JP2007233535A - Classification association method, classification association system and classification association program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable a retail dealer to easily utilize market data for analyzing market trend by rationally and highly precisely associating two classification systems different in classification in classifying articles into a plurality of categories by using a computer. <P>SOLUTION: In a method for associating two classification systems different in classification in classifying articles into a plurality of categories, a computer acquires information relating to the categories of the both classification systems, and finds out the category of 1 in one classification system and the category of 1 in the other classification system where the same articles are classified based on the acquired information, and decides whether or not the duplication rate of the articles classified into the both categories of 1 found out as mentioned above is larger than a duplication rate for decision (step S64), and when the duplication rate is larger than that for decision as the result of decision, the categories of 1 are associated and stored in a table (step S65). <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a method, system, and program for associating two classification systems having different ways of classifying articles such as products, and belongs to the technical field of information processing by a computer.

  In general, retailers such as home centers, supermarkets, drug stores, etc., sell products that take into account their sales policies, consumer groups, seasonal factors, etc., among the many products distributed in the market. It is important to sell assorted products, and as an aid to that, we obtain market data from market research companies, etc., to investigate what is selling well and what is not selling well in the entire market. Therefore, it is important to analyze the sales trends of all products in the market.

  In this regard, for example, Patent Document 1 discloses the sales quantity data of products acquired from a retailer's POS (Point of Sales) terminal and the received quantity data of products acquired from a warehouse terminal. In addition, a system is disclosed in which a supplier of a product such as a manufacturer analyzes a market trend quickly and easily from the stock status of the product in the entire market. The system can be used not only for inventory management but also as a system for capturing sales trends of products in the entire market.

  Further, in Patent Document 2, a merchandise supplier computer, an information provider computer, an operation processing agent computer, a retailer POS terminal and a computer are connected via a communication network to a retailer. Market research system configured so that the operation processing agent analyzes the sales of the product based on the sales information transmitted from the POS terminal, and transmits the result to the product supplier, the information provider and the retailer. Is disclosed. With this system, retailers can grasp market trends and product information suitable for their region in a timely and accurate manner, and devise accurate sales strategies. In addition, in the product master used in this research system, JAN (Japan Article Number) code, JICFS (JAN Item Code File Service) classification code, etc. are registered in addition to the product name and manufacturer name. Yes.

Japanese Patent Laying-Open No. 2005-008341 (paragraph 0009, paragraph 0041) JP 2002-298227 A (paragraph 0007, paragraph 041, paragraph 0108)

  By the way, as described above, each retailer has different types of products and display forms according to their sales policies and consumer groups. It is customary to have a system. For example, drugstores handle relatively large amounts of shampoos and rinses, and home centers have relatively small amounts. For drugstores, shampoos and rinses are classified into, for example, the major category “Daily Goods”. The categories “for women”, “for men”, “for adults”, and “for children” are subdivided into categories, but at home centers, they are grouped into large categories such as “bathrooms”. . In this way, even if it is the same product, it will be classified into categories with different names and hierarchies depending on the retailer, which is familiar to the company's own classification system for each retailer employee There is no inconvenience.

  However, the above-mentioned market data that retailers obtain from market research companies in an attempt to analyze the sales trends of all products in the market is a classification system that is conventionally adopted by the market research company that provides the market data. Or, according to the JICFS classification system, there is a problem that even if retailers look at the market data at a glance, it is different from their own classification system, so it is difficult to immediately use the market data for actual work. .

  In other words, even if the product is handled by the company and has a sales record, if the classification system is different, the product name, JAN code, etc. can be searched for in the market data. It becomes very difficult. In addition, if the product is not handled by the company and has no sales record, the product name, JAN code, etc. are unknown, so if the classification system is different, the product will be displayed even if it is included in the market data. I don't know how the product is related to the products sold by the company, but useful information (for example, information that sales of similar products classified in the same category as the products sold by the company are growing) ) May not be used at all.

  The present invention has been made in view of the above-mentioned problems in retailers who have their own product classification system according to their own assortment products. It is an object to associate two classification systems with a reasonably high accuracy, thereby enabling, for example, a retailer to easily use market data for analyzing market trends.

  In order to solve the above-mentioned problems, the present invention is configured as follows.

  First, the invention according to claim 1 of the present application is a method for associating two classification systems having different classification methods when an article is classified into a plurality of categories, in which the computer Information on the classification of the classification, and based on the obtained information, find one classification of one classification system and one classification of the other classification system in which the same article is classified, It is determined whether or not the duplication rate of articles classified in the category is larger than the duplication rate for determination, and if the duplication rate is larger than the duplication rate for judgment as a result of this determination, the categories of 1 are associated with each other. And storing it in a table.

  Next, the invention according to claim 2 is the classification corresponding method according to claim 1, wherein the computer is configured so that the same article is classified into one category of one classification system and one of the other classification system. When a category cannot be found or, as a result of the determination, the duplication rate is not greater than the duplication rate for determination, at least one of the classification of one classification system and the classification of the other classification system is classified as a lower hierarchy or It is characterized in that a change is made to a higher-level category, and for the changed lower-level category or higher-level category, a re-category heading or re-duplication rate determination is performed.

  Next, the invention according to claim 3 is the classification correspondence method according to claim 1 or 2, wherein the one classification system is a classification system created by a retailer for the handling product, and the other classification system is The classification system is a classification system created by the market research company for the survey target product including the products handled by the retailer, and the computer stores information on the sales performance of each product of the retailer and the product of the market research company for each product. Based on the association between the categories obtained from the survey results and stored in the table, the retailers do not sell, but the products sold in the market survey are classified into the categories of the retailers classification system. In the state of being classified and classified into this retailer's classification system, the survey results of the products sold by the retailer, and the retailer does not sell but are sold in the market research It characterized in that it created a survey report of retailers described in conjunction with the findings of the goods that.

  On the other hand, the invention according to claim 4 is a system for associating two classification systems having different classification methods when classifying an article into a plurality of classifications, wherein information on the classifications of both classification systems is obtained. An information acquisition means to acquire, and a search means for finding one classification of one classification system and one classification of the other classification system in which the same article is classified based on the information acquired by the information acquisition means; A determination unit that determines whether or not the duplication rate of the articles classified into both one category found by the search unit is larger than the duplication rate for determination, and the duplication rate as a result of the determination by the determination unit Is larger than the determination duplication rate, the apparatus further comprises classification correspondence table storage means for associating and storing the one category in a table.

