KR100783752B1 - Category automatic matching method and system - Google Patents

Abstract

The present invention relates to a category automatic matching method and a category automatic matching system for making an optimum category matching for a product based on a query word excluding a stopword repeatedly used without a direct relation with a product requested for registration. According to the present invention, it is possible to provide a category automatic matching method and a category automatic matching system for reducing a matching error and performing a more accurate category matching by recommending a category for a product based on a query word created without a stopword.

Product Brokerage, Product Registration, Stopwords, Kinswords, Category Recommendations

Description

Category automatic matching method and category automatic matching system {CATEGORY AUTOMATIC MATCHING METHOD AND SYSTEM}

1 is a diagram schematically illustrating a network connection relationship between a product brokerage server, a shopping mall server, and a product registrant according to the present invention.

2 is a flowchart illustrating an automatic category matching method according to an embodiment of the present invention.

3 is a diagram illustrating an example of a category database according to the present invention.

4 is a flowchart illustrating an example of a category recommendation process according to the present invention.

5 is a flowchart illustrating another example of a category recommendation process according to the present invention.

6 is a diagram illustrating an example of a screen counting the number of internal identifiers for each internal category according to a query of the present invention.

7 is a block diagram showing the configuration of the automatic category matching system according to an embodiment of the present invention.

<Explanation of symbols for main parts of the drawings>

700: Category Automatic Matching System

701: category database 702: receiving unit

703: query creation unit 704: category recommendation unit

The present invention relates to a category automatic matching method and a category automatic matching system for making an optimum category matching for a product based on a query word excluding a stopword repeatedly used without a direct relation with a product requested for registration.

In general, websites that provide a search service can be easily managed and searched by storing them hierarchically by category. For example, products such as mobile phones, computers, and MP3 players may be stored in the home appliances / cell phones, computers / peripherals and home appliances / sounds / MP3 categories, respectively. In addition, not only product information but also information such as multimedia files, news, and knowledge may be classified and stored and managed by category. Typically categories have a hierarchy.

However, as the Internet is activated, the number of sites providing information through the Internet is rapidly increasing, and each site classifies the information into categories according to its own classification criteria. For example, in classifying the same coupling product, Site A may classify the product in the category Fashion / Jewelry / 18K and Site B may classify the same product in the category Accessories / Gold / 18K. Therefore, the information search user may be confused by information search because of unified category classification criteria, and product brokerage service sites that provide functions such as price comparison, which are recently activated, are provided from respective sites. Difficulty in classifying the product information to be classified by category.

Commodity brokerage service refers to a service that brokers shopping malls that have entered a given brokerage website to a buyer of goods through a contract with a service operating company. It provides a variety of information to the buyer of the product, such as comparing the price of each shopping mall and presenting.

The product brokerage service may introduce products of the shopping malls to the product buyers in various ways, and when the product buyer selects a specific product sold by the specific shopping mall, the product buyer may be mediated to the selected shopping mall to actually sell the product. Is done in affiliate shopping malls.

For such a product brokerage service, product information registered in each shopping mall should be registered in a product brokerage server. The product information and category registered in each shopping mall and the product information and category registered in the product brokerage server are also different for the same product. It may be different and it becomes a problem.

Due to the nature of the Internet that anyone can share information freely, it is very difficult for all sites to store information in the same category. Therefore, to automatically classify and store information by category in a product brokerage service site that provides a function such as price comparison, an automatic category matching method for properly recommending an internal category corresponding to information provided by another site, and The need for the system is urgently needed.

In particular, in category matching, a new category matching model is urgently required to determine a category for a product based on a query word excluding a keyword that is repeatedly included in a product name or the like as a stopword for the promotion / marketing of the product.

The present invention has been made to solve the above problems, category automatic matching method and category to reduce the matching error and more accurate category matching by recommending the category for the product based on the query word created without the stopword It is an object to provide an automatic matching system.

In addition, the present invention is a category automatic matching method and category that can be controlled so that products that are contrary to the breezes or do not match the nature of the product brokerage server to stop intermediary sales by stopping category recommendation for a query containing a keyword set as a banned words It is an object to provide an automatic matching system.

