JP7108740B2 - Product Category Extraction Method for Shopping Search - Google Patents

Description

TECHNICAL FIELD The present disclosure relates to a product category extraction method and system for shopping search, specifically extracting a product category from a user query or search query input for shopping search, and determining product search ranking. It relates to a product category extraction method and system that can be reflected in

Recently, various services such as shopping, games, movies, and music are provided online in the Internet and mobile environment. In particular, online shopping is used by a large number of users because there is no time limit for purchasing products and it is possible to select and purchase products without directly visiting a product store.

Due to the convenience of online shopping, the number of users who use online shopping is increasing, and the number and types of products sold or handled online are increasing exponentially. Accordingly, there is a problem that it is difficult for the user to search for a product that matches his/her intention among various products sold online.

On the other hand, when a user connects to an online shopping system and inputs a query to search for a desired product, the product search results including product information related to the query and connection information regarding purchasers are displayed. I can provide a list. Since such an online shopping system simply compares a user query with product information in a product database to extract search results, it is difficult to derive product search results that match the intention of the user included in the query.

Korean Patent No. 10-1017016

The present disclosure has been made to solve the above problems, and provides a product category extraction method, a computer program, and an apparatus (system) for shopping search.

In addition, the present disclosure provides a method and apparatus (system) capable of more accurate product category extraction by using a statistically based classification model for product category extraction for shopping search.

The present disclosure can be embodied in a variety of ways including methods, apparatus (systems) or computer programs.

According to an embodiment of the present disclosure, a product category extraction method for shopping search performed by a computer system comprises: receiving a user query; extracting product search results for the user query from a product database; According to the classification model, calculating the relevance score of the user query and the plurality of product categories, and adjusting the product search ranking according to the product search result based on the calculated relevance score.

According to another embodiment of the present disclosure, a product category extraction method for shopping search performed on a user terminal comprises: receiving a user query; calculating a score; and adjusting a product search ranking for a user query based on the calculated relevance score.

According to still another embodiment of the present disclosure, there is provided a computer program for executing the product category extraction method for shopping search described above on a computer.

According to yet another embodiment of the present disclosure, a learning method for product category extraction for shopping search performed on a computer system comprises a plurality of first user queries and a plurality of queries associated with the plurality of first user queries. generating a plurality of first learning data including the first product category of and, based on the plurality of first learning data, determining the relevance score of one or more product categories corresponding to the second user query training a statistically-based classification model to compute.

According to still another embodiment of the present disclosure, there is provided a computer program for executing the learning method for product category extraction for shopping search described above on a computer.

According to yet another embodiment of the present disclosure, a product category extraction system for shopping search includes a communication module for receiving a user query, a memory, and computer readable instructions contained in the memory coupled to the memory. at least one processor configured to execute At least one processor extracts product search results for user queries from a product database, calculates relevance scores for user queries and a plurality of product categories using a statistically-based classification model, and selects products based on the calculated relevance scores. Adjust the product search ranking according to search results.

According to various embodiments of the present disclosure, it is possible to output to the user terminal a search result reflecting the product category intended by the user based on the user query. In addition, according to various embodiments of the present disclosure, by displaying products belonging to a product category related to a user query at the top of search results, it is possible to improve the click rate and purchase rate of users for the products. can.

According to various embodiments of the present disclosure, improving the accuracy of the relevance of product categories to user queries by using a statistically based classification model to calculate relevance scores of multiple product categories to user queries. can be made

According to various embodiments of the present disclosure, by dividing a user query into units of syllables and/or morphemes and calculating relevance scores for product categories using a statistically-based classification model, it is possible to It is also possible to improve the accuracy of calculating the relevance score of the product category for queries. Also, in various embodiments of the present disclosure, matching coverage of learning data is maximized, and the accuracy of calculating user queries and relevance scores of multiple product categories can be further improved.

According to various embodiments of the present disclosure, the learning data used for learning a statistically-based classification model that calculates the relevance scores of multiple product categories for a user query may be integrated search log data, shopping or product search log data, or the like. It can be generated from various search log data such as In addition, learning data can be automatically generated based on user queries, search results, and user actions in response to search results among search log data, so that learning data for the statistical-based classification model can be efficiently secured. .