  Next, the invention according to claim 5 is the classification corresponding system according to claim 4, wherein one category of one classification system and one of the other classification system in which the same article is classified by the search means. When the category is not found or, as a result of the judgment by the judging means, the duplication rate is not larger than the judgment duplication rate, at least one of the classification system and the other classification system is subordinated. Further comprising a hierarchy changing means for changing to a higher hierarchy classification or a higher hierarchy classification, and for the lower hierarchy classification or higher hierarchy classification changed by the hierarchy changing means, the search means re-categorized Or the determination means determines the overlap rate again.

  Next, the invention according to claim 6 is the classification corresponding system according to claim 4 or 5, wherein the one classification system is a classification system created by a retailer for the handling product, and the other classification system is The classification system is a classification system created by the market research company for the product to be surveyed including the products handled by the retailer, and the information acquisition means includes information on the sales performance of each product of the retailer and the market research company Based on the association between one category stored in the table by the classification correspondence table storage means, the retailer does not sell the information on the survey results for each product, but is sold in the market research. Classifiers that classify existing products into categories of the retailer classification system, and products that are sold by the retailer in a state that is classified into the categories of the retailer classification system by this classification means It further comprises a survey report creation means for creating a survey report for retailers that includes the survey results and the survey results of products not sold by the retailer but sold in the market survey. Features.

  In addition, each means of the classification corresponding | compatible system of the said Claims 4-6 is a hardware resource, such as a central processing unit of a computer, an input device, an output device, a memory | storage device, and each means using these hardware resources. And a program for realizing the operation.

  Furthermore, the invention according to claim 7 is a program for associating two classification systems having different classification methods when classifying an article into a plurality of classifications, wherein the computer is classified into both classification systems. Information acquisition means for acquiring information on, search means for finding one classification of one classification system and one classification of the other classification system in which the same article is classified based on the information acquired by the information acquisition means; The determination means for determining whether or not the duplication rate of the articles classified into both one category found by the search means is larger than the duplication rate for determination, and as a result of the determination by the determination means, the duplication When the rate is larger than the determination duplication rate, it is characterized by functioning as a classification correspondence table storage means for associating the one category with each other and storing them in a table.

  With the above configuration, according to the invention described in each claim of the present application, the following effects can be obtained.

  First, according to the classification correspondence method according to claim 1, for example, when trying to associate two classification systems having different classification methods when classifying articles such as products into a plurality of categories, First, information on the classifications of both classification systems (including information on classification names and articles classified into classifications, etc.) is acquired, and then the same articles are classified based on the acquired information. One category of one classification system and one category of the other classification system are found, and then the articles classified into one category of the one classification system found and the other It is determined whether or not the overlap rate with an item classified in one category of the classification system (ie, the rate at which the same item is classified) exceeds the determination overlap rate (for example, 90%). As a result of the judgment, the duplication rate When the determination overlap rate is exceeded, one classification of the one classification system and one classification of the other classification system are stored in the form of a table in association with each other, so the classification method is different. The two classification systems are associated with each other with a reasonably high accuracy by the computer (this accuracy depends on the determination duplication rate).

  This will be described in more detail with reference to the drawings. For example, as shown in FIG. 43, in one classification system 1, there is a middle classification category “b” below the major classification category “a”, and lower classification categories “c” and “d” are lower layers. It is assumed that product 1 and product 2 are classified in category “c”, and product 3, product 4, product 5 and product 6 are classified in category “d”. In the classification system 2, the middle category “f” is below the major category “e”, the lower categories are “g” and “h”, and the category “g” It is assumed that the product 1 and the product 2 are classified, and the product 5, the product 11 and the product 12 are classified in the category “h”. That is, FIG. 43 includes information on the classifications of both classification systems 1 and 2 (classification names “a” to “h”, information on articles classified into the categories “product 1”, “product 2”, etc.). And based on the acquired information, the state after having already found one division of one classification system 1 and one division of the other classification system 2 in which the same product is classified mutually. Show.

  In this case, when comparing the classification “c” of the classification system 1 and the classification “g” of the classification system 2, the same product 1 and product 2 are classified. Accordingly, the overlap rate is 100%. Therefore, it can be said that the classification “c” of the classification system 1 and the classification “g” of the classification system 2 have a high correspondence with each other. Therefore, for example, if a new product x is classified into the category “g” of the classification system 2, it can be inferred with high accuracy if the new product x is classified into the category “c” in the classification system 1. (Refer to claim 3 for this point).

  On the other hand, comparing the classification “d” of the classification system 1 and the classification “h” of the classification system 2, only the product 5 is the same product, and all other products are classified differently. However, the overlap rate is very low when viewed from the classification system 2 (for example, the overlap rate viewed from the classification system 1 is 25%, and the overlap rate viewed from the classification system 2 is 33%). It can be said that the category “d” and the category “h” of the classification system 2 do not have a high correspondence with each other. In this case, the simplest is to look for the classification of the small classification of the classification system 2 that includes as many products 3, 4 and 6 as possible from the classification system 1, and from the classification system 2, What is necessary is just to search the classification | category of the small classification of the classification system 1 containing as many goods 11 and goods 12 as possible (refer the following Claim 2).

  In that case, according to the method of claim 2, when the computer cannot find one section of one classification system and one classification of the other classification system in which the same article is classified, or As a result of the determination, when the overlap rate does not exceed the determination overlap rate, for example, automatically by a computer or according to an external operation using an input device such as a keyboard or a mouse, The category and / or category of the other classification system is changed to a lower hierarchy category or a higher hierarchy category, and the changed lower hierarchy category or higher hierarchy category is changed to the category header or again Since the overlap rate is determined, the above-mentioned determination overlap rate is satisfied by variously changing the category to a broad range (such as a large category) or a narrow range (such as a small category). Correspondence between the segments to each other so that the are always found by the computer. Or, by changing the determination overlap rate to a low value during the trial and error of changing the category, the correspondence relationship between the segments satisfying the determination overlap rate can be found early by the computer. .

  In this case, according to the method of claim 3, the computer is not classified in one classification system, but is classified in the other classification system, based on the correspondence between the one classification stored in the table. The classified goods are classified into one classification system, so if a computer classifies an unclassified article in one classification system, it will belong if it is classified in that classification system. Are classified with high accuracy (this accuracy also depends on the determination duplication rate).