In addition, the present invention provides a category automatic matching method and category automatic method for effectively matching an appropriate category for an external identifier by creating a query related to an external identifier and recommending an internal category that is highly related to the created query word as a category corresponding to the external identifier. It is an object to provide a matching system.

The automatic category matching method according to the present invention for achieving the above object, in connection with the request for registration of the product, receiving an external identifier and the external category to which the external identifier belongs, and a stopword from the external identifier or external category ( Excluding a stopword, creating a query using an external identifier or an external category excluding the stopword, and recommending a category based on the query.

In addition, as a technical configuration of the present invention for achieving the above object, the category automatic matching system, in connection with the registration request of the product, a receiving unit for receiving an external identifier and the external category to which the external identifier belongs, and the external identifier or external A query word creation unit for excluding a stopword from a category, a query word using an external identifier or an external category excluding the stopword, and a category recommendation unit for recommending a category based on the query word; do.

Hereinafter, the automatic category matching method and the automatic category matching system will be described with reference to the accompanying drawings.

The stopword continuously used in the present specification may refer to an external identifier of a product name input by an advertiser when registering a product to an intermediate server, or a keyword repeatedly included in an external category to which the external identifier belongs. For example, the stopword may include words such as a discount sale, a special price, a special edition, and the like, which are frequently used as product names. Such a stopword may provide a cause of mismatching when matching a category when included in the query, so that the automatic category matching system of the present invention creates a query using an external identifier or an external category excluding the stopword.

The internal identifier and internal category may refer to an identifier and a category which are used by the broker itself in an intermediary server for the sale of goods registered in a predetermined shopping server, and the external identifier and the external category are used in the shopping server. It may mean an identifier and a category. In this case, the identifier is used to identify a product, a multimedia file, knowledge or news, and may be a concept including a brand name, a file name, a keyword of knowledge, a title of news, and the like.

For example, the internal identifier and the internal category may be identifiers and categories that are used by themselves within a predetermined product brokerage server, and the external identifier and the external category may be identifiers and categories used by shopping mall servers associated with the product brokerage server. Can be.

1 is a diagram schematically illustrating a network connection relationship between a product brokerage server, a shopping mall server, and a product registrant according to the present invention.

The category automatic matching method according to an embodiment of the present invention may be performed by a predetermined category automatic matching system, and the category automatic matching system may include a product brokerage server 100, a shopping mall server 110, and a product registrant 120. ) Can be operated in conjunction with

In particular, the automatic category matching system of the present invention may create a query word excluding the set stop word, and accurately match a category for a product newly requested to be registered using the query word created by excluding the stop word. For example, the automatic category matching system sets a stopword that is repeatedly included in new registration of a product as a stopword, and excludes a keyword corresponding to the set stopword when the query is made.

In addition, the category automatic matching system of the present invention may utilize a banned word to identify a product that is not allowed to sell in the mediation server 100, product name (external identifier) or product category (external) Category), if a banned word is included, a series of procedures for rejecting the product registration may be performed. For example, the category automatic matching system may prevent the brokerage of products that are inconsistent with the breeze, and may not be intermediated, and the product registration may be performed for a product including condoms and love gels as a prohibition and including the condoms and love gels as product names Matching) can be prevented.

First, the commodity brokerage server 100 may provide information about a product to be searched to a searcher or to facilitate searcher access to a shopping mall server 110 that actually sells a product to be purchased. It may refer to a commercial search server that supports a search operation. That is, the product brokerage server 100 searches for a shopping mall server capable of providing product information (eg, product specification, product price, information on a shopping mall, etc.) for a specific product requested by the user, and creates a search result. The created search results are arranged on the search result list and presented to the searcher.

The shopping mall server 110 is a kind of service server that substitutes advertisements for goods to be sold through a goods brokerage service by the goods brokerage server 100. For example, the shopping mall server 110 is connected to a searcher guided by the goods brokerage server 100. The server may refer to a server that substantially performs a process related to selling a product, such as selling a product, performing a payment according to selling a product, or delivering a sold product to a searcher.

