JP5670490B2 - Category determination device, search device, category determination method, category determination program, and computer-readable recording medium storing the program - Google Patents

Description

  One embodiment of the present invention relates to an apparatus, a method, a program, and a recording medium for determining a product category, and an apparatus for searching for a product.

  In so-called virtual shopping streets where many virtual stores open via the Internet, a product category (hereinafter, also simply referred to as “category”) is often given to products traded in each store. By classifying products using product categories, the manager of the virtual shopping mall can systematically manage a huge number of products that each store purchases based on its own sales plan. Can be used to search and compare products from different stores at once. For example, the merchandise transaction system described in Patent Document 1 below has a plurality of search databases corresponding to a plurality of categories representing the types of merchandise. In this system, for each search database corresponding to each category, product specification information of that category is associated with daily life words that express the features of the product.

JP 2007-264747 A

  The association between the product and the category is often performed by the store, but the person in charge of the store may assign the wrong category to the product. Then, the product information includes an error, and when the user searches for a product from the category, the product that should be extracted is not extracted. This not only means that the store loses the opportunity to sell, but it also means that the use of the product information in the entire virtual shopping street may be remarkably deteriorated due to overlapping wrong category registrations.

  Therefore, it is required to correctly set the product category indicated by the product information.

  The category determination apparatus according to an aspect of the present invention, from the product display page of the virtual store on the network, an acquisition unit that acquires a product name and / or product description including a noun, a noun phrase, or a continuous noun sequence thereof, A compulsory word that matches the subject word with reference to an extraction unit that extracts a product-specific phrase from the merchandise name and / or product description as a subject word, and a dictionary database that stores dictionary information in which the compulsory word and the merchandise category are associated. Is a determination unit that identifies the product category associated with the compulsory word as the product category indicated by the product name, and one compulsory word uniquely identifies one product category. The determination unit, and an output unit that outputs category information indicating the product category specified by the determination unit.

  A category determination method according to an aspect of the present invention is a category determination method executed by a category determination device, and includes nouns, noun phrases, or continuous noun phrases thereof from a product display page of a virtual store on a network. An acquisition step for acquiring a product name and / or product description, an extraction step for extracting a product-specific phrase from the product name and / or product description as a subject word, and dictionary information in which compulsory words and product categories are associated are stored. When there is a compulsory word that matches the subject word, the product category associated with the compulsory word is identified as the product category indicated by the product name. The determination step in which one product category is uniquely specified by one compulsory word, and the product category specified in the determination step And an output step of outputting the category information indicating.

  The category determination program according to an aspect of the present invention includes an acquisition unit that acquires a product name and / or product description including a noun, a noun phrase, or a continuous noun sequence thereof from a product display page of a virtual store on the network; A compulsory word that matches the subject word with reference to an extraction unit that extracts a product-specific phrase from the merchandise name and / or product description as a subject word, and a dictionary database that stores dictionary information in which the compulsory word and the merchandise category are associated. Is a determination unit that identifies the product category associated with the compulsory word as the product category indicated by the product name, and one compulsory word uniquely identifies one product category. The computer is caused to execute the determination unit and an output unit that outputs category information indicating the product category specified by the determination unit.

  A computer-readable recording medium according to an aspect of the present invention acquires a product name and / or product description including a noun, a noun phrase, or a continuous noun sequence thereof from a product display page of a virtual store on a network. Section, an extraction unit that extracts a product-specific phrase from the product name and / or product description as a subject word, and a dictionary database that stores dictionary information in which compulsory words and product categories are associated with each other. When a compulsory word is present, the determination unit identifies the product category associated with the compulsory word as the product category indicated by the product name, and one product category is uniquely identified by one compulsory word. A category determination program that causes a computer to execute the determination unit specified in the above and an output unit that outputs category information indicating the product category specified by the determination unit. And stores the gram.

  According to such a form, a product-specific phrase is extracted from the product name and / or product description as a subject word, and it is determined whether or not a compulsory word that matches the subject word exists in the dictionary. And when such a compulsory word exists, the merchandise category uniquely specified from the compulsory word is output as the category of the merchandise. In this way, the product category indicated by the product information can be set correctly by specifying and outputting the product category uniquely derived from the subject word with reference to the dictionary of compulsory words that can identify one product category. It becomes possible.

  In the category determination device according to another aspect, the determination unit includes the second compulsory word when there is no first compulsory word that matches the subject word but there is a second compulsory word included in the subject word. The product category associated with the word may be specified as the product category indicated by the product name. By expanding the search range of forced words in this way, the possibility that the product category can be set correctly increases.

  In the category determination device according to another aspect, the extraction unit refers to an unnecessary word dictionary that stores unnecessary words to be excluded from the product name, and extracts phrases other than unnecessary words from the product name and / or the product description, Of the extracted phrases, the first word may be extracted as a theme word. By using the unnecessary word dictionary in this way, the subject words can be easily extracted.

  In the category determination device according to yet another aspect, the extraction unit refers to a rule dictionary that stores a description rule of a product name indicating the position of a subject word, and extracts some words from the product name and / or product description. Then, the first word of the extracted phrases may be extracted as the subject word. By using the rule dictionary in this way, the subject words can be easily extracted.

  In the category determination device according to another aspect, the extraction unit may extract a noun or a series of nouns from the product name by morphological analysis, and extract the first word as a subject word from the extracted words. Thereby, a subject word can be extracted without preparing a special rule or dictionary.

  In the category determination device according to another aspect, when the category of the product indicated by the product name is not specified by the determination unit, the word positioned next to the subject word is extracted from the extracted phrases. It may be extracted as the next subject word, and the determination unit may determine whether there is a compulsory word that matches the next subject word. In this way, by repeatedly searching for compulsory words while changing the subject word, the possibility that the product category can be correctly set increases.

  In the category determination device according to another embodiment, the acquisition unit receives product information that is about to be newly registered in the product database from the store terminal, and the output unit transmits the category information to the store terminal. The product information may be registered in the product database in response to an operation at the store terminal after the category information is received and displayed. In this way, by presenting the product category to the store where product information is to be registered, the store can easily set the correct product category.

  In the category determination device according to another embodiment, the acquisition unit reads out product information including the product category registered by the store as a registered category from the product database, and the determination unit is associated with a compulsory word that matches the subject word. And the product category that is different from the registered category is specified as the product category indicated by the product name, and the output unit transmits the category information to the store terminal, and the store terminal receives and displays the category information. In response to an operation on the terminal, the product information in the product database may be updated so that the product category specified by the determination unit is included in the product information separately from the registered category. Thereby, the correct product category can be set while leaving the product category (registered category) registered by the store.

  In the category determination device according to another embodiment, the acquisition unit reads out product information including the product category registered by the store as a registered category from the product database, and the determination unit is associated with a compulsory word that matches the subject word. In addition, the product category that is different from the registered category is specified as the product category indicated by the product name, and the output unit stores the product category specified by the determination unit in the product information separately from the registered category. The product information in the product database may be updated without notification. Thereby, it is possible to set a correct product category while leaving the product category (registered category) registered by the store without waiting for a response from the store.

  In the category determination device according to another aspect, the dictionary information is information in which a compulsory word, a product category, and information specifying a store or a store category are associated, and the determination unit stores the store corresponding to the product information. It may be determined whether or not there is a compulsory word that matches the subject word with reference to dictionary information that matches the information. Thereby, a merchandise category can be set in consideration of the characteristic of a store.