Embodiments of the present disclosure will be described with reference to the following accompanying drawings. Here, like reference numbers indicate like elements, but are not so limited.
FIG. 5 is a diagram illustrating a screen displaying search results reflecting product categories extracted from a user query by a product category extraction method for shopping searches according to an embodiment of the present disclosure; FIG. 5 is a diagram illustrating a screen displaying product search rankings adjusted based on the relevance scores of product categories to user queries calculated by a product category extraction method for shopping searches according to an embodiment of the present disclosure; 1 is a schematic diagram illustrating a system in which a plurality of user terminals and a product category extraction server are communicatively coupled to provide a product category extraction service for shopping search according to an embodiment of the present disclosure; FIG. 2 is a block diagram showing internal configurations of a user terminal and a product category extraction server according to an embodiment of the present disclosure; FIG. FIG. 10 is a diagram showing a configuration for calculating a product category related score for a user query using a statistically-based classification model according to an embodiment of the present disclosure; FIG. 4 is a diagram showing a configuration for calculating a relevance score using a classifier including a preprocessor and a statistically-based classification model according to an embodiment of the present disclosure; FIG. 5 is a diagram illustrating an example of generating learning data of a statistically-based classification model for product category extraction from various search results according to an embodiment of the present disclosure; 4 is a flowchart illustrating a product category extraction method for shopping search performed by a product category extraction server according to an embodiment of the present disclosure; 4 is a flowchart illustrating a product search ranking adjustment method for shopping search according to an embodiment of the present disclosure; 4 is a flowchart illustrating a product category extraction method for shopping search performed by a user terminal according to an embodiment of the present disclosure; 4 is a flowchart illustrating a statistically-based classification model learning method for product category extraction according to an embodiment of the present disclosure;

Hereinafter, specific contents for carrying out the present disclosure will be described in detail based on the accompanying drawings. however,
In the following description, specific descriptions of well-known functions and configurations are omitted when they may unnecessarily obscure the subject matter of the present disclosure.

In the accompanying drawings, the same or corresponding components are provided with the same reference numerals. Also, in the description of the embodiments below, duplicate descriptions of the same or corresponding components may be omitted. However, the omission of a description of an element should not be construed as omitting such element from being included in an embodiment.

Advantages and features of the disclosed embodiments, and the manner in which they are achieved, will become apparent with reference to the embodiments described below with reference to the accompanying drawings. This disclosure may, however, be embodied in a variety of different forms and should not be construed as limited to the embodiments set forth below. These examples are provided merely so that this disclosure will be complete and will enable those of ordinary skill in the art to accurately recognize the categories of the invention.

A brief description of terms used herein and a specific description of the disclosed embodiments are provided. The terms used in this specification have been selected as common terms that are currently widely used as much as possible while considering the function in this disclosure, but this is not the intention or precedent of an engineer engaged in the related field. , may change due to the emergence of new technologies. Also, certain terms may be arbitrarily chosen by the applicant, the meaning of which will be detailed in the description of the invention. Accordingly, terms used in this disclosure should be defined based on their meanings and overall context of this disclosure, rather than simply terminology.

In this specification, singular expressions may include pluralities and plural expressions may include the singular unless the context clearly dictates otherwise.

Throughout the specification, when a part 'includes' a certain component, this does not exclude other components, and may also include other components, unless specifically stated to the contrary. means you can.

Also, the term 'module' or 'unit' used in the specification means a software or hardware component, and the 'module' or 'unit' performs a certain role. However, 'module' or 'unit' is not meant to be limited to software or hardware. A 'module' or 'portion' may be configured to reside on an addressable storage medium and configured to run on one or more processors. Thus, by way of example, a 'module' or 'part' is defined as components such as software components, object-oriented software components, class components, task components, as well as processes, functions, categories, procedures, subroutines, programs. It may include at least one of a segment of code, driver, firmware, microcode, circuitry, data, database, data structure, table, array or variable. Components and 'modules' or 'units' may be combined with a smaller number of components and 'modules' or 'units', or may be combined with additional components and 'modules' or 'units'. ' can be further separated into