  Further, according to the method described in claim 3, the operation of the computer is made more concrete, and the classification system created by the retailer for the handling product, and the market research company sets the survey target product (the retailer handling product). Information related to the sales performance of each product of the retailer and information about the survey result of each product of the market research company, and table format Based on the association between the categories stored in step 1, the products that are not sold by the retailer but are sold in the market research are classified into categories of the retailer's classification system, and the retailer's classification system Combined with the survey results of products sold by retailers in the category of, and the survey results of products not sold by retailers but sold in market research We will create a survey report (or market data) for the retailers listed, and based on this report, retailers will be able to analyze market trends with their own and unique classification system they manage, It becomes easy to immediately use this research report for work.

  As a result, the following various effects are derived. (1) Products that are not sold in-house but sell well in the market can be easily identified, which is a great reference for the company's future product lineup. (2) It is easy to find products that are not selling well but are selling well in the market, and it is a great reference for investigating the cause of not selling well and examining future countermeasures. (3) Products that sell well in the company but not so well in the market can be easily understood, and can be a great reference for considering whether to continue or reduce sales in the future. (4) It is easy to understand the life cycle of products sold both in-house and in the market, and if the product is in its growth period, it will actively promote it, but if it is in decline, it will not do it. It will be a great reference for judgment. (5) It is easy to identify products that are not selling either in the company or in the market, and it will be a great reference for the decision that the company will withdraw from sales. (6) Comparison of average selling price and sales becomes easy, and comparing and analyzing the average selling price and sales of the market with the average selling price and sales of the company is a great reference for future pricing of the company. (7) Since the classification system of the survey report is the same as the classification system used in other sales analysis and market analysis conducted separately by the company, it should be used easily in synergy with other analysis results. Can do.

  On the other hand, according to the classification correspondence system according to claim 4, the information acquisition unit performs the information acquisition operation of the computer in claim 1, the search unit performs the category heading operation, and determines the duplication rate determination operation. Since the classification correspondence table storage means performs the storage operation in the table format, the same effect as in the first aspect can be obtained.

  Further, according to the classification corresponding system described in claim 5, since the hierarchy changing means performs the hierarchy changing operation of the computer in claim 2, the same effect as in claim 2 can be obtained.

  Further, according to the classification corresponding system of claim 6, the classification means performs the computer classification operation in claim 3, and the investigation report creation means performs the investigation report creation operation. The same effect as that of No. 3 is obtained.

  And according to the classification corresponding | compatible program of Claim 7, since the classification corresponding | compatible system similar to the system of the said Claim 4 is comprised by installing this program in a computer, the said Claim 4 and claim | claim The same effect as item 1 will be obtained.

  Hereinafter, a classification correspondence system according to an embodiment of the present invention will be described. In addition, this classification corresponding | compatible system implements the classification | category corresponding | compatible method concerning this invention, and the program which comprises this classification | category corresponding | compatible system with a computer comprises embodiment of the classification | category corresponding | compatible program based on this invention. .

  FIG. 1 shows an overview of a business model in which a classification support system according to the present invention is introduced. This business model provides market data from market research firms to retailers, taking into account their sales policies, consumer groups, seasonal factors, etc. from the many products distributed in the market. And retailers such as home centers, supermarkets, drug stores, etc., and system users who use the classification support system according to the present invention, what is selling well in the market as a whole? Market research firms that provide market data that investigates whether they are not selling well have joined this business model.

  First, the retailer tabulates the POS data (i) supplied from each store nationwide (transmission and reception of such data is performed via a computer network, various recording media, etc., unless otherwise specified). As long as it is the same), this is stored in the all-store POS database (ii). The retailer selects POS data for market research (for example, POS data for five stores representing the region) (iii) from this database (ii), and prompts the system user at appropriate intervals (for example, for one quarter). Supply)

  The system user converts the format of the market research POS data (i i i) for the market research company (operation I) and supplies it to the market research company. The market research company creates market data by using the POS data for market research (iii) whose format has been converted and the POS data supplied from other retailers in the same manner (operation II), Accumulate in the market data database (iv). Then, the market research company supplies the market data (iv) to other customers and the system user on demand.

  On the other hand, the retailer selects the POS data for survey report (for example, POS data of all stores with sales results of products) (v) from the all store POS database (ii) and supplies it to the system user. Then, the system user uses the market data (iv) from the market research company and the POS data (v) for the research report from the retailer for the retailers who have finished the classification of the products. A survey report X is created (operation III) and supplied to the retailer (in this case, a printed matter of the survey report X may be sent by mail).

  FIG. 2 shows the overall configuration of the classification corresponding system 10 according to the present embodiment. The classification correspondence system 10 is configured with a computer as a center, and a central processing unit 11 that causes the computer to execute predetermined processing according to a program, an input device 12 such as a keyboard and a mouse, and an output such as a display device and a printing device. A storage device 14 having a device 13, a program storage unit 14 a and a data storage unit 14 b for storing programs, various data files, tables and work tables for causing the computer to execute predetermined processing; It has a communication device 15 that transmits and receives information via a network with a computer of a commercial company and a computer of a market research company.

  FIG. 3 shows the procedure of the survey report creation operation (III) performed by the classification corresponding system 10 in the system user of FIG. First, the classification corresponding system 10 writes the POS data (v) for the survey report from the retailer in the POS data file. Here, as shown in FIG. 17, the POS data file records various data such as a department name, a product name, a product name, a JAN code number, and a product name. Here, the name of each item, the way of classifying products, etc. are determined and operated by retailers independently.

  Next, the classification corresponding system 10 converts the POS data (v) into a market data format (sub-operation III-1), and writes it into the converted POS data file. Here, as described below, the converted POS data file includes a list of items, names of items, and the like in the market data (iv) so that it can be easily compared with the market data (iv) from the market research company. As shown in FIG. 18, compared with FIG. 17, for example, the JAN code number and the product name (renamed from the product name) are located at the top, and the department name is classified as a retail large category. The product name is renamed to the retail category, and the product name is renamed to the retail small category.

  On the other hand, the classification corresponding system 10 writes the market data (iv) from the market research company to the market data file. Here, as shown in FIG. 19, the market data file records various data such as the week start date of the survey, the market classification, the product name, the JAN code number, the manufacturer name, and the attribute. Here, the name of each item, the way of classifying products, etc. are determined and operated by market research companies.

  Next, the classification corresponding system 10 edits the market data (iv) during the aggregation period of the POS data (sub operation III-2), and writes this in the edited market data file. Here, as will be described below, the edited market data file sets the survey target period to the POS data (v) aggregation period (for example, one quarter) so that it can be easily compared with the POS data (v) from the retailer. As shown in FIG. 20, compared with FIG. 19, the week start date of the first survey is changed to the period. Here, an example of a calendar table used for this editing is shown in FIG.