  In the category determination device according to another aspect, the dictionary database includes dictionary information in which the first type forced word and the product category are associated, dictionary information in which the second type forced word and the product category are associated, The dictionary information in which the word and the product category are associated is stored, one product category is uniquely specified by one type 1 compulsory word, and the type 2 compulsory word is more of the product category than the type 1 compulsory word. If the compulsory power is weak, the reference word is weaker in the product category than the second type compulsory word, and the determination unit includes the first type compulsory word, the determination unit is associated with the first type compulsory word. If there is a second type compulsory word that matches the subject word, the merchandise category associated with the second type compulsory word is identified, and a second subject with one or more subject words is identified. If it contains a seed compulsory word, the one or more second If the product category is identified based on the compulsory word and other words in the subject word, and the subject word does not include the first type compulsory word and the second type compulsory word and includes one or more reference words, A product category may be specified based on the one or more reference words.

  In this case, since the product category is determined in detail using the three types of phrases of the first and second types of compulsory words and reference words, an appropriate product category is selected from the product names having various description modes. Can be identified.

  In the category determination device according to another aspect, the extraction unit may extract a subject word from the product name with reference to a notation dictionary that stores information related to the shaking of the notation. Thereby, the correct product category can be set by absorbing the shaking of the notation.

  In the category determination device according to yet another aspect, the determination unit determines whether there is a compulsory word that matches the subject word with reference to a notation dictionary and a dictionary database that store information related to notation fluctuation. May be. Thereby, the correct product category can be set by absorbing the shaking of the notation.

  A search device according to an aspect of the present invention is a search device that can access a product database that stores product information updated by the category determination device, and matches a product category search condition received from a user terminal. A search unit that extracts product information from the product database and transmits the product information to the user terminal is included, and the search unit matches the search condition in addition to the product information whose registered category matches the search condition. Extract product information.

  According to such a form, the product information is searched with reference to not only the product category (registered category) registered by the store but also the product category specified by the category determination device. As a result, the quality of search results can be improved.

  According to one aspect of the present invention, a product category can be correctly set by specifying and outputting a product category uniquely derived from a subject word with reference to a dictionary of compulsory words that can specify one product category. it can.

It is a figure showing the whole electronic commerce (EC) system composition concerning an embodiment. It is a figure which shows the example of shop information. It is a figure which shows the example of the merchandise information in 1st Embodiment. (A)-(c) is a figure which shows the example of dictionary information (forced word dictionary). It is a figure which shows the hardware constitutions of EC server which concerns on 1st Embodiment. It is a block diagram which shows the function structure of EC server which concerns on 1st Embodiment. It is a flowchart which shows operation | movement of EC server shown in FIG. It is a figure which shows the structure of the category determination program which concerns on 1st Embodiment. It is a figure which shows the example of the merchandise information in 2nd Embodiment. It is a block diagram which shows the function structure of EC server which concerns on 2nd Embodiment. It is a flowchart which shows operation | movement of EC server shown in FIG. It is a figure which shows the structure of the category determination program which concerns on 2nd Embodiment. It is a figure which shows the example of the dictionary information in 3rd Embodiment. It is a block diagram which shows the function structure of EC server which concerns on 3rd Embodiment. It is a flowchart which shows operation | movement of EC server shown in FIG.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following embodiment, the category determination device and the search device according to the present invention are applied to an electronic commerce server (EC server). In the description of the drawings, the same or equivalent elements are denoted by the same reference numerals, and redundant description is omitted.

(First embodiment)
First, the function and configuration of the EC server 10 according to the first embodiment will be described with reference to FIGS. The EC server 10 is a component of the electronic commerce (EC) system 1.

  The EC system 1 is a computer system that provides a user (general consumer) with a website (EC site) of a virtual shopping street in which a plurality of virtual stores open. The user can search for a product or perform a purchase procedure through the site. The store operator registers the product to be sold in the virtual store in the virtual shopping mall, and delivers the product purchased by the user to the user. In addition, the kind of goods is not limited, Arbitrary tangibles and intangibles, such as a service, may be sufficient. In addition to the EC server 10, the EC system 1 includes a database group 20, terminals of each store operator, that is, store terminals Ts, and a large number of user terminals Tu. These devices are connected to each other via a communication network N.

  The EC server 10 is a computer system that manages EC sites. In response to a request from the user terminal Tu, the EC server 10 provides various web pages such as a main page and a product page, assists in product search, and purchase processing (addition and deletion of registered products based on input from the store terminal Ts) , Providing payment means, giving points to the user, etc.). In addition, the EC server 10 executes product information update processing (product registration or deletion, inventory addition, etc.) in response to a request from the store terminal Ts. Further, the EC server 10 determines a correct category (category that should be originally set) of a product handled in the virtual shopping mall.

  The database group 20 is a collection of various databases necessary for the EC system 1.

  The store terminal Ts is a computer owned by an operator of a store participating in the virtual shopping mall. The type of store terminal Ts is not limited, and may be, for example, a stationary or portable personal computer, or a portable terminal such as a high-function mobile phone (smart phone), a mobile phone, or a personal digital assistant (PDA). The number of store terminals Ts present in the EC system 1 is not limited.

  The user terminal Tu is a computer owned by a user (general consumer). Similar to the shop terminal Ts, the type of the user terminal Tu is not limited, and may be a stationary or portable personal computer, or a portable device such as a high-function mobile phone (smart phone), a mobile phone, or a personal digital assistant (PDA). It may be a terminal. The number of user terminals Tu existing in the EC system 1 is not limited.

  Based on the above, the details of the EC server 10 will be described. First, various databases accessed by the EC server 10 will be described in detail.

  The store database 21 is means for storing store information regarding each virtual store in the virtual shopping mall. The store information includes a store ID for uniquely identifying the store and attribute information indicating various attributes of the store. The attribute information includes the store name, address, category, and URL, but the information included in the attribute information is not limited. An example of store information is shown in FIG. From this figure, for example, it can be seen that the store with the store ID “S001” is “AB shop” and the category of this store is “alcoholic”. The store category may be systematically managed by a tree structure. One store may have a plurality of store categories.

  The product database 22 is a means for storing product information relating to products sold by each virtual store. The product information includes a store ID of a store that provides the product, a product ID that uniquely identifies the product, and attribute information that indicates various attributes of the product. The attribute information includes the store name, price, URL, inventory quantity, and category, but the information included in the attribute information is not limited. Here, the product category is a concept indicating the classification of products. The merchandise information is newly registered, updated, or deleted based on the instruction transmitted from the store terminal Ts. Here, the product name is the title displayed in the product name column of the product page (web page on which product information is posted) rather than the unique name or general name of the product, and is common to multiple types of products. (For example, a brand name, a manufacturer name, and an advertising phrase such as “free shipping”).

  An example of product information is shown in FIG. The registered category is a product category input and registered by the store. In this example, the store ID “S001” is associated with the product “BB beer” with the product ID “T002”, and the registered category of this product is specified as “food> Japanese sake / shochu”. Yes. The product categories are systematically managed by a tree structure. For example, “Food> Sake / Shochu” indicates the second hierarchy “Sake / Food” located under the first hierarchy “Food”.

  In this embodiment, each store sets not only the price of the product but also the product ID and the registration category, so that different product IDs and registration categories can be assigned to the same product between the stores. In the example of FIG. 3, the product ID of the product “AA liquor” is “T001” in the store “S001”, whereas “T101” is set in the store “S002”. Further, regarding the product “CC water”, both the product ID and the registration category are different between the stores “S001” and “S002”.

  In this way, in a situation where all stores are not forced to use a common product ID and product category, it is effective to determine the correct product category and take the consistency of the product category in the EC site as in this embodiment. It is. Even if the product ID is common to all stores (for example, even if a product ID common to stores is set based on a barcode such as a JAN code), if each store can freely set the product category, Category determination in the present embodiment is effective.