That is, the shopping mall server 110 may expose more advertisements for its products or shopping malls to the product brokerage server 100 that can easily secure a plurality of searchers, thereby facilitating more searchers to access the connected products. You can expect a high purchase success rate for the products you sell based on many searchers. In addition, the shopping mall server may provide the searcher through the product mediation server 100 with a search result including the product information of the product sold by the shopping mall server in response to the search request of the searcher connected to the product mediation server 100. have.

The product registrant 120 may refer to a commercial internet user that allows an advertisement / sale agency for a product sold by the shopping mall server 110 or a product newly registered on the shopping mall server 110 to be performed through the product mediation server 100. Can be.

The automatic category matching system maintains a category database including an internal identifier (product identifier) and an internal category (product category) to which the internal identifier belongs, and the predetermined shopping mall server 110. Receive an external identifier (product identifier) and an external category (product category) of the product requested to be registered.

In addition, the category automatic matching system generates a query word used when matching a category by using the received external identifier or an external category to which the external identifier belongs. At this time, the category automatic matching system may create a query using only an external identifier (product identifier), or may create a query using an external identifier and an external category to which the external identifier belongs. The query creation described above may be performed by various methods, and any method of writing a query related to an external identifier, as in the present invention, is not considered to be beyond the scope of the technical idea of the present invention. For example, the category automatic matching system may create a query by removing spaces from an external identifier, create a query in space units from an external identifier, or create a query by stemming the external identifier. In addition, the category automatic matching system may create a query by removing spaces from an external identifier and an external category, create a query in the unit of a space in the external identifier and the external category, or create a query by stemming the external identifier and the external category. It may be. In this case, when the external category has a hierarchical structure, the automatic category matching system may use only some hierarchies of the external category for query generation. For example, if the external category received in the product registration request is' Off-Market @ Life / Health / Life / Bathroom @ Cleaning / Laundry, the automatic category matching system will ensure that only the lowest level category, Cleaning / Laundry, is used for query creation. Can be.

Thereafter, the category automatic matching system recommends one or more of the internal categories stored in the category database as a category corresponding to the external identifier by using the written query. At this time, the category automatic matching system may automatically match the most suitable one internal category to the category corresponding to the external identifier.

In general, when identifiers such as product identifiers are similar, the categories to which the identifier belongs are often similar. Thus, a category database is searched using a query for an external identifier as in the category automatic matching method according to an embodiment of the present invention. By recommending a category including many internal identifiers having high similarity with the external identifier, a category for the external identifier such as a new product identifier can be set quickly and accurately.

2 is a flowchart illustrating an automatic category matching method according to an embodiment of the present invention.

The category automatic matching method according to the present embodiment may be performed by a predetermined category automatic matching system, and as described above, the category automatic matching system may operate on the commodity brokerage server and the shopping mall server or in conjunction with them. .

2, the automatic category matching system according to the present invention maintains a category database (S210).

3 is a diagram illustrating an example of a category database according to the present invention.

As shown in FIG. 3, the category database includes an internal identifier 301 used in the commodity brokerage server 100, an internal category 302 to which each internal identifier 301 belongs, and a search field 303. can do.

The internal identifier 301 may be, for example, a product identifier, and may refer to an identifier (ID) used in the product mediation server 100 to distinguish each product model. The internal category 302 is a category that is classified by itself within the product mediation server 100 and may mean a category within the product mediation server 100 to which the internal identifier 301 belongs. The internal category 302 may be stored in the form of an identifier representing the category. The search field 303 is a field to be searched when searching for a recommendation category using a query. The search field may include various information related to each product model, information such as brand name, model name, item name, keyword, and the like. For example, the search field may include only information about the internal identifier 301 or may include both information about the internal identifier 301 and information about the internal category 302.

In FIG. 3, the internal category 302 is recorded in the category database as a separate data field from the search field 303. However, this is merely an example. It may be included and recorded in the search field 303 instead of a separate data field.

In addition, the category automatic matching system receives an external identifier and an external category to which the external identifier belongs from the outside (S220).

For example, the external identifier may be a product identifier. The external identifier and the external category are used in the shopping mall server 110 other than the product mediation server 100. Each shopping mall server 110 may use different identifiers for the same product, and different for the same product. You can use categories. As described above, since the categories of each shopping mall server 110 are different from each other, they cannot be registered as such as the product brokerage server 100, and thus, the product brokerage server 100 needs to set and register an appropriate category for various products. .