  Then, using the converted POS data file and the edited market data file, the market data matched with the JAN code is combined with all the POS data (sub operation III-3). Then write it to the sales data file. Similarly, using the converted POS data file and the edited market data file, data that does not match the POS data in the JAN code is extracted for all market data (sub-operation III- 4). It is then written to a non-sales data file.

  Here, as shown in FIGS. 21 and 22, the sales data file is converted from the converted POS data file shown in FIG. 18 and the edited market data file shown in FIG. A POS data file matched with a JAN code (indicated by “*” in FIG. 20) is divided into a first half part and a second half part. In this case, duplicate JAN code numbers and Items such as product names or periods are grouped together.

  Further, as shown in FIG. 23, the non-sale data file is a file in which the JAN code does not match the converted POS data file shown in FIG. 18 from the edited market data file shown in FIG. 20 (FIG. 20). In which “*” is not attached).

  Then, using the sales data file and the non-sales data file, a survey report is created in accordance with the POS data product classification system (that is, the retailer product classification system) (sub-operation III-5). Then, it is written in the survey report data file. The survey report X is called from the survey report data file and is printed out by the output device 13 (see FIG. 2).

  Here, as shown in FIGS. 24 to 26, the survey report data file is added to the sales data file shown in FIGS. 21 and 22, and the non-sales data file shown in FIG. (For the sake of understanding, “◯” is added to the part derived from the sales data file, and “◎” is added to the part derived from the non-sales data file for the sake of understanding.) In that case, the non-sale data file in the lower part is market data for products that are not sold by the retailer but sold in the market, so that the product is still classified in the classification system of the retailer. It is not. However, as shown in the figure, in the survey report data file, the non-sale products are also classified according to the retailer classification system. This enables retailers to analyze market trends based on this research report X using their own unique classification system managed by the company, and it is easy to immediately use this research report X in the field. Become.

  Next, the classification correspondence table creation operation (IV) used for the survey report creation operation (III-5) will be described. This operation (IV) is also performed by the classification corresponding system 10 in this system user.

  As shown in FIG. 4, first, a matching table is created based on the JAN code (or JICFS classification) using the converted POS data file and the edited market data file (sub-operation IV-1). Here, as shown in FIG. 29, the matching table records various data such as a JAN code number, a product name, a large retail category, a retail category, a retailer category, a market category, a manufacturer name, and an attribute. is there.

  Based on this matching table and using the total number of hits work table and the number of hits work table, the classification hierarchy of the retailer classification system and / or the classification hierarchy of the classification system of the market research company can be varied. By changing, a classification correspondence table is created (sub-operation IV-2). Here, as shown in FIG. 28, the classification correspondence table is assigned a serial number, and is a classification system of retailers, which is a classification of retail large classification (“daily necessities” in the illustrated example) −classification of retail classification ( “Shampoo”, “Rinse”, “Soap”, etc. in the example)-A combination of each category of retail small categories (“Adult”, “Children”, “Body soap”, etc.) and market research Classification of the company's classification system, such as “Shampoo”, “Rinse Conditioner”, “Invasion Treatment”, “Body Shampoo”, etc. in the example)-Attribute categories (“Otonayo”, “Kodomoyo” in the example) , Etc.) are associated with each other.

  FIG. 5 shows the basic concept of the survey report creation operation III. As illustrated, in general, the number of merchandise sold by retailers is less than the number of merchandise sold throughout the market. Therefore, retailers sell products not sold by themselves but sold in the market (non-sale products) as well as sales trends of products sold by the company and sold in the market (sales products). The trend is more worrisome. Therefore, when creating the survey report X, not only the market data of the merchandise sold by the retailer but also the market data of the non-sales merchandise are posted together. A product that is not sold in the market but sold only by a retailer is, for example, a private brand product.

  Next, FIGS. 6 to 10 show the basic concept of the classification correspondence table creation operation IV. For example, as shown in FIG. 6, in the classification system of retailers, there are medium classification categories “rinse” and “shampoo” below the large classification category “daily necessities”, and small classification categories “ Assume that there are “rinse” and “conditioner”, and the product “R2” and the product “B3” are classified into the category “rinse”, and the product “R4” and the product “B4” are classified into the category “conditioner”. Further, there are subcategories “adult” and “children” below “shampoo”, the product “B6”, the product “S1” and the product “S2” are classified into the category “adult”, and the product “S4” It is assumed that the product S5 and the product S7 are classified.

  On the other hand, in the classification system of the market research company, the products R1 to R7 are classified into the market classification category “Rinse Conditioner”, and the products B1 to B7 are classified into the market classification category “Invas Treatment”. It is assumed that the products S1 to S7 are classified into the classification “shampoo”. In other words, FIG. 6 obtains information on the classification of the retailer classification system and information on the classification of the market research company classification system (the names of the categories “daily necessities”, “rinse”, “adult”, “rinse / "Conditioners" ... and information "R2", "B3", "S1" ... etc. regarding products classified into categories), and based on this acquired information, It shows the state after having already found one category of the classification system and one category of the market research company's classification system.

  In this case, any one of the classification categories “rinse”, “conditioner”, “adult”, “child” in the retailer classification system, and any one of the market classification categories in the market research company classification system “ Even if you compare `` Rinse Conditioner '', `` In Bath Treatment '', and `` Shampoo '', the overlapping rate of all the same products being classified is relatively low. The combination of the category and the category of the market research company's classification system cannot be found.

  Therefore, for example, as shown in FIG. 7, when the category “Otonayo” and “Kodomoyo” of the attribute 1 are set below the market classification “shampoo”, the product S1 and the product S2 are classified into the above category “Otonayo”. The product S4, the product S5 and the product S7 are classified into the category “Kodomoyo”, so that the category “adult” of the retailer classification system and the category “Otonayo” of the market research company By comparison, the same product S1 and product S2 are classified, so the overlap rate is relatively high both from the retailer's classification system and the market research company's classification system. It can be said that the classification system category “adult” and the market research company classification system “Otonayo” have a high correspondence with each other (sequential number “1” in the classification correspondence table of FIG. 28). Reference). Similarly, comparing the category “children” of the retailer classification system and the category “Kodomoyo” of the market research company classification, the same product S4, product S5 and product S7 are classified. The overlap rate is relatively high both from the classification system of the market research company and from the classification system of the market research company, so the classification of the classification system of the retailer “child” and the classification of the classification system of the market research company “Kodomoyo” Are in a high correspondence relationship (see serial number “2” in the classification correspondence table in FIG. 28). Therefore, if the attribute 1 of the retailer's non-sale product S3 or S6 is “Otonayo” or “Kodomoyo”, the non-sale product S3 or S6 is classified according to the category “adult” or “ “Children” are classified with high accuracy (in fact, see FIG. 24).