  The dictionary database 23 is a means for storing a compulsory word dictionary used for determining the product category. A compulsory word dictionary is a collection of dictionary information in which compulsory words (Definitive Category Words (DCW)) and product categories are associated, and a compulsory word is a phrase that uniquely identifies one product category. For example, it may be possible to set a word composed of a proper noun of a product, a general noun of an object, and a plurality of nouns as a forced word. The compulsory word dictionary is created in advance by the administrator of the EC server 10 and stored in the dictionary database 23.

  Three examples of forced word dictionaries are shown in FIG. The example of FIG. 4A is a compulsory word dictionary that does not consider the difference between stores and store categories. In this case, each record includes one compulsory word and one product category. From the example of FIG. 4A, the compulsory word “AA sake” is associated with only one product category “food> Japanese sake / shochu”, and the compulsory word “BB beer” has one product category “food> It can be seen that only “beer / Western sake” is associated. Note that the product category associated with the compulsory word may be a leaf category located at the end of the category tree, or a top-level or middle-level category. In addition, a product category that ends with “others” may be associated with a compulsory word.

  On the other hand, the examples of FIGS. 4B and 4C are compulsory word dictionaries created in consideration of the point that one word can be related to a plurality of product categories. The example of FIG. 4B is a case where a compulsory word dictionary is prepared for each store category, and each record includes a store category, a compulsory word, and a product category. The example of FIG. 4C is a case where a compulsory word dictionary is prepared for each store, and each record includes a store ID, a compulsory word, and a product category.

  For example, in the example of FIGS. 4B and 4C, considering that boots can be a product related to fashion or a product related to DIY (Do-it-yourself), for the compulsory word “boots”, Different store categories or different product categories are associated with each store. That is, in FIG. 4B, when determining the category of the commodity of the store that handles daily necessities, one commodity category “DIY> garden / DIY” corresponds to the compulsory word “boots” and handles fashion. When determining the category of merchandise in a store, one merchandise category “fashion> shoes” corresponds to the compulsory word “boots”. Also in FIG. 4C, as in the case of FIG. 4B, the product category corresponding to the compulsory word “boots” differs between the store with the store ID “S101” and the store with the store ID “S201”. Yes.

  The configurations of the store database 21 (store information), the product database 22 (product information), and the dictionary database (dictionary information) are not limited to those shown in FIGS. May be done.

  The database group 20 includes a user database that stores user information (member information), a purchase history database that stores purchase history of each user, etc., but for these databases that are not directly related to the determination of the product category, Detailed description is omitted.

  Next, the function and configuration of the EC server 10 will be described. The hardware configuration of the EC server 10 is shown in FIG. As shown in this figure, the EC server 10 includes a CPU 101 that executes an operating system, application programs, and the like, a main storage unit 102 that includes a ROM and a RAM, an auxiliary storage unit 103 that includes a hard disk, The communication control unit 104 includes a network card, an input device 105 such as a keyboard and a mouse, and an output device 106 such as a display.

  Each functional component of the EC server 10 described later reads predetermined software on the CPU 101 and the main storage unit 102, and operates the communication control unit 104, the input device 105, the output device 106, and the like under the control of the CPU 101. This is realized by reading and writing data in the main storage unit 102 and the auxiliary storage unit 103. Data and databases necessary for processing are stored in the main storage unit 102 and the auxiliary storage unit 103. Although FIG. 5 shows that the EC server 10 is configured by one computer, the functions of the EC server 10 may be distributed to a plurality of computers.

  In the present embodiment, when the store staff registers the product information, the EC server 10 automatically determines the product category and notifies the store of the category. In order to execute this category determination, the EC server 10 includes a reception unit (acquisition unit) 11, an extraction unit 12, a determination unit 13, a notification unit (output unit) 14, and a registration unit 15 as functional components. . These functional components are shown in FIG.

  The receiving unit 11 is means for receiving the product information input at the store terminal Ts from the store terminal Ts. Or the receiving part 11 may acquire merchandise information from the merchandise display page of the virtual store corresponding to the shop terminal Ts. The product category has not yet been set in this product information. The receiving unit 11 outputs the received product information to the extracting unit 12. The product information includes a product name and / or product description, and the product name and / or product description includes a noun, a noun phrase, or a series of nouns thereof.

  The extraction unit 12 is a means for extracting a product-specific phrase as a subject word from a product name that is a part of product information. As described above, the product name includes other words such as a brand name, a manufacturer name, and an advertising phrase in addition to the unique name or general name of the product. And since the brand name, maker name, and promotional phrase are common to several types of products, they are not suitable for determining the product category. Therefore, it is necessary to extract a product-specific phrase (subject word) from the product name. In the present embodiment, it is assumed that the subject word is extracted from the product name. However, the extraction unit 11 may extract the subject word from items other than the product name of the product information (for example, product description).

  The method for extracting the subject word is not limited. For example, the extraction unit 12 may delete a specific word / phrase from the product name by referring to a dictionary of unnecessary words prepared in advance, and may use the first word or phrase among the remaining words / phrases as a subject word. In this case, a database of unnecessary word dictionaries (unnecessary word list) is required, but this database may be included in the database group 20 or may be provided in the EC server 10. The unnecessary word can be defined as a phrase that cannot be said to be a product-specific phrase or a phrase that is common to a plurality of types of products. For example, it can be said that brand names and maker names are unnecessary words, and frequently used advertising phrases such as “free shipping” and “point n times” are also unnecessary words. In addition, quantities and code numbers can be unnecessary words.

  The extraction unit 12 refers to the unnecessary word dictionary and removes unnecessary words from the product name of the product information. Subsequently, the extraction unit 12 specifies the first word or phrase of the remaining character string as the subject word. For example, if the product name is “Brand X BB Beer Free Shipping” and “Brand X” and “Free Shipping” are registered in the unnecessary word dictionary, the extraction unit 12 uses the remaining word “BB Beer” as the subject. Identify as a word. In another example, when the product name is “Brand Y Denim Jacket” and “Brand Y” is registered in the unnecessary word dictionary, the extraction unit 12 is the first word of the remaining phrases “Denim Jacket”. Identify "denim" as the subject word. Alternatively, the extraction unit 12 may regard “denim jacket” as a theme word as one phrase.

  As another method, the extraction unit 12 may specify a subject word along the rule on the assumption that the product name is described under a predetermined rule. In this case, the extraction unit 12 holds information indicating the rule as a rule dictionary in advance. For example, if the product name description rule (or extraction rule) is “(brand name) + (subject candidate for subject word) + (other character string to be excluded)”, the product name to be processed is “brand X”. If it is “BB beer free shipping”, the extraction unit 12 extracts “BB beer” from the product name based on this rule, and extracts the word “BB beer” at the head of the extracted words as the subject word. (This is an example in which the candidate for the subject word includes only the subject word itself).

  The extraction unit 12 extracts a subject word from the product name of the input product information by any one of the methods described above, and outputs the product information and information indicating the subject word to the determination unit 13. By using an unnecessary word dictionary or a rule dictionary, a subject word can be easily extracted.

  The determination unit 13 is a means for determining a product category (product category to be originally registered) indicated by the product information by referring to the compulsory word dictionary using the subject word. First, the determination unit 13 refers to the dictionary database 23 to determine whether there is a compulsory word that matches the input subject word.

  When a compulsory word that matches the subject word is found, the determination unit 13 determines that the product category associated with the compulsory word is the product category indicated by the input product information. Then, the determination unit 13 outputs information indicating the specified product category to the notification unit 14 together with the product information.

  On the other hand, the process of the determination unit 13 when the compulsory word that matches the subject word is not found is not limited. For example, the determination unit 13 may output information indicating that a product category corresponding to the product name does not exist to the notification unit 14 when a compulsory word that matches the subject word is not found.

