information, and the latest third item included in the most recently purchased product information. Among the products sold during a predetermined period of time, ranking the number of items sold by the specified period selling goods that have been purchased during the most recent third predetermined period, The sale item prediction information is generated with the sale of the product sold in the latest third predetermined period from the highest number of purchasers to a predetermined rank as the future sale item.

  In the illustrated embodiment, the server device 40 associates information for identifying a customer who has purchased the product, which is purchase history information for the plurality of products, with sales information of the product by the customer purchasing the product. Purchase history information DB 43 for storing the information.

  Information processing performed in each of these units is controlled by a control unit 47 including a CPU, a RAM, a ROM, and the like.

2. Purchase history information tabulation processing of the product The purchase history information management unit 42 of the server device 40 manages the purchase history information tabulation processing for a plurality of products sold in the store.

  The purchase history information is information in which, for a plurality of products sold at a store, information for identifying a customer who has purchased the product is associated with sales information of the product by the customer purchasing the product. The aggregation of the information is performed by the purchase history information storage unit 42a included in the purchase history information management unit 42.

  In the illustrated embodiment, for example, the store-side terminal 30 such as the POS terminal reads information identifying a customer from a card presented when the customer purchases the product, and the server device 40 is information identifying the customer. Is acquired from the POS system or the customer system.

  The information that identifies the customer is information that is used to identify the purchaser of the merchandise that will be described later, and what and when the customer purchased the product, the purchased product, and the product. It is sufficient if information is stored in association with the purchased customer. For example, the card number presented by the customer when purchasing the product may be used.

  In the embodiment shown in the drawing, the sales of a product from, for example, a JAN (Japanese Article Number) code displayed on a tag of a product purchased by a customer at the store side terminal 30 such as the POS terminal used in the POS system. The server device 40 reads the information, and acquires the sales information of the product from the POS system or the customer system together with the information specifying the customer.

  In this way, the server device 40 acquires the sales information of the product due to the customer making a purchase action at the store and the information specifying the customer who purchased the product from the POS system or the customer system, Information for identifying the customer and sales information of the product obtained by the customer making a purchase at the store are associated with each other and stored in the purchase history information DB 43 and managed.

  The sales information of the product resulting from the purchase by the customer at the store includes information for specifying the product. As information for specifying the product, when the product whose sales are to be predicted is “clothing” as in this embodiment, for example, the brand, item name (jacket, cut-and-sew, skirt, coat, etc.), attribute of the product (Design, color, size, etc.). Moreover, it can also be set as the structure where the store which sold the said goods, the product number used in the said store, a manufacturer model number, etc. are contained.

  Such sales information of a product can be obtained by reading the JAN code attached to the product at a store where the product is purchased.

3. Extraction processing of past selling goods information The specified period selling goods information extraction means 44a in the specified period selling goods purchaser information management unit 44 of the server device 40, for example, from the purchase history information according to the conditions exemplified below. , Ranking the number of sales of the plurality of products sold at the store in the first predetermined period in the past, and selling the products from the top to the predetermined ranking It is extracted as merchandise information for a specific period that is merchandise information.

<Conditions for extracting past selling merchandise information>
Extraction month: Extraction date of each month: First Monday purchase history information tabulation period (first predetermined period): 1st of “current month-1” month to the last day of “current month-1” month Target: specify product Information (for example, information including product brand, item name, attribute, store selling the product, product number used at the store, manufacturer model number, etc.)
Ranking method: Order of extraction with the largest number of sales Target: Brand name of each store-Manufacturer's model number within the top 10 sales by item name

  In this embodiment, the product demand prediction is performed for each predetermined product group among a plurality of products sold at the store. Therefore, for example, a brand of a product, an item name, etc., which are information necessary for certifying that the product is a specific product group, are set as “information for specifying a product”.

  Further, when the product demand prediction at a specific store is performed, as described above, information related to the store where the product is sold is included in the “information specifying the product” and collected.

  The specific period selling goods information is used for information for predicting future selling goods as will be described later. Therefore, in this embodiment, the purchase history information is displayed for the period from “the first day of the“ current month-1 ”month to the last day of the“ current month-1 ”month, which is the most recent month for generating the sale item information for a specific period. This is a “first predetermined period” that is a period of aggregation (FIG. 8). The “predetermined first period”, which is the most recent predetermined period in the past for generating the selling information for the specified period, can be determined in consideration of the characteristics of the product for which the selling is predicted. The last two weeks or three weeks in the past in which the product information is generated can be set as the “first predetermined period”.