  On the other hand, FIG. 8 is obtained by removing the combination of FIG. 7 from the state of FIG. Again, there remains a state in which the combination of the classification system of the retailers and the classification system of the market research company, which have a high correspondence with each other, cannot be found. Therefore, for example, as shown in FIG. 9, when the classification “rinse” and “conditioner” of the retail small classification are aggregated into the classification “rinse” of the retail classification one level above, the products R2 and R4 are marketed. Since it is classified in the classification category “Rinse Conditioner” and in the aggregated retail category “Rinse”, the classification of the retailer classification system “Rinse” and the market research company Compared with the classification system “Rinse Conditioner”, the same product R2 and product R4 are classified in both categories, so both the retailer classification system and the market research company classification system overlap. The rate is relatively high, so the classification of the retailer classification system “Rinse” and the classification of the market research company classification system “Rinse Conditioner” are highly compatible with each other. It can be said that (see upper serial number of the class mapping table in FIG. 28, "3"). Therefore, the retailer's non-sale products R1 and R3 or R5 to R7 are classified with high accuracy into the classification “rinse” of the retailer's classification system (in fact, refer to FIG. 24).

  FIG. 10 is obtained by removing the combination of FIG. 9 from the state of FIG. Again, there remains a state in which the combination of the classification system of the retailers and the classification system of the market research company, which have a high correspondence with each other, cannot be found. In this case, for example, the classification “rinse” of the retailer classification system and the classification “invas treatment” of the classification system of the market research company may be forcibly associated by lowering the determination duplication rate (FIG. 28). In this case, the retailer's non-sale products B1 and B2 or B5 and B7 are also classified into the classification "Rinse" of the retailer's classification system. (In fact, see FIG. 24).

  Next, more detailed contents of the main sub-operations shown in FIGS. 3 and 4 will be described with reference to flowcharts.

  First, as shown in FIG. 11, the sub-operation III-3 for combining the market data with the POS data in FIG. 3 reads one POS data for all the POS data in FIG. 18 (step S11), and the POS data. It is determined through the JAN code whether the data is also present in the market data of FIG. 20, that is, whether the product is a product sold in the retailer or the market (step S12). If so, the market data is combined with the POS data (step S13), and the record is written in the sales data files of FIGS. 21 to 22 (step S14). However, if NO in step S12, that is, if the product is, for example, a private brand product, there is no market data, so step S13 is skipped and step S14 is executed.

  Further, the sub-operation III-4 for extracting the market data that does not match the POS data in FIG. 3 reads one market data for all the market data in FIG. 20, as shown in FIG. 12 (step S21). ) To determine whether the market data also exists in the POS data of FIG. 18 via the JAN code (step S22). As a result, when the market data does not exist, that is, the product is sold at the retailer. If the product is not sold in the market, the market data record is written in the non-sale data file of FIG. 23 (step S23). However, if YES in step S22 (corresponding to the case of YES in step S12 in FIG. 11), that is, if the product is a product that is sold in both the retailer and the market, step S23 is skipped.

  Further, in FIG. 3, the sub-operation III-5 for creating the survey report by the product classification system of the POS data is the retail corresponding to the serial numbers for all the serial numbers in the classification correspondence table of FIG. 28 as shown in FIG. The records of the sales data files of FIGS. 21 to 22 belonging to the classification are written into the survey report data files of FIGS. 24 to 26 (step S31). Further, the record of the non-sale data file of FIG. 23 belonging to the market classification corresponding to the serial number is written into the survey report data file of FIGS. 24 to 26 (step S32). Here, for the sake of understanding, in the sales data files of FIGS. 21 to 22 and the non-sales data file of FIG. 23, circled numbers are given in the order of sequential numbers written in the survey report data file.

  On the other hand, in the operation IV for creating the classification correspondence table in FIG. 3 (in other words, the operation in FIG. 4), as shown in FIG. 14, the matching table in FIG. 29 is first created (step S41: sub-operation IV in FIG. 4). -1) The hit rate (duplication rate) is calculated for all cases in this matching table, and the correspondence relationship between the categories in which the hit rate exceeds the hit rate condition value (judgment duplication rate) according to the calculation result Write to the classification correspondence table of FIG. 28 (step S42: sub-operation IV-2 of FIG. 4). Then, all the matching table entries that have not yet been written to the classification correspondence table are displayed on the screen (display device as the output device 13) (step S43). This screen display will be described in detail later.

  Then, the sub-operation IV-1 for creating the matching table in FIG. 4 (in other words, step S41 in FIG. 14) reads one POS data for all the POS data in FIG. 18, as shown in FIG. (Step S51) It is determined via the JAN code whether or not the POS data is also present in the market data of FIG. 20, that is, whether or not the product is a product sold in the retailer or the market (Step S52). As a result, if it exists, the POS data and the market data are combined (step S53), and the record is written in the matching table of FIG. 29 (step S54). However, if NO in step S52, that is, if the product is, for example, a private brand product, there is no market data, so steps S53 and 54 are skipped.

  Further, the sub-operation IV-2 for creating the classification correspondence table in FIG. 4 (in other words, step S42 in FIG. 14) first calculates the total number of hits by market classification from the matching table in FIG. 29, as shown in FIG. Calculate and write to the total hit number work table shown in FIG. 37 (step S61). For example, in FIG. 37A, as is apparent from FIG. 29, a total of three products B3, B4, and B6 hit the market classification category “Invas Treatment” (see FIG. 6). “Rinse conditioner” has a total of two products R2 and R4 (see FIG. 6), and the market classification “shampoo” has a total of five products S1, S2, S4, S5 and S7 (see FIG. 6). (See FIG. 6).