Thereafter, the category automatic matching system determines whether there is an internal identifier corresponding to the received external identifier (S230).

That is, for example, when both the external identifier and the internal identifier are product identifiers, the category automatic matching system corresponds to an external identifier received from the shopping mall server 110 among the internal identifiers used in the product mediation server 100. Determine if there is any.

If there is an internal identifier corresponding to the received external identifier (Yes direction in S230), the category automatic matching system selects a category to which the internal identifier belongs, as a category corresponding to the received external identifier (S240).

When the external identifier is related to the new product and there is no internal identifier corresponding to the received external identifier among the internal identifiers, the category automatic matching system creates a query related to the received external identifier (S250). In this case, the automatic category matching system excludes the stopword from the external identifier or the external category to which the external identifier belongs, and creates a query using the external identifier or the external category from which the stopword is excluded. As described above, the stopword is not directly related to the product, but may refer to a keyword that the product registrant 120 repeatedly includes in an external identifier or an external category to promote product sales. As an example of the above stopword, in this embodiment, a discount sale, a super sale, etc., which are inserted into a product name (external identifier), can be set as a stopword in order to promote the product.

That is, the category automatic matching system creates a query word without the stopword.

In the creation of the query, the category automatic matching system can create the query using only external identifiers. When creating a query using only an external identifier, the category automatic matching system may reduce the time required for creating the query.

In addition, the category automatic matching system may create a query using an external identifier and an external category. When creating a query using an external identifier and an external category, the category automatic matching system can more effectively calculate the similarity between the external identifier and the information in the category database. For example, when the external category has a three-level hierarchical structure, the automatic category matching system may create a query using the lowest layer (third layer) and the external identifier of the external category. As another example, the category automatic matching system may create a query using all layers (first to third layers) and external identifiers of the external category.

The automatic category matching system for generating the query may process the external identifier or the external category in various ways to perform the query generation. For example, a category automatic matching system may create a query by removing spaces from an external identifier and / or an external category, or create a query by extracting keywords in space units from the external identifier and / or external category, or create an external identifier and / or Alternatively, you can write queries by stemming the external categories. In addition, the category automatic matching system may generate a query by sequentially performing the plurality of query generation methods described above until a predetermined condition is satisfied.

For example, if a query is created by removing a space from an external identifier and / or an external category, and a search result of a desired condition is not obtained even if a search for an internal identifier described below is performed using the created query, a category automatic matching system May generate a query by extracting keywords in a space unit from an external identifier and / or an external category, and search for an internal identifier using the created query again. In addition, if the desired search result is not obtained even by the query word created by extracting the keyword by the spacing unit, a category automatic matching system forms a query word by morphological analysis of the keyword, and searches for an internal identifier using the created query word. Can be done again.

Next, the category automatic matching system recommends one or more of the internal categories stored in the category database as the category corresponding to the external identifier by using the created query (S260).

In step S260, the category automatic matching system may recommend only one category that is most suitable. At this time, the category automatic matching system may automatically set the recommended category to a category corresponding to the external identifier.

Also, the category automatic matching system may set the number of recommended categories according to the setting. At this time, the category automatic matching system may select the recommended category as one of the categories recommended by the product registrant 120 registering an external identifier such as a product identifier.

In particular, in the recommendation step of the category (S260), the automatic category matching system may determine whether the query word includes a banned word, and stop the category recommendation when the query word includes the banned word. As mentioned above, banned words are based on identifying commodities that are allowed to be traded. For example, the operator of the system may use keywords such as condoms, love gels, drugs, etc. that are contrary to morals or are not allowed to sell. Can be set as a banned word. That is, the category automatic matching system does not perform category automatic matching on products related to the creation of a query word including a banned word.

4 is a flowchart illustrating an example of a category recommendation process according to the present invention.

Referring to FIG. 4, the category automatic matching system searches for a category database using a written query and extracts internal identifiers having a similarity level or more with respect to the query (S410).

The category automatic matching system can measure the similarity in various ways. For example, the category automatic matching system can measure similarity by applying the Optimizing the Weight of Added Terms in Query Expansion (241-13 of the 9th Proceedings of the Korean Information Management Society Conference). See page 246).