  According to the conditions in this embodiment exemplified above, the specific period selling goods information extraction means 44a totals purchase history information for September 2014 from the purchase history information DB 43 as shown in FIG. Then, on the first Monday of October 6, 2014, the manufacturer's model number of the top 10 items of sales by brand-item name for each store in September 2014 is extracted.

  In this manner, the specific period selling goods information extraction means 44a can extract the specific period selling goods information for each predetermined product group among a plurality of products sold at the store.

  Further, as shown in FIG. 3, the specific period selling goods information extracting means 44 a is a predetermined item among a plurality of goods sold at a specific shop among the plurality of shops. Each group can be extracted.

  In this embodiment, since the product is clothing, the product group is “one manufacturer's model number with one item name of one brand of clothing”. In addition to this, the product group is a product such as “one manufacturer's model number with one item name of one brand of miscellaneous goods” or “one product part number with product classification of one manufacturer of grocery products”. It is possible to extract the selling item information for a specific period according to the type of the item.

4). Specific-period selling-sale product purchaser information generation process The specified-period selling-sell product purchaser information management unit 44b in the specified-period selling product purchaser information management unit 44 of the server device 40 is, for example, described above according to the conditions exemplified below. The customer who purchased the product related to the selling product information for the specified period in the second predetermined period in the past is identified as the selling product purchaser for the specified period that is the past selling product purchaser, and the selling for the specified period The product purchaser information for a specific period, which is information related to the product purchaser, is generated.

<Conditions for past purchasers'certification>
Approval month: Approval date of each month: Extraction range of customers required for approval on the first Monday (second predetermined period): Applicable for the first day of the "current month-4" month to the last day of the "current month-1" month: explained above For the first predetermined period, the manufacturer model number of the item within the top 10 sales by brand-item name for each store on the first day of the month “current month-1” to the last day of the month “current month-1” Period of certification as a customer-purchased product purchaser for a specific period of time purchased within a specified period: one year from the latest certification month

  In this embodiment, the “product” that predicts the best selling product is used as clothing. However, in the sale of clothing, the sales period is considered to be about 4 months at the longest. Therefore, in this embodiment, when generating the selling item purchaser information for a specific period in the past, the four months from “the first day of the“ current month-4 ”month to the last day of the“ current month-1 ”month, The second predetermined period, which is the survey range of the customer required for the certification, is set (FIG. 8).

  In other words, in this embodiment, the selling products in the last one month in the past (the selling products in the latest one month extracted by the processing operation performed by the specific period selling product information extraction means 44a) are changed to the past four months in the past. Customers who have purchased in the past, that is, customers who have purchased the best selling products in the past month in the past month, are also extracted as the best selling products purchasers for a specific period.

  In addition, if the “product” that predicts a popular product is a product other than clothing and the sale period of the product is considered to be about 6 months at the longest, sales in a specific period in the past When generating the product purchaser information, the second predetermined range, which is the survey range of customers required for certification, is from the first day of the “current month-6” month to the last day of the “current month-1” month. Considering the sales period that is considered to be the longest of the “product” that predicts the selling product, such as a period, etc., the second predetermined period that is the customer survey range required for past selling product purchaser certification Can be set.

  As will be described later, the information on the purchase of merchandise for a specific period that is generated in this step is information for predicting a future best-selling product. Therefore, it is desirable that the information be capable of enhancing the reliability of future hot selling product prediction. For this reason, it is desirable that the specific period selling item purchaser information generated in this process is information relating to a purchaser who has a record of purchasing a previously sold item. Therefore, the specified period selling goods product purchaser information generated in this process includes the “first predetermined period” in which the selling products are authorized, that is, the above-mentioned specified period selling goods information is generated. In addition, it is desirable that the information is related to a product purchaser who has purchased a popular product in the “first predetermined period” in a period preceding the “first predetermined period”.

  Therefore, in relation to the “first predetermined period”, the “second predetermined period” includes at least the “first predetermined period” and at least from the “first predetermined period”. A long period is desirable.