  Next, the number of hits for each retail category is calculated from the matching table of FIG. 29, and the ranking for each retail category and the hit number ranking are calculated and written in the hit number work table shown in FIGS. S62). For example, in FIG. 38, as is clear from FIG. 29, one product B3 classified into the market classification “Invas Treatment” is hit in the retail classification “Rinse” − “Rinse” (see FIG. 6). ), One product B4, which is classified in the retail category “rinse”-“conditioner”, and the market category “invas treatment” is hit (see FIG. 6), and the retail category “shampoo” — “adult” ”Hits one product B6 that is classified as“ Invas Treatment ”in the market classification (see FIG. 6), and is classified as“ shampoo ”as the market classification as“ shampoo ”-“ adult ”as the retail classification. The two products S1 and S2 are hit (see FIG. 6) and are classified into the retail category “Shampoo”-“children” and the market category “Shampoo”. One product S4, S5, S7 is hit (see FIG. 6), and one product R2 classified in the market classification category “Rinse Conditioner” is hit in the retail classification category “Rinse”-“Rinse” ( In FIG. 6), one product R4 classified into the market classification “Rinse Conditioner” is hit in the retail classification “rinse”-“conditioner” (see FIG. 6).

  Here, the ranking by retail category is the ranking of the number of hits by market category in the retail category, and the hit number ranking is the hit rate in the market category (calculated in the next step S63). ) Ranking.

  Next, the hit rate is calculated based on the data of the total hit number work table in FIG. 37 and the hit number work table in FIG. 38 to FIG. 41, and is written in the hit number work table (step S63). Here, the hit rate is the ratio of the number of hits to the total number of hits. For example, as can be seen from FIGS. 37 (a) and 38, since the total number of hits in the market classification “Invas Treatment” is 3, the hit rate of the retail classification with 1 hit is 33% respectively. Since the total number of hits in the category “Shampoo” is 5, the hit rate for the retail category with 2 hits is 40%, the hit rate for the retail category with 3 hits is 60%, and the market category “Rinse” -Since the total number of hits for "Conditioner" is 2, the hit rate for the retail category with 1 hit is 50% each.

  Then, it is determined whether or not the hit rate (duplication rate) exceeds the hit rate condition value (determination overlap rate) (step S64). The classification category and the market classification category are associated with each other and written to the classification correspondence table in FIG. 28 (step S65). However, if NO in step S64, that is, if the correspondence between the categories is not high, writing to the classification correspondence table is skipped. Thereafter, the process proceeds to step S43 in FIG. 14, and all the matching table items that have not yet been written to the classification correspondence table are displayed on the screen. Next, this screen display will be described in detail.

  FIG. 30 is an explanatory diagram of a screen displayed on the display device that is the output device 13 in step S43 of FIG. 14, and corresponds to the state of FIG. 6 (screen 1). Since this screen 1 is sorted by the retail classification, when the “market classification” button is pressed, the screen is sorted by the market classification (to screen 2).

  As shown in FIG. 31, on the screen 2, the contents of the total hit number work table shown in FIG. 37A and the contents of the hit number work table shown in FIG. 38 are displayed. Further, the hit number condition value (duplication rate for determination) can be input and specified in a changeable manner using the input device 12 (90% in the example). As described above with reference to FIG. 6, the hit rate (duplication rate) ranges from 33% to 60% at most, and does not reach 90% of the hit number condition value. The combination of the classification system classification and the market research company classification system classification cannot be found. Therefore, “shampoo” is selected from the pull-down menu at the lower right of the screen so that attribute 1 is set below the market classification category “shampoo”, and the “yes” button for attribute 1 is pressed. Then, the “execute” button at the center of the screen is pressed (to screen 3).

  As shown in FIG. 32, on the screen 3, the attributes “Otonayo” and “Kodomoyo” of the attribute 1 are set below the market classification “shampoo”, and as described above with reference to FIG. Market category category "Shampoo"-Attribute 1 category "Otonayo" has 2 hits, Market category "Shampoo"-Attribute 1 category "Kodomoyou" has 3 hits, both The hit rate becomes 100% (see FIG. 37B and FIG. 39). As a result, since the hit rate exceeds 90% of the hit number condition value, the two columns indicating the correspondence between the categories are highlighted and disappear from the screen (to screen 4). At the same time, the correspondence between the categories is written in the classification correspondence table (serial numbers 1 and 2) in FIG. 28 (see step S65 in FIG. 16).

  As shown in FIG. 33, on the screen 4, as described above with reference to FIG. 8, the hit rate is from 33% to 50% at most and does not reach 90% of the hit number condition value. The combination of the classification system of retailers and the classification system of market research companies, which have a high correspondence with each other, cannot be found. Therefore, the retail sub-category “rinse” and “conditioner” buttons are aggregated into the retail sub-category “rinse” of the retail category one level higher than the retail sub-category “rinse” and “conditioner” buttons. Press and press the “Edit subcategory” button. On the small screen that appears, select “Aggregate all minor categories in the same middle category” and press the “Aggregate” button. Then, the “execute” button at the center of the screen is pressed (to screen 5).

  FIG. 34A illustrates a small screen that appears when the “Edit middle category” button is pressed, and FIG. 34B illustrates a small screen that appears when the “Edit large category” button is pressed. Was illustrated.

  As shown in FIG. 35, in the screen 5, the categories “Rinse” and “Conditioner” of the retail small category are aggregated into the category “Rinse” of the retail category one level above, thereby referring to FIG. 9. As described above, the number of hits in the market classification “rinse conditioner” is 2, and the hit rate is 100% (see FIG. 40). As a result, since the hit rate exceeds 90% of the hit number condition value, one column indicating the correspondence between the sections is highlighted and disappears from the screen (to screen 6). At the same time, the correspondence between the categories is written in the classification correspondence table (upper row of serial number 3) in FIG. 28 (see step S65 in FIG. 16).

  As shown in FIG. 36, on the screen 6, as described above with reference to FIG. 10, the hit rate is from 33% to 66% at most and does not reach 90% of the hit number condition value. The combination of the classification system of retailers and the classification system of market research companies, which have a high correspondence with each other, cannot be found. However, after that, the category “Invas Treatment” in the market category is associated with the category “daily necessities”-“rinse” in the retail category, or the category “daily necessities”-“shampoo”-“adult” in the retail category. Since it is a selection, the association between “daily necessities” and “rinse” having a higher hit rate is selected with the “Yes” button, and the “execute” button is pressed (see the correspondence between the classification correspondence tables in FIG. 41 and FIG. 28). (Refer to the lower part of No. 3).

  The above is an example of creating the matching table in FIG. 29 via the JAN code (see the flowchart in FIG. 15), but as described in the sub-operation IV-1 in FIG. 4, via the JICFS classification. A matching table can also be created. That is, if the POS data (v) has a JICFS classification, it is extracted to create a POS data / JICFS classification correspondence table as shown in FIG. Similarly, if there is a JICFS classification in the market data (iv), it is extracted and a market data / JICFS classification correspondence table as shown in FIG. 42B is created. Then, a matching table as shown in FIG. 42 (c) may be created from these tables by mediating the JISFS classification.