As an example, the category automatic matching system in step S410 may extract internal identifiers whose similarity is measured to be 0.4 or more.

Also, in step S410, the category automatic matching system may extract only the internal identifiers whose price corresponding to the internal identifier is within a predetermined range from the price corresponding to the external identifier. For example, when the price corresponding to the external identifier is 5,900 won, the category automatic matching system may extract the internal identifier within the price range of 4,700 won to 7,080 won (price range ㅁ 20%). Typically, similar products in the same / similar categories often have similar price ranges. Thus, in this embodiment, by limiting the price range, the automatic matching system can quickly and simply extract internal identifiers that are related to external identifiers. To help.

Thereafter, the category automatic matching system recommends one or more of the internal categories as the categories corresponding to the external identifiers in the order of increasing the number including the extracted internal identifiers (S420).

5 is a flowchart illustrating another example of a category recommendation process according to the present invention.

In the example shown in FIG. 5, the category database exemplifies a search field generated by using an internal identifier and an internal category.

Referring to FIG. 5, the category automatic matching system searches a category database with a written query and extracts internal identifiers corresponding to search fields whose similarity with the query is greater than or equal to a predetermined level (S510).

As such, the category automatic matching system may recommend categories more appropriate to the external identifier by extracting the internal identifiers based on the similarity between the generated search field and the created query using the internal identifier and the internal category.

In measuring the similarity, the category automatic matching system can measure the similarity in various ways. For example, in operation S510, the category automatic matching system may extract internal identifiers having a similarity degree of 0.4 or more between the search field and the query word.

In extracting the internal identifier in step S510, the category automatic matching system may target only the internal identifiers whose price corresponding to the internal identifier is within a predetermined range from the price corresponding to the external identifier.

In addition, the category automatic matching system recommends one or more of the internal categories as a category corresponding to the external identifier in the order of increasing the number including the extracted internal identifiers (S520).

In the category recommendation in step S260 illustrated in FIG. 2, the automatic category matching system may generate the query in a different way according to the external identifier and the server transmitting the external category. That is, when the external identifier is a product identifier and the server transmitting the external identifier and the external category is a shopping mall server, the query automatic matching system may create a query word for each shopping mall in a different manner. For example, for the A shopping mall server, the category automatic matching system creates a query using only a product identifier (external identifier), and for the B shopping mall server, the category automatic matching system uses the lowest hierarchy and product identifier (external identifier) of the external category. ), And a category automatic matching system can generate a query using the entire external category and the product identifier. Accordingly, the automatic category matching system according to the present invention can set an appropriate querying method according to the external identifier and the characteristics of the server that transmits the external category, so that the category can be more effectively recommended.

In addition, in recommending a category in step S260 illustrated in FIG. 2, the automatic category matching system may vary the similarity with the query word based on the external identifier and the server transmitting the external category. For example, for the A shopping mall server, the category automatic matching system creates an internal identifier based on the similarity between the query word and the internal identifier. For the B shopping mall server, the category automatic matching system generates the internal identifier and the internal category. The internal identifier can be created based on the similarity between the generated search fields.

6 is a diagram illustrating an example of a screen counting the number of internal identifiers for each internal category according to a query of the present invention.

In FIG. 6, for the product registration, an external identifier 610 called 'discount sale food waste bin (pretty design) / sparkly / super special food waste bin' and 'open market @ life / health @' from an external shopping mall site (interpark) An external category 620, 'Living / Bathroom @ Cleaning / Laundry', is received.

First, the category automatic matching system excludes the stopword from the received external identifier or external category. That is, the category automatic matching system excludes stopwords such as keywords 'discount sale' and 'super special' set by the operator of the system from the external identifier or the external category so that accurate category matching can be performed for the product requested for registration. Afterwards, the category automatic matching system uses an external identifier 'food waste bin (pretty design) / slightly on the sink / food waste bin' and an external category 'open market @ life / health @ life / bathroom @ cleaning / laundry' excluding the stopword. To create a query.