  As another example, the determination unit 13 refers to the dictionary database 23 again, determines whether or not a compulsory word included in the subject word exists, and determines whether or not such a compulsory word exists. Also good. In this case, if one or more compulsory words included in the subject word are found, the determination unit 13 determines whether the merchandise category associated with each compulsory word is the merchandise category indicated by the input merchandise information. It is determined that it can be. And the determination part 13 outputs the information which shows the specified one or more goods category to the notification part 14 with goods information. For example, when the subject word is “product A accessory” and two compulsory words “product A” and “product A” are registered in the compulsory word dictionary, the determination unit 13 uses these compulsory words. Two corresponding product categories are set as category candidates. On the other hand, when the compulsory word included in the subject word is not found, the determination unit 13 outputs information indicating that the product category corresponding to the product name does not exist to the notification unit 14.

  As yet another example, the extraction unit 12 extracts another subject word from an item other than the product name of the product information (for example, product description or image), and the determination unit 13 selects a compulsory word corresponding to the subject word. You may search. That is, when a compulsory word cannot be found from the product name, the compulsory word may be searched based on other types of product information.

  In this way, by expanding the search range of compulsory words, compulsory words can be easily found, and the possibility that the product category can be set correctly increases accordingly.

  If the dictionary information is determined for each store or each store category as shown in FIG. 4B or 4C, the determination unit 13 determines the store ID of the product information or the store derived from this ID. The compulsory word corresponding to the subject word may be searched based on the store category of the store information in the database 21. In this case, the product category can be set in consideration of the store characteristics. In the case of FIG. 4B, when one store has a plurality of store categories, the determination unit 13 may refer to the dictionary information of each store category. In this case, if a plurality of product categories are extracted, the EC server 10 may recommend those product categories to the store and cause the store to select them.

  The notification unit 14 is means for transmitting the product category determined by the determination unit 13 to the store terminal Ts that has sent the product information. When information on one or more product categories is input from the determination unit 13, the notification unit 14 transmits the category information to the store terminal Ts as a recommended category of products. In this case, the merchandise category automatically determined by the EC server 10 is displayed on the store terminal Ts, and the store staff can adopt or select the merchandise category as it is.

  On the other hand, when information indicating that the product category does not exist is input from the determination unit 13, the notification unit 14 transmits the information to the store terminal Ts. In this case, since the product category is not automatically displayed on the screen of the store terminal Ts, the store staff needs to input the product category by himself.

  The registration unit 15 is means for receiving product information from the store terminal Ts and storing it in the product database 22. When an operation for registering product information including a product category is performed at the store terminal Ts, the store terminal Ts transmits the product information to the EC server 10. The registration unit 15 stores the product information in the product database 22.

  Next, the operation of the EC server 10 will be described with reference to FIG. 7, and the category determination method according to the present embodiment will be described.

  First, the receiving unit 11 receives merchandise information in which the merchandise category is not input from the store terminal Ts (step S11, acquisition step). Subsequently, the extraction unit 12 extracts a subject word from the product name in the product information (step S12, extraction step). The extraction unit 12 extracts a subject word using the above-described unnecessary word dictionary or description rule (extraction rule).

  Subsequently, the determination unit 13 refers to the compulsory word dictionary (dictionary database 23) using the subject word to determine a product category that should be registered (determination step). The determination unit 13 first determines whether or not there is a compulsory word that matches the subject word, and if such a compulsory word exists (step S13; YES), the product category corresponding to the compulsory word is determined in the dictionary. Obtained from the database 23 (step S14a). Subsequently, the notification unit 14 transmits information on the product category as a compulsory message to the store terminal Ts (step S16a, output step). Here, the compulsory message is intended to force the store to set the product category.

  When there is no compulsory word that matches the subject word, but there is a compulsory word included in the subject word (step S13; NO and step S15; YES), the determination unit 13 identifies the one or more compulsory words identified. One or more product categories corresponding to is acquired from the dictionary database 23 (step S14b). Subsequently, the notification unit 14 transmits the information on the product category to the store terminal Ts with reference to the information (Step S16b, output step). In this case, not a message forcing the store to use the product category, but a message for the purpose of recommending the product category as a reference is sent to the store.

  On the other hand, when a compulsory word is not finally found (step S15; NO), the notification unit 14 notifies the store terminal Ts that there is no product category (step S17).

  Thereafter, product information including a product category is input at the store terminal Ts. And the registration part 15 stores the merchandise information in the merchandise database 22 according to registration operation of the said shop terminal Ts (step S18).

  As described above, the process of step S15 can be omitted. In this case, when there is no compulsory word that matches the subject word (step S13; NO), the notification unit 14 notifies the store terminal Ts that there is no merchandise category (step S17).

  Next, a category determination program P1 for causing a computer to function as the EC server 10 will be described with reference to FIG.

  The category determination program P1 includes a main module P10, a reception module P11, an extraction module P12, a determination module P13, a notification module P14, and a registration module P15.

  The main module P10 is a part that comprehensively controls the category determination function. The functions realized by executing the reception module P11, the extraction module P12, the determination module P13, the notification module P14, and the registration module P15 are respectively the reception unit 11, the extraction unit 12, the determination unit 13, the notification unit 14, and the like. The function is the same as that of the registration unit 15.

  The category determination program P1 is provided after being fixedly recorded on a tangible recording medium such as a CD-ROM, DVD-ROM, or semiconductor memory. Moreover, the category determination program P1 may be provided via a communication network as a data signal superimposed on a carrier wave.

  As described above, according to the present embodiment, a product-specific word / phrase is extracted from the product name as a subject word, and it is determined whether or not a compulsory word that matches the subject word exists in the dictionary database 23. . If such a compulsory word exists, the merchandise category uniquely identified from the compulsory word is notified to the store as the merchandise category indicated by the merchandise name. In this way, by referring to the compulsory word dictionary that can identify one product category and identifying the product category that is uniquely derived from the subject word, it is possible to notify the store of the product category that should be originally registered. . As a result, the store can be guided to register the correct product category, and the reliability of the product information can be further improved. Further, by using the compulsory word dictionary, it is possible to easily determine the product category that should be registered.

  In the present embodiment, since the product category is presented to the store where the product information is to be registered, the store can easily set the correct product category.

(Second Embodiment)
Next, the function and configuration of the EC server 30 according to the second embodiment will be described with reference to FIGS. In this embodiment, the EC server 30 does not determine the product category when the store attempts to register product information, but whether or not the product category of the product information already registered in the product database 22 is correct. Determine whether. In the following, description of items that are the same as or similar to those in the first embodiment will be omitted, and items specific to the EC server 30 will be specifically described.

  In the present embodiment, the product database 22 (product information) is configured as shown in FIG. That is, the product information includes a determination category automatically set by the EC server 30 in addition to a registration category that is a product category input and registered by the store. In the example of FIG. 9, the store ID “S001” is associated with the product “BB beer” with the product ID “T002”, and the registration category and the determination category of this product are “Food> Sake / Sake” “Shochu” and “Food> Beer / Western Sake”. The determination category may not be registered as in the record specified by the store ID “S001” and the product ID “T001”.

  The hardware configuration of the EC server 30 is the same as that in the first embodiment (see FIG. 5). As shown in FIG. 10, the EC server 30 includes an acquisition unit 31, an extraction unit 32, a determination unit 33, a registration unit (output unit) 34, and a search unit 35 as functional components.

  The acquisition unit 31 is means for reading out product information that is a target for determining a product category from the product database 22. The acquisition unit 31 outputs the read product information to the extraction unit 32.