  In this embodiment, the product demand prediction is performed for each predetermined product group among a plurality of products sold at the store. Therefore, information that is necessary to certify that it is a specific product group, for example, product brands, item names, etc., are collected as “information for specifying products”, and past purchases of popular products The authorization of the person also certifies the past purchaser of the merchandise for a specific product group.

  For example, past selling merchandise purchasers are certified for each brand-item name which is “information for identifying a product”.

  In addition, when forecasting product demand for a specific product group in a specific store, as described above, information on the store that sells the product included in the “information specifying the product” Referring to each specific store, a past selling item purchaser is certified for each brand-item name which is “information for specifying a product”.

  In accordance with the conditions in this embodiment exemplified above, the specific period selling goods purchaser information generation means 44b obtains the number of sales by brand-item name for each store in September 2014 extracted as described above. Customers who purchased the manufacturer model number of the top 10 items are extracted from the purchase history information DB 43 in the range from June 1, 2014 to the end of September 2014, and on the first Monday of October 6, 2014 The extracted customer is recognized as a purchaser of a sale item for a specific period that is a purchaser of a sale item in the past.

  The specific-period selling product purchaser information generating unit 44b generates the specified-period selling product purchaser information, which is information related to the specified-period selling product purchaser, after identifying the specified-period selling product purchaser.

  The specific period selling product purchaser information generation means 44b generates a plurality of continuous first predetermined periods in the past (specific period selling item information should be generated under the conditions in this embodiment exemplified above). In the “first predetermined period” in which purchase history information is aggregated (from the first day of the “current month-1” month to the last month of the “current month-1” month), the product purchaser information for the specific period is generated. Perform the process.

  Then, the selling item purchaser information for the specific period generated in the latest first predetermined period is stored in the selling item purchaser information DB 45.

  At this time, among the plurality of specified period selling product purchaser information in the plurality of first predetermined periods stored in the selling product purchaser information DB 45, the specified period selling product purchaser information of the oldest period is stored. It is possible to perform the process of deleting.

  The specific period selling item purchaser information is used for information for predicting a future selling item as will be described later. Therefore, it is desirable to use the selling merchandise purchaser information in a period closer to the time when the prediction is performed.

  Thus, when storing the selling item purchaser information for the specific period generated in the most recent first predetermined period in the selling item purchaser information DB 45, it is stored in the selling item purchaser information DB 45 at the same time. Among the plurality of specified period selling item merchandise purchaser information in the first predetermined period, the specified period selling item merchandise purchaser information of the oldest time is deleted.

  In the condition of this embodiment exemplified above, the period for qualifying as a purchaser for a certain period selling goods is set to “one year from the latest certification month”.

  Therefore, when the above-exemplified conditions in this embodiment are stored in the selling product purchaser information DB 45 for the specific period selling product purchaser information generated in the most recent first predetermined period (one month). At the same time, among the plurality of specified period selling item purchaser information in the first predetermined period stored in the selling item purchaser information DB 45, the specified period selling item purchaser information of the oldest period That is, the customer who has been identified as the purchaser of the sale item for the specific period 13 months ago, and the customer who is once recognized as the purchaser of the sale item for the specific period, then purchases the sale item for the specific period for 12 months (one year). If the authorization period for 12 months (one year) expires, it will continue to be authorized as a buyer and will be sold for a specific period stored in the selling product purchaser information DB 45. Delete from user information We are.

  As a result, a customer who has been certified as a selling product purchaser for a specific period in a specific month in the past and stored in the selling product purchaser information DB 45 is again in a specific month before 12 months have passed. If it is certified as a purchaser for a certain period of time, it will continue to be certified as a purchaser of a sale item for a specified period for 12 months (one year) and stored in the sale item purchaser information DB 45. However, once a customer has been identified as a seller for a specified period of time, but has not been certified as a buyer for a specified period of time within the next 12 months (one year) When the authorization period of 12 months (one year) from the first authorization month expires, information is deleted from the sale item purchaser information for a specific period stored in the sale item purchaser information DB 45.

  In this embodiment, the item for which the best selling is predicted is “clothing”, and the product purchaser (customer) once certified as the best selling product purchaser for a specific period is recorded / saved in the best selling product purchaser information DB 45. The stored period is set to 12 consecutive first predetermined periods (that is, one year), but this period can be changed in various ways in consideration of the characteristics of the commodity that predicts the best selling. . For example, it can be variously determined at a predetermined predetermined number of first predetermined periods.