In the creation of the query, the category automatic matching system first uses the external identifier 'food waste bin (beautiful design) / slightly on the sink / food waste bin' which excludes the stopword. Recycle Bin '631 can be created. Looking at the search result 641 which extracts the internal identifier for each internal category based on the query term 631 written using only the external identifier, the automatic category matching system matches the query term 631 within the internal category having the category ID '07020404'. It can be seen that the internal identifiers having a similarity of 0.4 or more are counted as four.

Accordingly, for the query `` food waste bin beautiful design brightly food trash on the sink '' (631), the automatic category matching system, 'life / kitchen @ kitchen supplies @ kitchen supplies of category ID' 07020404 'that counted the most internal identifier @Food waste bin is recommended as a category.

In addition, the category automatic matching system uses the external identifier 'food waste bin (pretty design) / brightly on the sink / food waste bin' and the 'lowest class' of the external categories, 'stopping / cleaning', which excludes the stopword. Food Trash Can Pretty Design A food trash can on the sink '(632) can be created. Looking at a search result 642 of extracting an internal identifier for each internal category based on the query 632 written using the external identifier and the lowest hierarchy of the external category, the category automatic matching system includes: i) an internal category having a category ID of '07050003'. Counting 409 internal identifiers having a similarity level of at least 0.4 as the query term 632, and ii) counting internal identifiers having a similarity level of at least 0.4 level with the query term 632 at 237 counts within the internal category of the category ID '02041000', i). In the internal category of the category ID '00030700', it can be seen that the number of internal identifiers having a similarity of 0.4 or more to the query word 632 is counted as 111.

Accordingly, for the query 'cleaning laundry food waste bin beautiful design brightly the food waste on the sink' (632), the automatic category matching system, the living / kitchen @ laundry / cleaning of category ID '07050003' that counted the most internal identifier 'Cleaning bin' / trash box is recommended as a category.

In addition, the category automatic matching system is an external identifier 'food waste bin (pretty design) / brightly on the sink / food waste bin' and the external category 'open market @ life / health @ life / bathroom @ cleaning / laundry' Using the questionnaire 'open market life health life bathroom cleaning laundry food waste bin beautiful design bright food sink on the sink' (633) can be written. Looking at the search result (643) of extracting the internal identifier for each internal category based on the query word 633 created using the external identifier and the external category, the automatic category matching system iii) in the internal category of category ID '07050003', 409 counts internal identifiers having a similarity level of 0.4 or more to the query term 633, and ii) counts 322 internal identifiers having a degree of similarity of 0.4 or more counts to the query category 633 within the internal category of the category ID '02041000', and i) category ID ' In the internal category 11010505 ', it can be seen that the number of internal identifiers having a similarity of 0.4 or more to the query word 633 is counted as 252.

Accordingly, for the query 'open market living health living bathroom cleaning laundry food waste bin beautiful design brightly food waste bin on the sink' (633), the category automatic matching system is the 'life of category ID' 07050003 'that counted the most internal identifiers / Kitchen washing / cleaning supplies cleaning supplies trash / trash cans as a category.

Accordingly, the automatic category matching system according to the present invention can perform accurate category matching on the product requested to be registered based on an external identifier excluding a stopword or a query written using an external category.

As another embodiment of the present invention, the automatic category matching system may determine whether a query includes a banned word, and stop recommending a category when the query includes a banned word. For example, if the written query includes condoms, love gels, drugs, etc., included in the banned words, the category automatic matching system generates a registration rejection signal relating to the rejection of the brokerage for the product and stops recommending the category. Can be.

7 is a block diagram showing the configuration of the automatic category matching system according to an embodiment of the present invention.

Referring to FIG. 7, a category automatic matching system 700 according to an embodiment of the present invention includes a category database 701, a receiver 702, a query word generator 703, and a category recommender 704.

The category database 701 includes an internal identifier and an internal category to which the internal identifier belongs. At this time, the internal identifier and the internal category are identifiers and categories which are used by themselves in a predetermined product intermediation server, and are distinguished from those of the external identifier and the external category which are the identifiers and categories used by the shopping mall server. In this case, the external identifier / internal identifier may be used to identify a product, a multimedia file, knowledge, or news, and may be a concept including a brand name, a file name, a keyword of knowledge, a title of news, and the like.