  Since the functions of the extraction unit 32 and the determination unit 33 are the same as the functions of the extraction unit 12 and the determination unit 13 in the first embodiment, a detailed description thereof will be omitted. However, the determination unit 33 outputs the specified product category information or information indicating that the product category does not exist to the registration unit 34 together with the product information.

  The registration unit 34 is means for storing the determined product category (category information) in the product database 22 as a determination category. The registration unit 34 determines whether the determined product category is different from the registered category of the product information, and registers the determination category only when they are different.

  The registration method of the judgment category is not limited. For example, the registration unit 34 may register the product category in the product database 22 only when permission is obtained from the store for registration of the product category associated with the compulsory word that matches the subject word. In this case, the registration unit 34 transmits inquiry information including the determined product category and product information to the store terminal Ts. When the response information indicating that registration is permitted for the inquiry information is received from the store terminal Ts, the registration unit 34 records the determined product category in a corresponding record in the product database 22. As another method, the registration unit 34 may record the determined product category in a corresponding record in the product database 22 without making an inquiry to such a store.

  For example, assuming that the phrase “BB beer” is obtained as the subject word for the product “BB beer” in FIG. 9, the product category “food> beer / Western sake” is obtained from the dictionary database 23 shown in FIG. Therefore, as shown in FIG. 9, this “food> beer / Western liquor” is registered in the product database 22 as a determination category.

  When one or more product categories are acquired based on the compulsory word included in the subject word that does not match the subject word, the registration unit 34 transmits the product category as a recommended category to the store terminal Ts. When one merchandise category is designated at the store terminal Ts in response to this transmission, the registration unit 34 registers the merchandise category as a determination category in a corresponding record in the merchandise database 22.

  If the determined product category is the same as the registered category of the product information, the registration unit 34 ends the process without registering the determined category. Further, even when the determination unit 33 does not find a product category corresponding to the product information, the registration unit 34 does not register the determination category.

  The search unit 35 is means for searching for product information according to the search conditions received from the user terminal Tu and transmitting the extracted product information to the user terminal Tu. In the following, a process for searching for a product that matches the product category indicated by the search condition will be described.

  The search unit 35 reads product information whose registration category matches the search condition from the product database 22, and also reads product information whose determination category matches the search condition from the product database 22. Then, the search unit 35 transmits the read product information to the user terminal Tu as a search result. The search unit 35 may lower the search score for product information whose registered category matches the search condition and whose registered category is different from the determination category as compared to other product information. This process is performed in order to display the product in which the wrong product category is registered by the store below the other products when the search result is displayed on the user terminal Tu.

  For example, assuming that product information is stored as shown in FIG. 9 and the search condition is “food> beer / sake”, a product “BB beer” whose determination category matches the condition is searched. Even when the search condition is “food> Japanese sake / shochu”, the product “BB beer” whose registered category matches the condition is searched. In this case, the search score of “BB beer” is low. For example, “BB beer” is displayed below “AA liquor” on the search screen.

  Next, the operation of the EC server 30 will be described with reference to FIG. 11, and the category determination method according to the present embodiment will be described.

  First, the acquisition unit 31 acquires product information from the product database 22 (step S21, acquisition step). Subsequently, the extraction unit 32 extracts a subject word from the product name in the product information (step S22, extraction step). The process of step S22 is the same as step S12 in the first embodiment.

  Subsequently, the determination unit 33 refers to the compulsory word dictionary (dictionary database 23) using the subject word to determine a product category that should be registered (determination step). First, the determination unit 33 determines whether or not there is a compulsory word that matches the subject word. If such a compulsory word exists (step S23; YES), the product category corresponding to the compulsory word is determined in the dictionary. Obtained from the database 23 (step S24). If the product category is different from the registered category, the registration unit 34 stores the product category as a determination category in the product database 22 (step S25, output step). As described above, the registration unit 34 may inquire of the store terminal Ts in advance whether the determination category needs to be registered, or may register the determination category without making an inquiry.

  If there is no compulsory word that matches the subject word, but there is a compulsory word included in the subject word (step S23; NO and step S26; YES), the determination unit 33 identifies the one or more compulsory words specified. One or more product categories corresponding to is acquired from the dictionary database 23 (step S27). And the registration part 34 notifies the merchandise category to the store terminal Ts, and stores one merchandise category as a determination category in the merchandise database 22 in accordance with an instruction from the store terminal Ts (step S28, output step).

  If no compulsory word is found from the subject word (step S26; NO), the product information is not updated.

  The processes in steps S21 to S28 are repeatedly executed until all product information is processed (see step S29). As in the first embodiment, the processes in steps S26 to S28 can be omitted. In this case, when there is no compulsory word that matches the subject word (step S23; NO), processing for the next product information starts.

  Next, a category determination program P2 for causing a computer to function as the EC server 30 will be described with reference to FIG.

  The category determination program P2 includes a main module P20, an acquisition module P21, an extraction module P22, a determination module P23, a registration module P24, and a search module P25.

  The main module P20 is a part that comprehensively controls the category determination function. The functions realized by executing the acquisition module P21, the extraction module P22, the determination module P23, the registration module P24, and the search module P25 are the acquisition unit 31, the extraction unit 32, the determination unit 33, the registration unit 34, respectively. And the function of the search unit 35 is the same.

  The category determination program P2 can also be distributed by various methods in the same manner as the category determination program P1.

  As described above, according to the present embodiment, a product-specific word / phrase is extracted from the product name as a subject word, and it is determined whether or not a compulsory word that matches the subject word exists in the dictionary database 23. . And when such a compulsory word exists, the merchandise category uniquely specified from the compulsory word is registered as the category of the merchandise indicated by the merchandise name. In this way, the product category can be set correctly by specifying and registering the product category uniquely derived from the subject word with reference to the compulsory word dictionary that can identify one product category. As a result, the reliability of the product information can be further improved. Further, by using the compulsory word dictionary, it is possible to easily determine the product category that should be registered.

  In the present embodiment, the correct product category can be set while leaving the registered category without waiting for a response from the store or after obtaining an explicit answer from the store.

  In the present embodiment, the product information is searched with reference to not only the registered category but also the determination category, so that the quality of the search result can be improved.

(Third embodiment)
Next, the function and configuration of the EC server 10A according to the third embodiment will be described. In the present embodiment, the EC server 10A does not use one type of compulsory words to determine the product category, but uses two types of compulsory words (first type compulsory words and second type compulsory words) and reference words. . In the following, description of items that are the same as or similar to those in the first embodiment will be omitted, and items specific to the EC server 10A will be particularly described.

  In the present embodiment, the dictionary information stored in the dictionary database 23 is information indicating in which product category each phrase is set as a first type forced word, a second type forced word, or a reference word. It can be said that this dictionary information is information in which the product category is associated with at least one of the first type forced word, the second type forced word, and the reference word. The dictionary information is created in advance by the administrator of the EC server 10A and stored in the dictionary database 23.

  The first type forced word is a phrase that uniquely identifies one product category. If the first type forced word is included in the subject word extracted from the product name, the product is included in the corresponding product category. being classified. The second type forced word is a phrase that is referred to when the first type forced word does not exist in the subject word. A reference word is a phrase that is used to determine a product category when neither a first-type forced word nor a second-type forced word exists in the subject word.

  The compulsory power of the product category is the strongest in the first type compulsory word, next the second type compulsory word, and the weakest reference word. Specifically, when it is determined that a certain product belongs to the product category corresponding to the first type compulsory language, the product is always associated with the product category, and the store side may change this association. Can not. When it is determined that a certain product belongs to the product category corresponding to the second type compulsory language, the store is determined conditionally (for example, on the condition that the approval of the administrator of the EC site is obtained). The product can be associated with a product category other than the product category. When it is determined that a certain product belongs to the product category corresponding to the reference word, the store can unconditionally associate the product with a different product category from the determined one. Of course, each store can also associate a product with a product category corresponding to the second type compulsory word or reference word.