According to one embodiment of the present disclosure, a 'module' or 'unit' may be embodied in a processor and memory. 'Processor' should be interpreted broadly to include general purpose processors, central processing units (CPUs), microprocessors, digital signal processors (DSPs), controllers, microcontrollers, state machines, and the like. In some circumstances, a 'processor' may also refer to an application specific semiconductor (ASIC), programmable logic device (PLD), field programmable gate array (FPGA), or the like. A 'processor' may be, for example, a combination of a DSP and a microprocessor, a combination of multiple microprocessors, a combination of one or more microprocessors combined with a DSP core, or any other such configuration. can also be referred to as a combination of processing devices. Also, 'memory' should be interpreted broadly to include any electronic component capable of storing electronic information. 'Memory' includes RAM (Random Access Memory), ROM (Read Only Memory), NVRAM (Non-Volatile Random Access Memory), PROM (Programmable Read-Only Memory), EPROM (Erasable Programmable Read-Only Memory), EEPROM ( It can also refer to various types of processor-readable media such as electrically erasable programmable reda-only memory), flash memory, magnetic or optical data storage devices, registers, and the like. Memory is said to be in electronic communication with a processor when the processor can read information from, or record information in, the memory. Memory integrated with the processor is in electronic communication with the processor.

In the present disclosure, 'user query' can include, but is not limited to, text (eg, words, phrases, sentences, etc.) that can be entered and searched using a search engine. For example, a user query may include text that a user inputs through an input UI of a search program such as a web browser, search application, or shopping application, or text that the user can select by clicking from a query list provided by the search program. can contain In other examples, user queries are not limited to text, but can include text or information that can be extracted from multimedia content such as images, videos, sounds, voices, and the like.

In the present disclosure, 'product category' may indicate the type, type, category, or similar product information determined to classify the product. The product category is divided into major, medium, small, and small categories according to the size of the category to which the product belongs. In addition, among the product categories, a major category can include one or more middle categories, and a middle category can include one or more middle and small categories or small categories. For example, 'fashion', which is a large category of fashion products, can include medium categories such as 'men's fashion' and 'ladies' fashion', and 'ladies' fashion' is 'dress' and 'dress'. Minor categories such as 'knits/sweaters' can be included. It is determined to which of the above-mentioned graded product categories the user query for product search is related, and if a product belonging to the graded product category is searched, it is close to the user's intention included in the user query. Product search results can be derived.

In this disclosure, a 'statistics-based classification model' is a classifier trained to classify new unlabeled information based on its relevance to known labeled data, or Classification models can be presented generically. For example, statistically based classification models include language model, SVM (support vector machine), maximum entropy model, CRF model (conditional random field model), FFNN (feed-forward neural network), RNN (recurrent neural network), GRU (gated recurrent unit), LSTM (long short-term memory), transformer model, BERT model (bidirectional encoder representations from transformers model), etc. It can be any one of a neural network model.

In this disclosure, 'embedding' can refer to the technique of transforming a categorical or discrete variable into a continuous vector. Embedding can be used to reduce the dimensionality of information and derive meaning from categorical variables, such as user queries, by transforming them into continuous vectors. Also, if a user query is represented as a vector in an embedded vector space, neighboring (spatially close) vectors, etc. on that vector space can share similar or related properties. For example, a method of embedding user queries in a vector space can be implemented by training an artificial neural network.

FIG. 1 is a diagram illustrating a screen displaying search results 130 reflecting product categories 120 extracted from a user query 110 by a product category extraction method for shopping searches according to an embodiment of the present disclosure. As shown in the figure, when a user inputs a user query 110 into a search window of a shopping site, search results 130 reflecting product categories 120 extracted from the user query 110 can be output. The search results 130 include images and names 132, 136 and product categories 134, 138 that indicate the searched product. For example, when a user inputs 'knit dress' as a search query, the product category 120 of 'dress', which is the product included in the query (for example, fashion as a major category, women's fashion as a middle category, and minor category). The search ranking may be adjusted so that the products 132 , 136 related to the one piece) are displayed at the top of the search results 130 . That is, in the search result 130, the products 132 and 136, which are 'dress 1' and 'dress 2' corresponding to dress products, are preferentially displayed at the top of the search result 130 among various women's fashion products. obtain. Therefore, through the search result 130, the user can easily and quickly check information on the products 132 and 136 related to the product category 120 that can be extracted from the product search query input by the user. In addition, the user can select or click any one of the products 132 and 136 as necessary to obtain more specific information about the selected product, or perform online purchases.