  FIG. 4 is a diagram for explaining the certification of a specific period selling goods purchaser who is a past selling goods purchaser and the certification period. In this figure, it is as follows.

  -Since customer A purchased the selling goods in December 2012, which is the first predetermined period, during the period from September 1, 2012 to the end of December 2012, which is the second predetermined period. On the first Monday of January 2013, it is certified as a seller for a certain period of selling goods. The certification period is from January 2013, which is the certification month, to the end of December 2013.

  -Since customer B purchased a selling product in December 2012, which is the first predetermined period, during the period from September 1, 2012 to the end of December 2012, which is the second predetermined period. On the first Monday of January 2013, it is certified as a seller for a certain period of selling goods. In addition, since the products sold in May 2013, the first predetermined period, were purchased during the period from February 1, 2013 to the end of May 2013, the second predetermined period. On the first Monday of the month, it is re-certified as a purchaser for a certain period. Therefore, the period during which customer B is authorized as a product purchaser for a specified period is initially from January 2013 to the end of December 2013. Since the product has been certified as a product purchaser, the period for which the product purchaser has been certified for a specific period is from June 2013, which is the latest certification month, to the end of May 2014, 12 months.

  -Since customer C purchased the selling goods in January 2013, which is the first predetermined period, during the period from October 1, 2012 to the end of January 2013, which is the second predetermined period. On the first Monday of February 2013, the product was certified as a buyer for selling goods for a specific period. After that, since it was never certified as a seller for a certain period selling goods for 12 months, the certification period expired in January 2014. From February 2014, the selling goods purchaser information DB 45 It is not included in the purchase information of merchandise sold for a specific period of time. And since the selling goods in March 2014, which is the first predetermined period, were purchased during the period from December 1, 2013 to the end of March 2014, which is the second predetermined period, On the first Monday of the month, it is certified as a product purchaser for a specified period. In this case, the period for which the product is purchased as a selling product for a specific period is until the end of March 2015.

  Note that, as described above, in this embodiment, the product demand prediction is performed for each predetermined product group among a plurality of products sold in the store, so as to certify that it is a specific product group. The product brand, item name, etc., which are necessary information for the product, are collected as “information for identifying the product”, and the past selling item purchaser's certification is “information for identifying the product”. It is done by name.

  FIG. 5 is an example of a specific-period selling product table representing past selling products included in the specified-period selling product purchaser information. In the example shown in the figure, the manufacturer model number of the product within the top 10 sales by brand-item name for each store in September 2014 is shown.

  FIG. 6 is an example of a specific-period selling product purchaser table representing customers who have purchased past selling products included in the specified-period selling product purchaser information. In the example shown in the figure, the manufacturer model number of the top 10 items sold by brand-item name for each store in September 2014 is purchased during the period from June 1, 2014 to the end of September 2014. On the first Monday of October 6, 2012, customers who have been certified as purchasers for a specific period are shown.

  The specific period selling goods purchaser information generated as described above is stored in the specified period selling goods purchaser information DB 45.

  Since the above-described past selling product information extraction processing and past selling product purchaser authorization processing are performed every first predetermined period (one month) described above, under the conditions of this embodiment, the specified period In the selling merchandise purchaser information DB 45, monthly generated selling merchandise purchaser information is stored and accumulated every month. Under the conditions of this embodiment, when storing the selling product purchaser information for the specified period in the most recent first predetermined period (one month) in the selling product purchaser information DB 45, Among the plurality of specified period selling item purchaser information in the first predetermined period stored in the line item purchaser information DB 45, the oldest specified period selling item purchaser information, that is, 13 months The information on the purchaser of merchandise sold for a specific period is deleted.

  As a result of accumulating the information on the purchase of merchandise for a specific period, the information on the purchase of merchandise for a specific period of time used in the forecast process for a future sale of merchandise will be stored. Generated. The specified period selling item merchandise purchaser accumulation information is obtained by accumulating the generated specific period selling item merchandise purchaser information in a plurality of consecutive first predetermined periods in the past. As shown in FIG. 7, a plurality of specified period selling item purchaser accumulation information in which the specified period selling item purchaser information for the last 12 months including the month in which the prediction process is performed is accumulated.