  One word may be a second type compulsory word in a plurality of product categories, and one word may be a reference word in a plurality of product categories. In some cases, a product category with one word becomes the second type compulsory word and a reference word in another product category. However, one type 1 compulsory word does not become a type 1 compulsory word, a type 2 compulsory word, or a reference word in other product categories.

  An example of dictionary information is shown in FIG. From this example, it is understood that the compulsory word “model X” is a second type compulsory word of the product category A and a reference word of the product category B. It can also be seen that the compulsory word “accessory” is registered as a reference word in the product categories B to G. Note that the product category associated with the compulsory language may be a leaf category, or a top or middle category.

  The hardware configuration of the EC server 10A is the same as that in the first embodiment (see FIG. 5). As shown in FIG. 14, the EC server 10A is different from the EC server 10 in the first embodiment in that a determination unit 13A is provided instead of the determination unit 13. The function of the determination unit 13A will be described below.

  First, the determination unit 13A compares the input subject word with the first type compulsory word in the dictionary database 23. When there is a first type compulsory word that matches or is included in the subject word, the determination unit 13A determines that the product category associated with the first type compulsory word is the input product information. It is determined that the product category is displayed.

  When the first type compulsory word corresponding to the subject word does not exist, the determination unit 13A compares the subject word with the second type compulsory word in the dictionary database 23. If there is a second type compulsory word that matches the subject word, the determination unit 13A determines that the product category associated with the second type compulsory word is the product category indicated by the input product information. Is determined.

  When the subject word includes two or more second-type forced words, the determination unit 13A sets a plurality of product categories corresponding to the plurality of second-type forced words as category candidates.

  If the subject word includes other words in addition to the second type forced word, the determination unit 13A determines the product category as follows.

  If the second type compulsory word corresponding to the product category X also serves as a reference word for another product category Y and the other words are reference words for the product category Y, the determination unit 13A determines the product category Y is determined to be the category of the product indicated by the input product information.

  When the second type compulsory word corresponding to the product category X also serves as a reference word for another product category Y and the other words are not reference words for the product category Y, the determination unit 13A determines that the product category is Judge that it does not exist. Even when the second type compulsory word does not serve as a reference word for another product category, the determination unit 13A determines that the product category does not exist.

  If the other word is not a reference word, the determination unit 13A determines whether or not the other word can be deleted with reference to a predetermined deletion possible list. Note that words to be registered in the deleteable list may be arbitrarily determined. If the other word is registered in the list and can be deleted, the determination unit 13A determines that the product category corresponding to the second type compulsory word is the product category indicated by the input product information. Is determined. On the other hand, if the other word is not registered in the list and cannot be deleted, the determination unit 13A determines that the product category does not exist.

  When the first type compulsory word and the second type compulsory word corresponding to the subject word do not exist, the determination unit 13A compares the subject word with the reference word in the dictionary database 23. When the subject word includes only one reference word, the determination unit 13A determines that the product category associated with the reference word is the category of the product indicated by the input product information. When the subject word includes two or more reference words, the determination unit 13A sets a plurality of product categories corresponding to the plurality of reference words as category candidates. On the other hand, when the subject word does not include the reference word, the determination unit 13A determines that the product category does not exist.

  When the determination unit 13A can identify one product category by the above processing, the notification unit 14 transmits information on the product category to the store terminal Ts. When the determination unit 13A extracts a plurality of product categories, the notification unit 14 transmits information on these product categories to the store terminal Ts. When the determination unit 13A determines that the product category does not exist, the notification unit 14 transmits information indicating the absence to the store terminal Ts in order to entrust setting of the product category to the store side.

  Some classification examples are shown on the premise of the dictionary information shown in FIG. If the subject word extracted from a certain product is “model X”, it matches the type 2 compulsory word of the product category A, so the product is classified into the product category A.

  If the subject word is “accessory for model X”, this subject word includes the second type compulsory word “model X” of product category A, and this “model X” also serves as a reference term for product category B. The other word “accessory” is a reference word for the product category B. Therefore, the product corresponding to this subject word is classified into the product category B.

  If the subject word is “model X type bonus”, the subject word includes the second type compulsory word “model X” of the product category A, and the other word “bonus” is not a reference word of the product category B. Therefore, it is determined that there is no product category corresponding to the subject word.

  It is assumed that the subject word is “cute model X” and the word “cute” is registered in the erasable list. In this case, the subject word includes the second type compulsory word “model X” of the product category A, and the other word “cute” can be deleted. Therefore, the product corresponding to this subject word is classified into the product category A.

  If the subject word matches or includes “screwdriver”, the product is classified into the product category D because the term is a first type compulsory word.

  If the subject word includes “driver”, since this word is the second type compulsory word of the product categories L and M, two product categories L and M are extracted as product category candidates corresponding to the subject word. Is done.

  The operation of the EC server 10A will be described using FIG. 15 and the category determination method according to the present embodiment will be described. The process after step S12 in FIG. 7 differs between the present embodiment and the first embodiment.

  When the extraction unit 12 extracts a subject word from the product name in the product information (step S12, extraction step), the determination unit 13A refers to the dictionary database 23 using the subject word to determine a product category that should be originally registered. Then, the notification unit 14 notifies the store terminal Ts based on the determination result (output step).

  If there is a first type compulsory word or a second type compulsory word that matches the subject word (step S31; YES), the determination unit 13A selects one product category corresponding to the compulsory word in the dictionary database. 23, the notification unit 14 notifies the store terminal Ts of the product category (step S32). When such a compulsory word does not exist (step S31; NO), the determination unit 13A determines whether or not the subject word includes the second type compulsory word (step S33).

  When the subject word includes the second type compulsory word (step S33; YES), the determination unit 13A determines the product category by the above method based on the number of the second type compulsory word and the presence of other words. (Step S34). If one or more product categories can be extracted (step S35; YES), the determination unit 13A acquires the product category from the dictionary database 23, and the notification unit 14 notifies the product category to the store terminal Ts. (Step S36). On the other hand, when the product category cannot be extracted (step S35; NO), the notification unit 14 notifies the store terminal Ts of the absence of the product category (step S37). In this case, the product category is manually set in the store (manual setting).

  When the subject word does not include the second type compulsory word (step S33; NO), the determination unit 13A determines whether the subject word includes a reference word (step S38). If the subject word includes a reference word (step S38; YES), the determination unit 13A acquires one or more product categories corresponding to the reference word from the dictionary database 23, and the notification unit 14 selects the product category. The store terminal Ts is notified (step S39). On the other hand, when the subject word does not include the reference word (step S38; NO), the notification unit 14 notifies the store terminal Ts of the absence of the product category (step S37), and the product category is manually set on the store side. Set (manual setting).

  In the third embodiment, the same effect as that of the first embodiment can be obtained. In addition, in the present embodiment, since the product category is determined in detail using the three types of phrases such as the first type and the second type of compulsory words and reference words, an appropriate product is selected from the product names having various descriptions. The category can be specified and presented to the store. The method of the present embodiment can also be applied as a modification of the second embodiment.

  The present invention has been described in detail based on the embodiments. However, the present invention is not limited to the above embodiment. The present invention can be variously modified without departing from the gist thereof.