In one embodiment, product category 120 may include category groups 122, 124, 126 divided into a predetermined number of tiers, such as major, minor, minor, and minor categories. For example, as shown in FIG. 1, the first tier category group, Category 1 (122), may include 'Fashion', 'Maternity/Childcare', etc., which represent a wide range of products. In addition, category 2 (124), which is a second category group showing a narrower range of products than the first category group, may include 'ladies' fashion', 'men's fashion', and the like. Also, category 3 (126), which is a third tier category group that indicates a narrower range of products than the second tier category group, may include 'dress', 'blouse/shirt', and the like. Category 3 (126) can then be subcategories of category 2 (124), and category 2 (124) can be subcategories of category 1 (122).

In one embodiment, when a user query 110 is input from a search window of a shopping site or an integrated search site, each of the category groups 122, 124, and 126 included in the graded category groups 122, 124, and 126 for the user query 110 is The relevance or relevance probability of product categories (eg, 'fashion', 'maternity/childcare', 'women's fashion', 'men's fashion' dress', 'blouse/shirt', etc.) can be calculated. In the product category extraction method, when the user query 110 is 'knit dress', not only a plurality of product categories related to fashion, but also a plurality of other product categories such as home appliances, fashion miscellaneous goods, and life and health products. It is also possible to calculate the relevance for When the user query 110 is 'knitted dress', the relevance to large, medium and small categories (eg, 'fashion', 'women's fashion', and 'dress') related to one-piece women's fashion. Although it is calculated to be the highest, the relevance to other product categories of fashion and product categories related to other types of products can be calculated to be low.

On the other hand, a product search result 130 for the user query 110 is obtained from the product database. A product search based on the user query 110 can be executed by extracting product information highly relevant to the user query 110 from product information stored in the product database. In the product search results 130 extracted in this manner, product information highly relevant to the user query 110 may be displayed at a higher rank than product information less relevant. In addition, it is possible to adjust the product search ranking according to the product search result 130 based on the relevance (or related score) of the product category to the user query 110 calculated as described above. Accordingly, when the user query 110 is 'knit dress', information on products 132 and 136 belonging to product categories related to women's fashion may be displayed at the top of the product search result 130 .

According to one embodiment, the relevance of each product category included in the user query 110 and the graded category groups 122, 124, and 126 may be determined by calculating a relevance score. The user query 110 and product category association score may be a probability value that the user query 110 is associated with the product category. For example, when 'knitted dress' is input as the user query 110, the relevance score of 'knitted dress' and the product category 'fashion' included in category 1 (122) may be '0.4'. The relation score of 'knit dress' and the product category 'maternity/childcare' may be '0.1'. Accordingly, among the product categories of category 1 (122), products belonging to the 'fashion' category with the highest related score may be displayed in the search results 130. FIG.

In another embodiment, the search result 130 may display all products belonging to the product category most relevant to the user query 110 among the product categories included in the graded category groups 122 , 124 , and 126 . For example, when 'knit dress' is received as the user query 110, the 'fashion' category with the highest related score among the product categories included in category 1 (122) and the category 2 (124) include Search results 130 are products that fall under both the 'ladies' fashion' category with the highest related score among the product categories and the 'dress' category with the highest related score among the product categories included in category 3 (126). can be displayed in According to the configuration as described above, the product of the category desired by the user is estimated from the received user query 110 and displayed at the top of the search result 130, thereby improving the click rate of the search result.

In addition, according to the product category extraction method for the user query 110, the user query 110 is divided into syllable and/or morpheme units for analysis, and a plurality of product categories 120 for the user query 110 are identified. Relevant score can be calculated. In one embodiment, the product category extraction method may calculate the relevance score of the plurality of product categories 120 with respect to the user query 110 by analyzing the user query 110 by dividing it into syllable units. In addition, the product category extraction method can calculate the relevance score of a plurality of product categories 120 with respect to the user query 110 by dividing the user query 110 into morpheme units and analyzing them. The two relevance scores extracted as described above are used as the final relevance score by summing or combining.

The product category extraction method for the user query 110 can calculate the relevance score of the product category 120 using a statistically based classification model. A statistically-based classification model can be trained to calculate the probability that a user query 110 is associated with multiple product categories 120 . In one embodiment, the statistically-based classification model takes as input a user query 110 divided into units of syllables and/or morphemes, and calculates probabilities (or association scores) associated with all possible product categories 120. can learn For example, when 'knit dress' is input as the user query 110, the probability value corresponding to 'fashion', 'women's fashion', and 'dress', which are major, middle and small categories related to this, is It can be calculated to be larger than the probability value corresponding to other product categories.