  As described above, the specific-period selling product purchaser information generating unit 44b generates the specific-period selling product purchaser information for each predetermined product group among the plurality of products sold in the store. Can do.

  Further, as shown in FIG. 5 and FIG. 6, the specific period selling goods product purchaser information generating means 44 b sells the specified period selling goods purchaser information at a specific store among the plurality of stores. It can be generated for each predetermined product group among a plurality of products.

  In this embodiment, since the product is a clothing item, the product group is "one manufacturer model number with one item name of one brand among clothing items". In addition to this, the product group may be, for example, “one manufacturer model number with one item name of one brand of miscellaneous goods”, “one product item number with product classification of one manufacturer of grocery products”, etc. According to the type of the product, it is possible to generate the selling product purchaser information for a specific period for each product group.

5. Predictive processing of future selling products In the selling product prediction unit 46, the latest purchase product information extracting means 46a is used by a plurality of specified period selling product purchasers stored in the specified period selling product purchaser information DB 45. The most recently purchased product information, which is information related to products purchased during the latest third predetermined period, is extracted using the purchase history information stored in the purchase history information DB 43.

  Next, the selling product prediction information generating unit 46b in the selling product prediction unit 46 generates selling product prediction information based on the most recently purchased product information generated as described above.

  This processing is performed for a specific period of time when a purchase activity was performed during the latest third predetermined period among the products sold in the latest third predetermined period included in the most recently purchased product information. Ranks the number of line item purchasers, and sells the products sold in the third predetermined period from the top to the predetermined rank in the number of items sold for a specific period in the future. Is extracted as

  On the condition of this embodiment, a plurality of specified period selling item merchandise purchasers stored in the specified period selling item merchandise purchaser information DB 45 certify the selling item on the first day of “current month-1” month to “ For the first predetermined period (one month) in the past, the last day of the current month -1 ”month, the products that sold a lot are sold for the second predetermined period (the day of the month of“ current month -4 ”to be certified). (4 months, the last day of the "current month-1" month), and a customer who has purchased the products of the past selling goods. Such a particular period selling goods purchaser is extracted every month to make a large group for 12 months, and the customers belonging to this group are the last third predetermined period (for example, the most recent week) The product purchased in the past is used as the latest purchase product information used for future product demand prediction.

  Further, for each of the products sold in the latest third predetermined period included in the latest purchase product information, customers belonging to a large group for the 12 months for which the product purchase act was performed (specific period Products sold in the third predetermined period from the top to the predetermined rank from the highest number of purchasers for a specific period of time. It is what.

  As described above, the purchasers of a certain period of selling goods that are in a large group for 12 months are referred to as the first day of the “current month -1” month to the last day of the “current month -1” month for certifying the sold goods. The products that sold a lot in the first predetermined period (one month) in the last past are sold in the second predetermined period (from the 1st of the “current month-4” month to the month of the current month to the “current month-1” month. It is a customer who has made a purchase during the last four days and has a track record of purchasing past selling items.

  Therefore, by performing such certification processing, future demand forecasts can be made into the product purchase trends of customers who have purchased in advance before the increase in sales of products that actually sold a lot in the past. It becomes possible to predict based on this. Thereby, it becomes possible to perform future demand prediction with higher reliability than the future demand forecast that relies on the intuition and experience of the buyer.

  In this embodiment, the extraction process of the latest purchased product information and the prediction process of the future selling product are performed under the conditions described below.

<Extracting process of latest purchased product information>
Extraction week: Every week Extraction day: Monday Extraction range (third predetermined period): Monday of “this week −1” week to “this week −1” week in the current month in which processing for predicting future selling products is performed Sunday extraction target: Products purchased by a customer who is a plurality of specified period selling item purchasers stored in the specified period selling item purchaser information DB 45 within the third predetermined period

  In this embodiment, the sale item purchaser information for a specific period is generated for each specific product group (for example, by brand-item name).

  Therefore, for each specific product group (for example, by brand-item name), the third predetermined period described above (Monday of “this week −1” week in the current month in which processing for predicting future selling products is performed) The latest purchase product information on Sunday of “this week -1” week) was generated.