  As described above in detail with specific examples, according to the present embodiment, the articles “R1”, “R2”..., “B1”, “B2”..., “S1”, “S2”. When classifying a plurality of categories into two or more categories, a computer (classification) is used to associate two different classification systems (that is, a retailer classification system and a market research company classification system) with each other. The correspondence system 10) first provides information on the classifications of both classification systems (class names “daily necessities”, “rinse”, “adult”, “rinse conditioner”, etc., and information “R2” on articles classified into categories. , “B3”, “S1”, etc. are included) (POS data file in FIG. 17, market data file in FIG. 19), and then the same goods are classified based on this acquired information One minute One category of the system and one category of the other classification system are found (FIG. 6), and then the articles classified into the one category of the one classification system found and the other classification It is determined whether or not the overlap rate (that is, the rate at which the same product is classified) with an item classified in one category of the system exceeds the determination overlap rate (which may be 100%) (see FIG. 16). Step S64) And, as a result of the determination, when the overlap rate exceeds the determination overlap rate, one section of the one classification system and one section of the other classification system are associated with each other. Stored in the table format (classification correspondence table in FIG. 28) (step S65 in FIG. 16), two classification systems having different classification methods can be reasonably highly accurate by the computer (classification correspondence system 10) ( This accuracy is So that the associated at depends on the titration, overlap ratio).

  In that case, when one section of one classification system and one classification of the other classification system in which the same article is classified by the computer (classification correspondence system 10) is not found, or as a result of the determination, When the overlap rate does not exceed the determination overlap rate (FIGS. 6 and 8), for example, automatically by a computer (classification correspondence system 10) or an external operation using the input device 12 such as a keyboard or a mouse Depending on the classification, the classification of one classification system and / or the classification of the other classification system is changed to a lower hierarchy or an upper hierarchy (FIGS. 7 and 9), and the changed lower Since the division heading or the duplication rate determination is performed again for the division of the hierarchy or the division of the higher hierarchy, the division is classified into a wide range (such as a large category) or a narrow range (such as a small category or attribute). By variously changing the sections or the like), so that the correspondence between the segment that satisfy the judgment overlapping rate is always found by a computer (class mapping system 10). Alternatively, by changing the determination overlap rate to a low value during the trial and error of the change of the category, the correspondence relationship between the segments satisfying the determination overlap rate can be improved early by the computer (classification support system 10). Will be found.

  In this case, the computer (classification correspondence system 10) is not classified in one classification system based on the association between one category stored in the table (classification correspondence table in FIG. 28), but the other classification Since articles classified in the system (for example, “R1”, “B1”, “S3”..., Etc.) are classified into one classification system (the survey report data file in FIGS. 24 to 26), the computer By the (classification correspondence system 10), an unclassified article (for example, “R1”, “B1”, “S3”, etc.) in one classification system will belong if it is classified in that one classification system. The classification will be classified with high accuracy (this accuracy also depends on the duplication rate for determination).

  Therefore, when a retailer wants to associate the classification system created for the product handled with the classification system created by the market research company for the surveyed product (including the retailer's handled product), the computer (For example, “R1”, “B1”, “S3”...), Etc., information on the sales performance of each product of the retailer (POS data (v)) and information on the survey results of each product of the market research company (market data) (Iv)) and a product that is not sold by a retailer but is sold in a market research based on the association between one category stored in a table format (classification correspondence table in FIG. 28) Is classified into a category of the retailer's classification system, and in the state of being classified into the category of the retailer's classification system, the survey results of the products sold by the retailer and the retail business Is not sold, but in the market research, a survey report (namely, market data) X for retailers is created together with the survey results of the products sold (the surveys in FIGS. 24 to 26) Based on this survey report X, retailers will be able to analyze market trends with their own familiar classification system managed by the company, and immediately use this survey report X in their operations. It becomes easy.

  The above embodiment is the best embodiment of the present invention, but various modifications can be made without departing from the scope of the claims. For example, in the survey report data file of FIGS. 24 to 26, a part (“◯” mark) derived from the sales data file of FIGS. 21 to 22 and a part derived from the non-sales data file of FIG. ◎ ”) are clearly divided into upper and lower parts, but instead, for example, a part derived from a non-sales data file is mixed into a part derived from a sales data file according to retail classification. (For example, in FIG. 24, the product S3 is described next to the product S2, the product S6 is described next to the product S5, the product R1, the product R3, the product R5 to the product R7, the product B1, the product B2, Product B5 and Product B7 are described next to Product B4, etc.). Further, in the survey report data file of FIGS. 24 to 26, the arrangement of the items, the names of the items, and the like may be returned to the level of the first POS data file of FIG.

  As described above, according to the present invention, two classification systems having different classification methods can be associated with each other with a reasonably high accuracy by using a computer. As a result, for example, from a market research company to a retailer Has a broad industrial applicability to business models that provide market data to