By the method described above, the search ranking of the products 132 and 136 included in the search result 130 can be adjusted based on the related score of the product category 120 calculated by the statistically based classification model. The product search ranking (or product search ranking score) may be determined based on the lexical similarity of the product name or product category to the user query 110, user popularity of the product, and the like. At this time, the search ranking score of each product 132, 136 and the related score of the product category may be applied as a sum or weighted value to determine the final search ranking score of the product 132, 136. FIG. In one embodiment, the search rankings of the products 132 and 136 may be adjusted based on the highest relevance score among the relevance scores of the product category 120 calculated for the user query 110 . Alternatively, the search rankings of the products 132 and 136 may be adjusted based on the top n (n is an integer greater than 1) related scores among the related scores of the product category 120 calculated for the user query 110 .

In another embodiment, the product category extraction method for user query 110 is used to generate or update product categories for product information from a product database. For example, the product category extraction method calculates the relevance scores of all possible product categories based on the product name of the product information stored in the product database, and selects the product category associated with the highest relevance score as the product of the product information. Can be saved or updated as categories.

In the example shown in FIG. 1, product search results 130 or product search rankings determined by the product category extraction method for shopping search of the present disclosure are provided via a website or application for product search or shopping. However, it is not limited to this. Methods according to various embodiments of the present disclosure can also be provided via a search service of an integrated search site or a search application executed by a user terminal. In addition, although FIG. 1 shows that the product category 120 includes three graded category groups 122, 124, and 126, it is not limited to this. Product categories 120 may include four or more tiered category groups. Also, the number of categories related to each product type may be different. For example, a product related to 'dress' is associated with a total of three categories: 'fashion' as a major category, 'ladies' fashion' as a middle category, and 'dress' as a minor category, and is associated with 'tablet PC'. The products to be sold are associated with a total of two categories of 'digital/consumer electronics' as a major category and 'tablet PC' as a minor category.

FIG. 2 is a screen displaying search rankings of products 212 and 214 adjusted based on the relevance score of the product category to the user query calculated by the product category extraction method for shopping search according to an embodiment of the present disclosure. It is a figure which illustrates. In one embodiment, when a user query is input, a search result 220 in which search rankings of products 212 and 214 belonging to product categories related to the input user query are adjusted may be displayed. As shown in the figure, the major, medium, and small categories of products 212 can be 'fashion', 'ladies' underwear/pajamas', and 'pajamas/loungewear', respectively. The categories of class, middle class and small class can be 'fashion', 'women's fashion' and 'dress' respectively. At this time, in the existing search results 210 that do not reflect the category relevance score, the product 212 containing the same or similar vocabulary as the user query may be displayed above other products 214 .

Meanwhile, among the received user query 'knit dress' and the product categories included in the level category group, the product categories with the highest related scores may be 'fashion', 'women's fashion' and 'dress'. . At this time, if the search rankings of the products 212 and 214 are adjusted based on the calculated related scores, the products 214 that belong to all three product categories with high related scores are combined into one product category with high related scores (for example, ' The product search order can be adjusted so that the product 212 belonging only to fashion') is displayed above the product 212.

In one embodiment, the products 212 and 214 belonging to the product category with the highest relevance score (eg, 'fashion') among the product categories included in the higher level category 1 (eg, 122 in FIG. 1) are searched for products. You can adjust the product search ranking so that it appears at the top of the results. Additionally or alternatively, the product 214 belonging to the product category with the highest relevance score (e.g., 'ladies' fashion') among the product categories included in the middle-level category 2 (e.g., 124 in FIG. 1) is the other product. The product search order can be adjusted so that the product is displayed higher than 212. In this way, products belonging to one or more product categories with the highest relevance score among the product categories included in one or more of the predetermined number of tiered category groups are displayed at the top of the product search results. You can adjust the product search ranking as follows.

In one embodiment, the search ranking of the products 212, 214 can be adjusted by the product category relevance scores calculated for the user query and product category weights that determine the importance of each relevance score. Specifically, based on the product relevance score (or product search ranking score) calculated by considering factors such as the lexical similarity between user queries and product names, and the user popularity of the products 212 and 14, The final product search ranking score can be calculated by adding the calculated product category relevance score and product category weight value. For example, the product compatibility score calculated considering the user query of the product 214 and the lexical similarity of the specific product name is '0.6', and the calculated category relevance score is '0.5'. , if the product category weight is '0.3', the final product search ranking score can be '0.6+(0.5×0.3)=0.75'. The search rankings of the products 212 and 214 can be adjusted according to the final product search ranking score calculated in this way.