  The third predetermined period, which is the extraction range for generating the most recently purchased product information, is the most recent week in the current month in which processing for predicting future selling products is performed for the purpose of improving the reliability of future demand forecasts ( “Monday of“ this week-1 ”week to Sunday of“ this week-1 ”week) (FIG. 8).

  In this way, from the viewpoint of improving the reliability of future demand forecasting, the “third predetermined period” is related to the “first predetermined period” in which the above-mentioned specific period selling product information is generated. In other words, it is the most recent period in the current month in which processing for predicting future selling items is performed, and the length of the period is preferably shorter than the “first predetermined period”.

  In this embodiment, when the most recently purchased product information extracting means 46a generates the most recently purchased product information, there is only a single product of a specific period selling product purchaser who has made a purchase during the third predetermined period described above. The group was not included in the most recently purchased product information in order to increase the reliability of future demand forecasts.

  That is, the latest purchase product information generated by the latest purchase product information extraction means 46a is a product group in which there are at least a plurality of purchasers of specific-period selling products that have made a purchase during the third predetermined period described above. It was made to be included in.

  Therefore, a plurality of purchase actions for a certain product group (brand-item) are performed, but the product group in the case where it is only a plurality of purchase activities by the same one-period selling product buyer for a specific period. In addition, the specific product group in the case where there is a single purchaser of a specific product for a specific period purchasing a specific product group (for example, a specific brand-item) is generated by the latest purchase product information extraction unit 46a. It is not included in the latest purchase product information. This is to increase the reliability of future demand forecasts.

  Further, only the product group within three weeks from the sales start week is included in the latest purchase product information generated by the latest purchase product information extraction means 46a.

  In other words, in this embodiment, the latest purchase product information extraction unit 46a has the plurality of specific-period selling product purchasers stored in the specified-period selling product purchaser information DB 45 in the latest third predetermined period. The purchase history information DB 43 stores the most recently purchased product information that is information related to the product purchased in the fourth period from the sales start week (within three weeks from the sales start week in FIG. 8). Extracted using purchase history information.

  In this embodiment, future demand is predicted for the product “clothing”. Referring to FIG. 8, in clothing, the week in which the number of items sold reaches the peak is the average 4 to 5 weeks from the first sales week. Therefore, from the viewpoint of improving the reliability of future demand prediction, in this embodiment, the near-purchased product information extracting unit 46a stores the most recently purchased product information in the marketed product purchaser information DB 45 for a specific period. Products purchased by a plurality of point-of-sale merchandise for a specific period of time that are purchased during the most recent third predetermined period, and related to products for the fourth predetermined period (within three weeks from the sales start week) from the sales start week The information is extracted using the purchase history information stored in the purchase history information DB 43. In this way, it is possible to know which products are selling before the peak of the number of products sold.

  The period from the first sales week to the peak week of product sales differs for each product for which demand is predicted. Therefore, for each product for which future demand prediction is performed, the fourth predetermined period, which is the period from the first sales week to the week when the number of sales of the product reaches its peak, is taken into consideration in the selling product purchaser information DB 45 for a specific period. The most recently purchased products that are purchased by the purchasers of a plurality of specified periods in the last third predetermined period and are information about the products in the fourth predetermined period from the sales start week Information can be extracted by using the purchase history information stored in the purchase history information DB 43 by the latest purchase product information extraction means 46a.

  As described above, the “third predetermined period” is determined in a range in which the above-described “fourth predetermined period starting from the sales start period of a product for which future demand prediction is performed” is specified. Therefore, from the viewpoint of increasing the reliability of future demand forecasting, the “third predetermined period” is related to the “fourth predetermined period starting from the sales start period of products for which future demand forecasting is performed”. In other words, it is preferable that the period is shorter than the “fourth predetermined period” in the “fourth predetermined period”.

  Further, in relation to the aforementioned “third predetermined period”, the “fourth predetermined period starting from the sales start period of the product for which future demand prediction is performed” is “the third predetermined period”. Average period that is at least a period longer than the “predetermined third period” in the most recent past and reaches the peak of the number of sales of the product from the initial sales period of the product for which future demand forecast is made It is desirable that the period is shorter than the period.

  An example of the process of extracting the latest purchased product information in this embodiment will be described with reference to FIGS.