  A database (notation dictionary) that stores information related to notation shaking may be used to absorb the notation shaking of words (the presence of multiple notations for one word). For example, if the English word “interface” is written in Japanese, the notation “interface (intafe-su)” and “interface (inta-feisu)” may occur. Therefore, when extracting the subject word from the product name or searching for the compulsory word from the subject word, the correct product category can be set by absorbing the notation fluctuation by referring to the notation dictionary.

  In each of the above embodiments, when the product category cannot be specified using the subject word, the determination unit may repeatedly search for the product category while changing the subject word. That is, when the product category cannot be specified by the first word in the phrase extracted using the unnecessary word dictionary or the rule dictionary, the extraction unit selects the second word in the phrase as the next subject. A word. And the determination part searches a compulsory word using the said next subject word, and specifies a goods category. If the product category cannot be specified by this process, the extraction unit sets the third word as the next subject word, and the determination unit performs the same process again using the subject word. In this way, by repeatedly searching for compulsory words while changing the subject word, the possibility that the product category can be correctly set increases.

  The extraction unit may extract a noun or a series of nouns by performing a morphological analysis on the product name, and may set the first word of the extracted phrase as a subject word. Thereby, a subject word can be extracted without preparing the above-mentioned unnecessary word dictionary and rule dictionary. However, in this case, since it is unclear whether the word is a product-specific phrase, it is preferable to search for a forced word repeatedly while changing the subject word as described above.

  In each of the above embodiments, the category determination device of the present invention is applied to an EC server. However, a category determination server may be provided separately from the EC server. For example, the category determination function (the acquisition unit 31, the extraction unit 32, the determination unit 33, and the registration unit 34) and the search function (search unit 35) in the second embodiment may be separately implemented on two servers. .

  From the above description, the present invention can be defined as follows.

An acquisition unit for acquiring product information including a product name;
An extraction unit for extracting a product-specific phrase from the product name as a subject word;
A dictionary database that stores dictionary information in which a compulsory word and a product category are associated is referred to, and when the compulsory word that matches the subject word exists, the product category associated with the compulsory word is A determination unit that identifies the product category indicated by the product name, wherein the one product category is uniquely identified by one compulsory word; and
A category determination apparatus comprising: an output unit that outputs category information indicating a product category specified by the determination unit.

The product associated with the second compulsory word when the first compulsory word matching the subject word does not exist but the second compulsory word included in the subject word exists. A category is specified as the category of the product indicated by the product name.
Item 4. The category determination device according to Item 1.

The extraction unit refers to an unnecessary word dictionary that stores unnecessary words to be excluded from product names, extracts words other than the unnecessary words from the product names included in the product information, and extracts the first of the extracted words Extracting a word as the subject word;
Item 3. The category determination device according to Item 1 or 2.

The extraction unit refers to a rule dictionary that stores a description rule of a product name indicating the position of the subject word, extracts a part of the phrase from the product name included in the product information, and among the extracted phrases Extracting the first word as the subject word;
Item 3. The category determination device according to Item 1 or 2.

The extraction unit extracts a noun or a series of nouns from the product name by morphological analysis, and extracts the first word of the extracted words as the subject word,
Item 3. The category determination device according to Item 1 or 2.

When the category of the product indicated by the product name is not specified by the determination unit, the extraction unit extracts a word positioned next to the subject word from the extracted words as the next subject word. ,
The determination unit determines whether or not the compulsory word that matches the next subject word exists;
Item 6. The category determination device according to any one of Items 3 to 5.

The acquisition unit receives the product information that is about to be newly registered in the product database from the store terminal,
The output unit transmits the category information to the store terminal,
The product information is registered in the product database in response to an operation at the store terminal after the category information is received and displayed by the store terminal.
Item 7. The category determination device according to any one of Items 1 to 6.

The acquisition unit reads the product information including the product category registered by the store as a registered category from the product database,
The determination unit identifies a product category that is associated with the compulsory word that matches the subject word and that is different from the registered category, as a product category indicated by the product name,
The output unit transmits the category information to a store terminal,
In response to an operation at the store terminal after the category information is received and displayed by the store terminal, the product category specified by the determination unit is included in the product information separately from the registered category, The product information in the product database is updated;
Item 7. The category determination device according to any one of Items 1 to 6.

The acquisition unit reads the product information including the product category registered by the store as a registered category from the product database,
The determination unit identifies a product category that is associated with the compulsory word that matches the subject word and that is different from the registered category, as a product category indicated by the product name,
The output unit updates the product information in the product database without notifying the store such that the product category specified by the determination unit is included in the product information separately from the registered category.
Item 7. The category determination device according to any one of Items 1 to 6.

The dictionary information is information in which the compulsory word, the product category, and information specifying a store or a store category are associated,
The determining unit refers to the dictionary information that matches the store information corresponding to the product information, and determines whether or not the compulsory word that matches the subject word exists;
Item 10. The category determination device according to any one of Items 1 to 9.

The dictionary database includes dictionary information in which a first type forced word and a product category are associated, dictionary information in which a second type forced word and a product category are associated, and a dictionary in which a reference word and a product category are associated. Remember information,
One merchandise category is uniquely identified by one of the first type compulsory words, the second type compulsory word has a lower force of merchandise category than the first type compulsory word, and the reference word is the second type The product category is less compelling than compulsory,
The determination unit is
If the subject word includes the first type compulsory word, identify the product category associated with the first type compulsory word;
If there is a second type compulsory word that matches the subject word, identify the product category associated with the second type compulsory word;
If the subject word includes one or more type 2 compulsory words, the product category is identified based on the one or more type 2 compulsory words and other words in the subject word;
When the subject word does not include the first type compulsory word and the second type compulsory word and includes one or more reference words, the product category is specified based on the one or more reference words.
Item 4. The category determination device according to Item 1.

The extraction unit extracts the subject word from the product name with reference to a notation dictionary that stores information related to the shaking of the notation;
Item 12. The category determination device according to any one of Items 1 to 11.

The determination unit determines whether or not the compulsory word that matches the subject word exists with reference to a notation dictionary that stores information on notation fluctuation and the dictionary database.
Item 13. The category determination device according to any one of Items 1 to 12.

A search device capable of accessing the product database storing product information updated by the category determination device according to Item 8 or 9,
A search unit that extracts the product information that matches the search condition of the product category received from the user terminal from the product database and transmits the product information to the user terminal;
The search unit extracts the product information in which the specified product category matches the search condition in addition to the product information in which the registered category matches the search condition;
Search device.

A category determination method executed by a category determination device,
An acquisition step for acquiring product information including the product name;
An extraction step of extracting a product-specific phrase from the product name as a subject word;
A dictionary database that stores dictionary information in which a compulsory word and a product category are associated is referred to, and when the compulsory word that matches the subject word exists, the product category associated with the compulsory word is A determination step for specifying the category of the product indicated by the product name, wherein the one product category is uniquely specified by one compulsory word; and
A category determination method including an output step of outputting category information indicating the product category specified in the determination step.

An acquisition unit for acquiring product information including a product name;
An extraction unit for extracting a product-specific phrase from the product name as a subject word;
A dictionary database that stores dictionary information in which a compulsory word and a product category are associated is referred to, and when the compulsory word that matches the subject word exists, the product category associated with the compulsory word is A determination unit that identifies the product category indicated by the product name, wherein the one product category is uniquely identified by one compulsory word; and
The category determination program which makes a computer perform the output part which outputs the category information which shows the goods category specified by the said determination part.

An acquisition unit for acquiring product information including a product name;
An extraction unit for extracting a product-specific phrase from the product name as a subject word;
A dictionary database that stores dictionary information in which a compulsory word and a product category are associated is referred to, and when the compulsory word that matches the subject word exists, the product category associated with the compulsory word is A determination unit that identifies the product category indicated by the product name, wherein the one product category is uniquely identified by one compulsory word; and
A computer-readable recording medium that stores a category determination program that causes a computer to execute an output unit that outputs category information indicating a product category specified by the determination unit.