FIG. 3 illustrates a system 300 in which a plurality of user terminals 310_1, 310_2, 310_3 and a product category extraction server 330 are communicatively coupled to provide a product category extraction service for shopping search according to an embodiment of the present disclosure. It is a schematic diagram showing As shown in the figure, the system 300 includes a product category extraction server 330 that provides a product category extraction service for shopping searches, and a plurality of user terminals 310_1 and 310_2 connected to the product category extraction server 330 via a network 320. , 310_3 and a search log database 340 . According to one embodiment, product category extraction server 330 is one or more computer-executable programs (e.g., downloadable applications) that can store, provide, and execute data for providing product category extraction services. It can include server devices and/or databases, and one or more distributed computer devices and/or distributed databases of cloud computing service infrastructure. The product category extraction service provided by the product category extraction server 330 can be provided to the user via a search application or web browser installed in each of the plurality of user terminals 310_1, 310_2, 310_3.

A plurality of user terminals 310_1, 310_2, 310_3 can communicate with the product category extraction server 330 via the network 320. FIG. Network 320 can be configured to allow communication between multiple user terminals 310 and product category extraction servers 330 . Depending on the installation environment, the network 320 may be a wired network such as Ethernet, PLC (Power Line Communication), telephone line communication device and RS-serial communication, mobile communication network, WLAN (Wireless LAN), It can consist of wireless networks such as Wi-Fi®, Bluetooth® and ZigBee®, or a combination thereof. The communication method is not limited, and the user terminals 310_1 and 310_2 as well as the communication method utilizing a communication network (eg, mobile communication network, wired Internet, wireless Internet, broadcast network, satellite network, etc.) that can include the network 320. , 310_3 is also included. For example, the network 320 includes a PAN (personal area network), a LAN (local area network), a CAN (campus area network), a MAN (metropolitan area network), a WAN (wide area network), a BBN (broadband network), the Internet, and the like. Any one or more of the networks can be included. Network 320 may also include any one or more of network topologies including bus networks, star networks, ring networks, mesh networks, star-bus networks, tree or hierarchical networks, etc. It is not limited to this.

In FIG. 3, a mobile phone or smart phone 310_1, a tablet computer 310_2 and a laptop or desktop computer 310_3 are shown as examples of user terminals, but the user terminals 310_1, 310_2, and 310_3 may be wired and/or wireless. It can be any computing device capable of communication and on which a search application, mobile browser application or web browser can be installed and executed. For example, the user terminal 310 includes a smart phone, a mobile phone, a navigation terminal, a desktop computer, a laptop computer, a digital broadcast terminal, a PDA (Personal Digital Assistants), a PMP (Portable Multimedia Player), a tablet computer, a game console (game console), wearable device, internet of things (IoT) device, virtual reality (VR) device, augmented reality (AR) device, and the like. In addition, although FIG. 2 shows that three user terminals 310_1, 310_2, and 310_3 communicate with the product category extraction server 330 via the network 320, the present invention is not limited to this, and different numbers of user terminals communicate with the network 320. It may also be configured to communicate with the product category extraction server 330 via.

According to one embodiment, the product category extraction server 330 can receive one or more user queries input from a plurality of user terminals 310_1, 310_2, and 310_3. According to one embodiment, the product category extraction server 330 can receive user queries entered into a search engine of a search site displayed by a web browser running on a plurality of desktop or laptop computers 310_3. According to another embodiment, the product category extraction server 330 retrieves user queries entered into search engines of search sites displayed by mobile browser applications running on a plurality of mobile phones, smart phones and tablet computers 310_1, 310_2. can receive According to another embodiment, the product category extraction server 330 can receive user queries entered into a search engine of an online shopping site that sells products.

In addition, the product category extraction server 330 may calculate relevance scores of a plurality of product categories with respect to the received user query. In one embodiment, the product category extraction server 330 may include a statistically based classification model trained to calculate the relevance score of product categories. The product category extraction server 330 divides the received user query into units of syllables and/or morphemes, and uses the divided user queries of units of syllables and/or morphemes as inputs for a statistical-based classification model that calculates a relevance score. can. In addition, the product category extraction server 330 can receive learning data for learning a statistical-based classification model from the search log database 340 connected through the network 320 . For example, the learning data received from the search log database 240 may include a pair of user queries and product information (e.g., category information of specific products) selected by the user among product search results corresponding to the user queries. can be done.