  In the case of performing the latest purchase product information extraction process in December 2013 shown in FIG. 4, the above-mentioned specific period selling product purchaser accumulation information from January 2013 to December 2013 is used (FIG. 7). Accordingly, customers A, B, and C correspond to customers who are certified as buyers of the selling products for the specific period in the extraction target.

  When the latest purchase product information extraction process is performed in January 2014 shown in FIG. 4, the selling product purchaser accumulation information for the specific period from February 2013 to January 2014 is used (FIG. 7). Therefore, the customers B and C correspond to the customers who are recognized as the purchasers of the selling products for the specific period being extracted.

  When the latest purchase product information extraction process is performed in February 2014 shown in FIG. 4, the selling product purchaser accumulation information for the specific period from March 2013 to February 2014 is used (FIG. 7). Therefore, only the customer B corresponds to the customer who has been recognized as the particular-period selling goods purchaser in the extraction target.

  In accordance with the above conditions, the most recently purchased product information extracting means 46a is, for example, that a plurality of customers who have been identified as specified-period selling product purchaser information DB 45 stored in the specified-period selling product purchaser information DB 45 The manufacturer model number of the product of the same brand as the brand purchased in the week, and the latest purchase product information about the product within three weeks from the sales start week is obtained from the purchase history information DB 43 on the third week of Monday in September 2014. Extract.

  As described above, the latest purchase product information extraction unit 46a can generate the latest purchase product information for each predetermined product group among a plurality of products sold at the store.

  Further, the latest purchase product information extraction unit 46a can generate the latest purchase product information for each predetermined product group among a plurality of products sold at a specific store among the plurality of stores. .

  In this embodiment, since the product is a clothing item, the product group is "one manufacturer model number with one item name of one brand among clothing items". In addition to this, the product group may be, for example, “one manufacturer model number with one item name of one brand of miscellaneous goods”, “one product item number with product classification of one manufacturer of grocery products”, etc. The most recently purchased product information for each product group can be generated according to the type of product.

<Prediction process for future selling products>
The selling merchandise product prediction information generating means 46b, based on the most recently purchased product information, among the products sold in the most recent third predetermined period included in the most recently purchased product information, Ranking the number of purchasers of the specified period selling goods sold during the predetermined period, and ranking the number of the purchasers of the specified period selling goods from the top to the predetermined rank The sale item prediction information is generated by using the sale item for the predetermined period as a future sale item.

  FIG. 9 is an example in which the selling product prediction information generated by the selling product prediction information generation means 46b is displayed on the output screen of the computer.

  In the example shown in the figure, the manufacturer model number of a product within the top 10 sales by brand-item name for each store in February 2014 extracted by the past selling product information extraction process is the past selling product purchase. A plurality of customers who purchased between November 1st, 2013 and the end of February 2014 by the author's authorization process, and were certified as selling products for a specific period, are expected to use the forecast processing for future selling products in 2014. A product of the same brand as the above-mentioned certified brand purchased between March 10 and March 16, and a manufacturer model number within three weeks from the sales start week is a future selling item. This represents the predicted state on Sunday and Monday.

  In FIG. 9, future selling products for the brand A at a specific store X are displayed by “item name—manufacturer model number”.

  In this way, the selling product prediction information generating unit 46b can generate the selling product prediction information for each predetermined product group among a plurality of products sold at the store.

  Further, as shown in FIG. 9, the selling merchandise product prediction information generating unit 46b is configured to store the selling merchandise product prediction information at a predetermined one of a plurality of products sold at a specific store among the plurality of stores. It can be generated for each product group.

  In this embodiment, since the product is a clothing item, the product group is "one manufacturer model number with one item name of one brand among clothing items". In addition to this, the product group may be, for example, “one manufacturer model number with one item name of one brand of miscellaneous goods”, “one product item number with product classification of one manufacturer of grocery products”, etc. It is possible to generate selling product prediction information for each product group according to the type of the product.

  FIG. 10 summarizes the number of sales for the future selling products displayed in FIG. 9 from March 17, 2014 to April 20, 2014 (5 weeks). The status is displayed by item name.

  As shown in FIG. 10, in the item name of the product of brand A at a specific store X that is predicted as a future selling product, among the nine manufacturer model numbers, the product number of 8 models is the highest in sales. Can be confirmed.