  DESCRIPTION OF SYMBOLS 10 ... EC server (category determination apparatus), 11 ... Reception part (acquisition part), 12 ... Extraction part, 13, 13A ... Determination part, 14 ... Notification part (output part), 15 ... Registration part, 20 ... Database group, DESCRIPTION OF SYMBOLS 21 ... Store database, 22 ... Merchandise database, 23 ... Dictionary database, 30 ... EC server (category determination apparatus, search device), 31 ... Acquisition part, 32 ... Extraction part, 33 ... Determination part, 34 ... Registration part (output part) ), 35: Search unit, P1: Category determination program, P10: Main module, P11: Reception module, P12: Extraction module, P13: Determination module, P14: Notification module, P15: Registration module, P2: Category determination program, P20 ... main module, P21 ... acquisition module, P22 ... extraction module, P23 ... determination module, P2 ... registration module, P25 ... search module, Ts ... the shop terminal, Tu ... user terminal.

Claims (17)

An acquisition unit for acquiring a product name and / or product description including a noun, a noun phrase or a continuous noun sequence thereof from a product display page of a virtual store on the network;
An extraction unit that extracts a product-specific phrase from the product name and / or product description as a subject word;
A dictionary database that stores dictionary information in which a compulsory word and a product category are associated is referred to, and when the compulsory word that matches the subject word exists, the product category associated with the compulsory word is A determination unit that identifies the product category indicated by the product name, wherein the one product category is uniquely identified by one compulsory word; and
A category determination apparatus comprising: an output unit that outputs category information indicating a product category specified by the determination unit. The product associated with the second compulsory word when the first compulsory word matching the subject word does not exist but the second compulsory word included in the subject word exists. A category is specified as the category of the product indicated by the product name.
The category determination apparatus according to claim 1. The extraction unit refers to an unnecessary word dictionary that stores unnecessary words to be excluded from product names, extracts words other than the unnecessary words from the product name and / or product description, and extracts the first of the extracted words Extracting a word as the subject word;
The category determination apparatus according to claim 1 or 2. The extraction unit refers to a rule dictionary that stores a description rule of a product name indicating the position of the subject word, extracts some words from the product name and / or product description, and among the extracted words Extracting the first word as the subject word;
The category determination apparatus according to claim 1 or 2. The extraction unit extracts a noun or a series of nouns from the product name by morphological analysis, and extracts the first word of the extracted words as the subject word,
The category determination apparatus according to claim 1 or 2. When the category of the product indicated by the product name is not specified by the determination unit, the extraction unit extracts a word positioned next to the subject word from the extracted words as the next subject word. ,
The determination unit determines whether or not the compulsory word that matches the next subject word exists;
The category determination apparatus as described in any one of Claims 3-5. The acquisition unit receives product information that is about to be newly registered in the product database from the store terminal,
The output unit transmits the category information to the store terminal,
The product information is registered in the product database in response to an operation at the store terminal after the category information is received and displayed by the store terminal.
The category determination apparatus as described in any one of Claims 1-6. The acquisition unit reads product information including a product category registered by the store as a registered category from the product database,
The determination unit identifies a product category that is associated with the compulsory word that matches the subject word and that is different from the registered category, as a product category indicated by the product name,
The output unit transmits the category information to a store terminal,
In response to an operation at the store terminal after the category information is received and displayed by the store terminal, the product category specified by the determination unit is included in the product information separately from the registered category, The product information in the product database is updated;
The category determination apparatus as described in any one of Claims 1-6. The acquisition unit reads product information including a product category registered by the store as a registered category from the product database,
The determination unit identifies a product category that is associated with the compulsory word that matches the subject word and that is different from the registered category, as a product category indicated by the product name,
The output unit updates the product information in the product database without notifying the store such that the product category specified by the determination unit is included in the product information separately from the registered category.
The category determination apparatus as described in any one of Claims 1-6. The dictionary information is information in which the compulsory word, the product category, and information specifying a store or a store category are associated,
The determination unit refers to the dictionary information that matches store information corresponding to product information, and determines whether or not the compulsory word that matches the subject word exists;
The category determination device according to any one of claims 1 to 9. The dictionary database includes dictionary information in which a first type forced word and a product category are associated, dictionary information in which a second type forced word and a product category are associated, and a dictionary in which a reference word and a product category are associated. Remember information,
One merchandise category is uniquely identified by one of the first type compulsory words, the second type compulsory word has a lower force of merchandise category than the first type compulsory word, and the reference word is the second type The product category is less compelling than compulsory,
The determination unit is
If the subject word includes the first type compulsory word, identify the product category associated with the first type compulsory word;
If there is a second type compulsory word that matches the subject word, identify the product category associated with the second type compulsory word;
If the subject word includes one or more type 2 compulsory words, the product category is identified based on the one or more type 2 compulsory words and other words in the subject word;
When the subject word does not include the first type compulsory word and the second type compulsory word and includes one or more reference words, the product category is specified based on the one or more reference words.
The category determination apparatus according to claim 1. The extraction unit extracts the subject word from the product name with reference to a notation dictionary that stores information related to the shaking of the notation;
The category determination device according to any one of claims 1 to 11. The determination unit determines whether or not the compulsory word that matches the subject word exists with reference to a notation dictionary that stores information on notation fluctuation and the dictionary database.
The category determination device according to any one of claims 1 to 12. A search device capable of accessing the product database for storing product information updated by the category determination device according to claim 8 or 9,
A search unit that extracts the product information that matches the search condition of the product category received from the user terminal from the product database and transmits the product information to the user terminal;
The search unit extracts the product information in which the specified product category matches the search condition in addition to the product information in which the registered category matches the search condition;
Search device. A category determination method executed by a category determination device,
An acquisition step of acquiring a product name and / or product description including a noun, a noun phrase or a continuous noun sequence thereof from a product display page of a virtual store on the network;
An extraction step of extracting product-specific phrases from the product name and / or product description as subject words;
A dictionary database that stores dictionary information in which a compulsory word and a product category are associated is referred to, and when the compulsory word that matches the subject word exists, the product category associated with the compulsory word is A determination step for specifying the category of the product indicated by the product name, wherein the one product category is uniquely specified by one compulsory word; and
A category determination method including an output step of outputting category information indicating the product category specified in the determination step. An acquisition unit for acquiring a product name and / or product description including a noun, a noun phrase or a continuous noun sequence thereof from a product display page of a virtual store on the network;
An extraction unit that extracts a product-specific phrase from the product name and / or product description as a subject word;
A dictionary database that stores dictionary information in which a compulsory word and a product category are associated is referred to, and when the compulsory word that matches the subject word exists, the product category associated with the compulsory word is A determination unit that identifies the product category indicated by the product name, wherein the one product category is uniquely identified by one compulsory word; and
The category determination program which makes a computer perform the output part which outputs the category information which shows the goods category specified by the said determination part. An acquisition unit for acquiring a product name and / or product description including a noun, a noun phrase or a continuous noun sequence thereof from a product display page of a virtual store on the network;
An extraction unit that extracts a product-specific phrase from the product name and / or product description as a subject word;
A dictionary database that stores dictionary information in which a compulsory word and a product category are associated is referred to, and when the compulsory word that matches the subject word exists, the product category associated with the compulsory word is A determination unit that identifies the product category indicated by the product name, wherein the one product category is uniquely identified by one compulsory word; and
A computer-readable recording medium that stores a category determination program that causes a computer to execute an output unit that outputs category information indicating a product category specified by the determination unit.

Images