The product category extraction server 330 can extract product categories associated with one or more high-order related scores from a plurality of product categories, and search product information belonging to the product categories from a product database (not shown). The ranking of the retrieved product information can be readjusted based on the relevance score of the product category. In addition, the product category extraction server 330 may adjust the product search order based on the product category weighting value that determines the importance of the product category related to the top relevance score and the relevant relevance score. Here, the product category weight value can be changed according to the user's response to the product search result (for example, the number of times the user selects the top product search result).

FIG. 4 is a block diagram showing internal configurations of the user terminal 310 and the product category extraction server 330 according to an embodiment of the present disclosure. The user terminal 310 can represent any computer device capable of running a web browser or an application and capable of wire/wireless communication. It may include a computer 310_2, a laptop or desktop computer 310_3, and so on. As shown in FIG. 4, user terminal 310 may include memory 412 , processor 414 , communication module 416 and input/output interface 418 . Similarly, product category extraction server 330 may include memory 432 , processor 434 , communication module 436 and input/output interface 438 . As shown in FIG. 4 , user terminal 310 and product category extraction server 330 can be configured to communicate information and/or data over network 320 using respective communication modules 416 , 436 . Input/output device 420 is also configured to input information and/or data to user terminal 310 and output information and/or data generated from user terminal 310 via input/output interface 418 . can.

Memory 412, 432 may include any non-transitory computer-readable medium. According to one embodiment, the memory 412, 432 is a permanent memory such as random access memory (RAM), read only memory (ROM), disk drives, solid state drives (SSD), flash memory, and the like. A permanent mass storage device may be included. As another example, a permanent large-capacity storage device such as a ROM, SSD, flash memory, disk drive, etc. may be used as a separate permanent storage device separate from the memories 412 and 432. 330. In addition, the memories 412 and 432 store an operating system and at least one program code (for example, a code for executing a product category extraction method installed and driven by the user terminal 310 or the product category extraction server 330). obtain.

Such software components can be loaded from a computer-readable medium separate from memory 412,432. Such separate computer-readable recording media can include recording media directly connectable to such user terminal 310 and product category extraction server 330, such as floppy drives, discs, tapes, DVD/CD- Computer readable media such as ROM drives and memory cards may be included. As another example, software components and the like can be loaded into memory 412, 432 via communication modules 416, 436, which are not computer-readable media. For example, at least one program is a computer program installed by a file provided via the network 320 by a developer or a file distribution system that distributes an application installation file (for example, a product category extraction service for shopping searches). It can be loaded into the memory 412, 432 based on the application serving).

The processors 414, 434 can be configured to process computer program instructions by performing basic arithmetic, logic and input/output operations. Instructions may be provided to processors 414 , 434 by memory 412 , 432 or communication modules 416 , 436 . For example, the processors 414,434 can be configured to execute instructions received by program code stored in a storage device, such as the memory 412,432.

The communication modules 416 and 436 can provide configurations and functions for the user terminal 310 and the product category extraction server 330 to communicate with each other via the network 320 so that the user terminal 310 and/or the product category extraction server 330 can communicate with each other. configuration or functionality for communicating with other user terminals or other systems (eg, separate search engine systems, etc.). For example, the processor 414 of the user terminal 310 may request or data generated by program code stored in a recording device such as the memory 412 (e.g., a search result output request considering a product category for a user query) may be sent to the communication module. Control at 416 may be communicated to product category extraction server 330 via network 320 . Conversely, control signals or instructions provided by the control of processor 434 of product category extraction server 330 may be received by user terminal 310 via communication module 416 of user terminal 310 via communication module 436 and network 320 . For example, the user terminal 310 can receive the product search result for the user query, the related scores of multiple product categories, the product search result reflecting the calculated related score, etc. from the product category extraction server 330 through the communication module 416 . .