  As described above, in this embodiment, before reaching a predetermined period from the sales start week, which is the peak of sales of products (for example, the fourth to fifth weeks from the sales start week), Can be ascertained before the peak period arrives (for example, within 3 weeks from the sales start week of the product). Thus, buyers and the like can purchase goods that can become a selling point with higher reliability, reduce defective inventory of goods, and improve sales.

  And in this embodiment, paying attention to the customer who purchased the past selling product extracted by the extraction processing of the past selling product information, the customer is identified as the past selling product purchaser, and the future Reliability of products purchased by the past buyers of the previously sold products during the most recent specified period through the selling product prediction process, that is, “new products that have been purchased by customers who have purchased past products of sale” In addition, it uses information that is very likely to become a selling point to predict future selling items. Therefore, it is possible to provide a highly reliable product demand forecasting system that is different from the information for forecasting popular products such as the number of sales of past simple products, temperature, weather, and trend information.

  The preferred embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to such embodiments, and various forms are possible within the technical scope grasped from the description of the claims. Can be changed.

DESCRIPTION OF SYMBOLS 10 Product demand prediction system 20 Client terminal 30 Shop side terminal 40 Server apparatus 41 Information transmission / reception means 42 Purchase history information management part 42a Purchase history information storage means 43 Purchase history information DB
44 Specific period selling goods product purchaser information management part 44a Specific period selling goods information extraction means 44b Specific period selling goods purchaser information generation means 45 Specific period selling goods purchaser information DB
46 selling product prediction unit 46a latest purchase product information extracting means 46b selling product prediction information generating means 50 communication network

Claims (4)

Purchasing history information storage means for storing, for a plurality of products sold in a store, information for identifying a customer who has purchased the product and sales history information in which sales information of the product by the customer purchasing the product is associated. When,
Based on the purchase history information, the number of sales of the plurality of products sold at the store in the first predetermined period in the past is ranked, and the products from the top to the predetermined ranking Specific period selling product information extraction means for extracting as a specified period selling product information that is past selling product information,
The customer who purchased the product related to the selling product for the specified period in the second predetermined period in the past is recognized as the selling product purchaser for the specified period which is the past selling product purchaser, and sold for the specified period. Specific period selling goods purchaser information, which is information relating to the line item purchaser, is generated, and the specified period selling goods purchaser information is stored in the specified period selling goods purchaser information DB. Generating means;
The purchase information on the latest purchase product information, which is information related to the product purchased by the plurality of sale products for the specified period sold in the specific period sale product purchaser information DB, in the latest third predetermined period. Latest purchase product information extraction means to extract from history information,
Based on the latest purchase product information, among the products sold in the latest third predetermined period included in the latest purchase product information, a purchase act is performed during the latest third predetermined period. Ranking the number of the purchasers of the selling products for the specified period that were made, and the number of selling products for the specified period selling products was sold in the third predetermined period from the top to the predetermined rank A commodity demand forecasting system comprising a commodity sales forecast information generating means for generating commodity sales forecast information by using a commodity as a future commodity. The selling product forecast information is based on the specified period selling product information for each predetermined product group among the plurality of products sold in the store, and the specified period selling product information for each predetermined product group. Of a plurality of products sold at the store, which are generated based on the specific period selling goods purchaser information for each predetermined product group generated and the most recently purchased product information for each predetermined product group. The commodity demand forecasting system according to claim 1, wherein the commodity demand forecasting information is for each of the predetermined commodity groups. The point-of-sales product prediction information is specified at the specific store generated based on the point-of-sale item information at the specific store in the plurality of stores and the point-of-sale product information at the specific store. The selling product prediction information at a specific store among the plurality of stores, which is generated based on the period selling product purchaser information and the latest purchase product information at the specific store. The product demand prediction system according to 1 or 2. The specified period selling product purchaser information generating means performs a process of generating the specified period selling product purchaser information in each of the plurality of consecutive first predetermined periods in the past. When storing the selling product purchaser information for the specific period generated in one predetermined period in the selling product purchaser information DB, a plurality of the first selling items stored in the selling product purchaser information DB are stored. 4. The process of deleting the specified period selling product purchaser information of the oldest period in the specified period selling product purchaser information in one predetermined period. The commodity demand forecasting system according to claim 1.