The category database 701 may further include a search field as described with reference to FIG. 3. The internal category may be stored in the form of an identifier indicating the category. The search field is a field to be searched when searching for a recommendation category using a query. The search field may include various information related to each product model, information such as brand name, model name, item name, keyword, and the like. For example, the search field may include only information about an internal identifier, or may include both information about an internal identifier and information about an internal category.

The reception unit 702 receives an external identifier to which an external identifier belongs to and an external category to which the external identifier belongs in relation to a product registration request.

For example, the receiver 702 may receive an external identifier and an external category from a shopping mall server.

The query creating unit 703 creates a query related to an external identifier. In particular, the present embodiment excludes a stopword from an external identifier or an external category, and creates a query using the external identifier or the external category excluding the stopword. do.

In query generation, the query generation unit 703 may generate a query using only an external identifier from which the stopword is excluded. When creating a query using only an external identifier, the query preparing unit 703 may reduce the time required to prepare the query. In addition, the query generator 703 may generate a query using an external identifier and an external category. As an example, in the case of an external category having a three-level hierarchical structure, the query generation unit 703 may create a query using an external identifier and a lowest hierarchy of the external category. As another example, in the case of an external category having a three-level hierarchical structure, the query generator 703 may generate a query using an external identifier and all of the top, middle, and bottom hierarchies of the external category.

In preparing the query, the query generator 703 may process the external identifier or the external category in various ways to perform the query. For example, the query generator 703 may create a query by removing a space from an external identifier and / or an external category, or create a query by extracting keywords in a space unit from an external identifier and / or an external category, or an external identifier. And / or stemming the external category to create a query. In addition, the query generation unit 703 may sequentially generate the query by performing the plurality of query generation methods described above until a predetermined condition is satisfied.

For example, if a query is created by removing spaces from an external identifier and / or an external category, and a search result of a desired condition is not obtained even if a search for an internal identifier described later is performed using the created query, the query creation unit ( 703 may generate a query by extracting a keyword in units of spaces from an external identifier and / or an external category, and perform a search for an internal identifier using the created query. In addition, if the desired search result is not obtained even by the query word created by extracting the keyword in the spacing unit, the query word creation unit 703 forms the query word by morphological analysis of the keyword, and uses the generated query word in the internal identifier. You can perform the search again.

The category recommender 704 recommends one or more of the internal categories stored in the category database 701 as a category corresponding to the external identifier using the written query.

In category recommendation, the category recommender 704 may recommend only one category that is most suitable. In this case, the category recommender 704 may automatically set one recommended category to a category corresponding to the received external identifier.

In addition, the category recommendation unit 704 may recommend a plurality of categories according to the setting. In this case, the category recommending unit 704 may select one of the categories recommended by the product registrant registering the external identifier of the product identifier as the recommendation category.

In particular, the category recommendation unit 704 according to the present invention stops category recommendation for a query including a keyword set as a banned word, so that products which are contrary to the morals or do not conform to the characteristics of the product mediation server are not intermediated. Can be controlled.

The category recommendation unit 704 may operate according to the operation flow described with reference to FIGS. 4 and 5.

The category recommender 704 searches the category database 701 with the written query to extract internal identifiers having a predetermined level or more similar to the created query, and includes one of the internal categories according to the order of the number of the internal identifiers. The above can be recommended as a category corresponding to the external identifier. In addition, the category recommendation unit 704 may extract only the internal identifiers whose price corresponding to the internal identifier is between the upper limit and the lower limit of the predetermined range from the price corresponding to the external identifier.

Embodiments of the invention include a computer readable medium containing program instructions for performing various computer-implemented operations. The computer readable medium may include program instructions, local data files, local data structures, or the like, alone or in combination. The media may be those specially designed and constructed for the purposes of the present invention, or they may be of the kind well-known and available to those having skill in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical recording media such as CD-ROMs, DVDs, magnetic-optical media such as floppy disks, and ROM, RAM, flash memory, and the like. Hardware devices specifically configured to store and execute the same program instructions are included. The medium may be a transmission medium such as an optical or metal wire, a waveguide, or the like including a carrier wave for transmitting a signal specifying a program command, a local data structure, or the like. Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like.