It is explanatory drawing which shows the whole image of the business model which introduced the classification corresponding | compatible system which concerns on embodiment of this invention. It is a block diagram which shows the whole structure of the said classification corresponding | compatible system. It is explanatory drawing which shows the procedure of investigation report creation operation | movement III of FIG. It is explanatory drawing which shows the procedure of the classification | category corresponding | compatible table creation operation IV of FIG. It is explanatory drawing which shows the basic concept of the said investigation report preparation operation | movement III. It is explanatory drawing which shows the basic concept of the said classification | category corresponding | compatible table creation operation IV. It is explanatory drawing following FIG. It is explanatory drawing following FIG. It is explanatory drawing following FIG. It is explanatory drawing following FIG. It is a flowchart which shows the content of the sub operation | movement III-3 of FIG. It is a flowchart which shows the content of the sub operation | movement III-4 of FIG. 5 is a flowchart showing the contents of sub-operation III-5 in FIG. 3. 14 is a flowchart showing the contents of the classification correspondence table creation operation IV. FIG. 5 is a flowchart showing the contents of sub-operation IV-1 in FIG. 4 when a matching table is created based on a JAN code. 5 is a flowchart showing the contents of sub-operation IV-2 of FIG. It is explanatory drawing which shows the structure of the POS data file of FIG. It is explanatory drawing which shows the structure of the converted POS data file of FIG. It is explanatory drawing which shows the structure of the market data file of FIG. It is explanatory drawing which shows the structure of the edited market data file of FIG. It is explanatory drawing which shows the structure of the sales data file of FIG. 3, Comprising: It is a part derived from POS data. It is also a part derived from market data. It is explanatory drawing which shows the structure of the non-sales data file of FIG. It is explanatory drawing which shows the structure of the investigation report data file of FIG. It is explanatory drawing following FIG. It is explanatory drawing following FIG. It is explanatory drawing which shows the structure of the calendar table of FIG. It is explanatory drawing which shows the structure of the classification | category corresponding table of FIG. It is explanatory drawing which shows the structure of the matching table of FIG. 14 is an explanatory diagram of a screen displayed on the output device (display device) in step S43 of the flowchart of FIG. 14 and corresponds to FIG. 6 (screen 1). It is explanatory drawing following FIG. 30, Comprising: It sorted by market classification (screen 2). It is explanatory drawing following FIG. 31, Comprising: It corresponds to FIG. 7 (screen 3). It is explanatory drawing following FIG. 32, Comprising: It corresponds to FIG. 8 (screen 4). In screen 4 shown in FIG. 33, (a) is a small screen that appears when editing the middle category, and (b) is an explanatory diagram of the small screen that appears when editing the major category. It is explanatory drawing following FIG. 33, Comprising: It corresponds to FIG. 9 (screen 5). It is explanatory drawing following FIG. 35, Comprising: It corresponds to FIG. 10 (screen 6). It is explanatory drawing which shows the structure of the total number-of-hits work table of FIG. 4, (a) is a thing when the screen 2 shown in FIG. 31 is displayed, (b) is the screen 3 shown in FIG. Is when It is explanatory drawing which shows the structure of the hit number work table of FIG. 4, Comprising: It is a thing when the screen 2 shown in FIG. 31 is displayed. Similarly, the screen 3 shown in FIG. 32 is displayed. Similarly, the screen 5 shown in FIG. 35 is displayed. Similarly, the screen 6 shown in FIG. 36 is displayed. 4A is an explanatory diagram showing the configuration of a POS data / JICFS classification correspondence table, and FIG. 4B is a market data / JICFS classification correspondence table when a matching table is created based on the JISFS classification in sub-operation IV-1 of FIG. (C) is an explanatory view showing the structure of the matching table created from (a) and (b). It is explanatory drawing of the effect of this invention.

Explanation of symbols

10 Computer (Classification compatible system)
11 Central processing unit 12 Input device 13 Output device 14 Storage device 14a Program storage unit 14b Data storage unit iv Market data v POS data X Survey report

Claims (7)

A method of associating two classification systems having different classification methods when classifying an article into a plurality of categories,
Computer
Get information about the classifications of both classification systems,
Based on this acquired information, find one classification of one classification system and one classification of the other classification system in which the same article is classified,
It is determined whether or not the duplication rate of the articles classified in both of the found 1 categories is larger than the duplication rate for determination,
As a result of the determination, when the overlap rate is larger than the determination overlap rate, the classification correspondence method is characterized in that the first sections are associated with each other and stored in a table. In the classification corresponding method according to claim 1,
Computer
When one category of one classification system and one classification of the other classification system in which the same article is classified cannot be found or, as a result of the determination, the overlap rate is not greater than the determination overlap rate, Change at least one of the classification of one classification system and the classification of the other classification system to the classification of the lower hierarchy or the classification of the higher hierarchy,
A classification correspondence method, wherein the classification of the lower hierarchy or the classification of the upper hierarchy is changed, and the classification heading or the duplication rate is judged again. In the classification corresponding method according to claim 1 or 2,
The one classification system is a classification system created by retailers for the products handled,
The other classification system is a classification system created by the market research company for the survey target products including the products handled by the retailer,
Computer
Obtain information on the sales performance of each product of the retailer and information on the survey results of each product of the market research company, and
Based on the correspondence between one category stored in the table, the retailer does not sell, but the market research classifies products sold in the retailer classification system,
Recorded together with survey results of products sold by retailers in the classification system of this retailer classification system and survey results of products not sold by retailers but sold in market research A classification response method characterized by creating a survey report for a retailer. A system for associating two classification systems having different classification methods when classifying articles into a plurality of categories,
Information acquisition means for acquiring information on the classification of both classification systems;
Search means for finding one classification of one classification system and one classification of the other classification system in which the same article is classified based on the information acquired by the information acquisition means;
Determination means for determining whether or not the duplication rate of articles classified into both one category found by the search means is greater than the duplication rate for determination;
A classification correspondence system comprising: a classification correspondence table storage means for associating the one category with each other and storing them in a table when the overlap ratio is greater than the determination overlap ratio as a result of the determination by the determination means. . In the classification corresponding system according to claim 4,
When the search means does not find one category of one classification system and one classification system of the other classification system in which the same article is classified, or as a result of determination by the determination means, the duplication rate is for determination If it is not greater than the overlap rate, it further comprises hierarchy changing means for changing at least one of the classification of one classification system and the classification of the other classification system to a classification of a lower hierarchy or a classification of an upper hierarchy,
With respect to the lower hierarchical level or the higher hierarchical level changed by the hierarchical level changing means, the search means performs the re-categorized heading or the determining means determines the duplication rate again. Classification support system. In the classification correspondence system according to claim 4 or 5,
The one classification system is a classification system created by retailers for the products handled,
The other classification system is a classification system created by the market research company with respect to the survey target products including the products handled by the retailer,
The information acquisition means acquires information on sales results for each product of a retailer and information on a survey result for each product of a market research company, and
Based on the association between one category stored in the table by the classification correspondence table storage means, the retailer does not sell, but the product sold in the market research is classified into categories of the retailer classification system. Classification means;
With this classification means classified into the classification of retailer classification system, the survey results of products sold by retailers and the survey results of products not sold by retailers but sold in market research A classification correspondence system, further comprising: a survey report creating means for creating a survey report for a retailer described together. A program for associating two classification systems having different classification methods when classifying articles into a plurality of categories,
Computer
Information acquisition means for acquiring information on the classification of both classification systems;
Search means for finding one classification of one classification system and one classification of the other classification system in which the same article is classified based on the information acquired by the information acquisition means;
Determining means for determining whether or not the duplication rate of the articles classified into both one category found by the search means is larger than the duplication rate for determination; and
As a result of the determination by the determination means, if the overlap rate is larger than the determination overlap rate, the classification correspondence program is made to function as a classification correspondence table storage means for associating the one category with each other and storing them in a table .

Images