While specific embodiments of the present invention have been described so far, various modifications are possible without departing from the scope of the present invention. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be defined not only by the claims below, but also by those equivalent to the claims.

As can be seen from the above description, according to the present invention, a category automatic matching method and a category automatic matching method to reduce a matching error and perform a more accurate category matching by recommending a category for a product based on a query word created without a stopword. A matching system can be provided.

Also, according to the present invention, by automatically stopping category recommendation for a query including a keyword set as a banned word, the category automatic matching method can be controlled so as not to sell a product that is contrary to the morals or does not meet the characteristics of the product brokerage server. And a category automatic matching system.

In addition, according to the present invention, a category automatic matching method for effectively matching the appropriate category for the external identifier by creating a query associated with the external identifier, and recommending an internal category that is highly related to the created query as a category corresponding to the external identifier; A category automatic matching system can be provided.

Claims (17)

In the category automatic matching method, Receiving an external identifier and an external category to which the external identifier belongs, in association with a registration request of a product; Excluding a stopword from the external identifier or external category; Creating a query using an external identifier or an external category excluding the stopword; Determining whether a queried word is included in the written query word; And As a result of the determination, if a banned word is not included, recommending a category based on the query word Category automatic matching method comprising a. The method of claim 1, And the stopword is a keyword which is repeatedly included in the external identifier or the external category at every registration request, without being directly related to the product. The method of claim 1, Stopping the recommendation of the category when a banned word is included as a result of the determination The category automatic matching method further comprising. The method of claim 1, And the banned word is a word associated with generating a registration rejection signal for the registration request. The method of claim 1, The step of recommending a category, Maintaining a category database comprising an internal identifier and an internal category to which the internal identifier belongs; And Recommending one or more of the internal categories stored in the category database using the query as a category corresponding to the external identifier Category automatic matching method comprising a. The method of claim 5, The step of recommending in a category corresponding to the external identifier, Searching the category database with the query to extract internal identifiers having a degree of similarity with the query at least a predetermined level; And Recommending one or more of the internal categories as a category corresponding to the external identifier in the order of increasing the number including the extracted internal identifiers. Category automatic matching method comprising a. The method of claim 5, The category database further includes a search field generated using the internal identifier and the internal category, The step of recommending in a category corresponding to the external identifier, Searching the category database and extracting internal identifiers corresponding to the search fields having a predetermined level or more similar to the query word; And Recommending one or more of the internal categories as a category corresponding to the external identifier in the order of increasing the number including the extracted internal identifiers. Category automatic matching method comprising a. The method according to claim 6 or 7, Extracting the internal identifiers, Wherein the price corresponding to the internal identifier extracts only internal identifiers that fall between an upper limit of the first range from a price corresponding to the external identifier and a lower limit of a second range from a price corresponding to the external identifier. Category automatic matching method. The method according to claim 6 or 7, The step of recommending in a category corresponding to the external identifier, And automatically matching the largest number including the extracted internal identifiers among the internal categories to a category corresponding to the external identifier. The method of claim 1, The step of writing a query, And automatically generating the query using only external identifiers excluding the stopword. The method of claim 1, The step of writing a query, And automatically generating the query using an external identifier and an external category excluding the stopword. The method of claim 1, The external category consists of hierarchical fields, The step of writing a query, And generating the query using an external identifier excluding the stopword and a field of some layer of the external category. The method of claim 1, The step of recommending a category, And extracting the query word in different ways according to the external identifier and the server transmitting the external category. The method of claim 5, And the internal identifier and the external identifier are both product identifiers. A computer-readable recording medium in which a program for executing the method of any one of claims 1 to 7 and 10 to 14 is recorded. In the category automatic matching system, A receiving unit which receives an external identifier and an external category to which the external identifier belongs, in association with a registration request of a product; A query word generator to exclude a stopword from the external identifier or the external category, and to create a query using the external identifier or the external category from which the stopword is excluded; And A category recommending unit that determines whether a queried word is included in the created query word and recommends a category based on the query word when the queried word is not included as a result of the determination; Category automatic matching system comprising a. The method of claim 16, The category recommendation part The category automatic matching system, characterized in that to stop the category recommendation, if a banned words included.

Images