Input/output interface 418 may be a means for interfacing with input/output devices 420 . By way of example, input devices may include devices such as cameras including image sensors, keyboards, microphones, mice, etc., and output devices may include devices such as displays, speakers, haptic feedback devices, etc. . As another example, the input/output interface 418 may be a means for interfacing with a device such as a touch screen that integrates configuration and functions for performing input and output. For example, processor 414 of user terminal 310 may be configured using information and/or data provided by product category extraction server 330 or other user terminal 310 when processing computer program instructions loaded in memory 412. Service screens and product search results can be displayed on the display via the input/output interface 418 . Although the input/output device 420 is not included in the user terminal 310 in FIG. The input/output interface 438 of the product category extraction server 330 also interfaces with the product category extraction server 330 or with devices (not shown) for input and output that the product category extraction server 330 may include. It can be a means for interfacing. Although FIG. 4 shows the input/output interfaces 418 and 438 as elements configured separately from the processors 414 and 434, the present invention is not limited to this and the input/output interfaces 418 and 438 are configured to be included in the processors 414 and 434. You can also

User terminal 310 and product category extraction server 330 may include more components than those shown in FIG. However, it is not necessary to explicitly show most prior art components. According to one embodiment, user terminal 310 can be implemented to include at least a portion of input/output device 420 described above. In addition, the user terminal 310 may further include other components such as a transceiver, a GPS (Global Positioning System) module, a camera, various sensors and a database. For example, if the user terminal 310 is a smart phone, it can generally include components that a smart phone has, such as an acceleration sensor, a gyro sensor, a camera module, various physical buttons, buttons using a touch panel, input/output ports, and The user terminal 310 may be embodied to further include various components such as a vibrator for vibration.

According to one embodiment, the processor 414 of the user terminal 310 can be configured to run a search application, mobile browser application, or web browser capable of connecting to a product search or shopping site. At this time, program code associated with the application or web browser can be loaded into the memory 412 of the user terminal 310 .

The processor 414 of the user terminal 310 receives information and/or data provided from the input/output device 420 via the input/output interface 418, and receives information and/or data from the product category extraction server 330 via the communication module 416. can be received and the received information and/or data can be processed and stored in memory 412 . Such information and/or data may also be provided to product category extraction server 330 via communications module 416 .

According to one embodiment, the processor 414 can receive input or selected text, images, actions, etc. through an input/output device 420 such as a touch screen or keyboard coupled to the input/output interface 418 . The received text, images and/or actions may be stored in memory 412 or provided to product category extraction server 330 via communications module 416 and network 320 . For example, the processor 414 can receive one or more user queries and/or product categories corresponding to the user queries via an input/output device such as a touch screen, keyboard, or the like, and process the received user queries and/or product categories. , etc., can be provided to product category extraction server 330 via communications module 416 and network 320 .

According to one embodiment, processor 414 can provide a user query received by an input device or a request for search results for the user query to product category extraction server 330 via network 320 and communication module 416 . Upon receiving the user query, the product category extraction server 330 may calculate relevance scores of multiple product categories from the user query. In addition, the product category extraction server 330 may provide information on product search results or product search rankings reflecting the calculated relevance score of the product category.

According to another embodiment, the processor 414 can provide the user query received by the input device or a request for search results for the user query to the product category extraction server 330 via the network 320 and the communication module 416 . Processor 414 can also receive product search results associated with the user query from product category extraction server 330 or product databases via network 320 and communication module 416 . The processor 434 may calculate relevance scores of a plurality of product categories with respect to the user query according to a user query or a search result request for the user query, and apply the calculated relevance scores to product search results to adjust product search rankings. .

According to another embodiment, product category extraction server 330 can receive user queries and user-selected product categories from user terminal 310 via network 320 and communication module 416 . Alternatively, the product category extraction server 330 can receive data pairs of user queries and user-selected product categories from the search log database (340 in FIG. 3) via the network 320 . The product category extraction server 330 generates learning data from the received user query and product category data pairs, and uses the generated learning data to calculate the user query and product category relevance score. can be learned.

FIG. 5 is a diagram showing a configuration for calculating related scores 532, 542, 552, and 562 of product categories 530, 540, 550, and 560 for a user query 510 using a statistically-based classification model 520 according to an embodiment of the present disclosure. is. Statistical-based classification model 520 can be generated, stored, or executed by product category extraction server 330 or user terminal 310 shown in FIGS. As shown in FIG. 5, the statistically-based classification model 520 can output relevance scores 532, 542, 552, and 562 for a plurality of product categories 530, 540, 550, and 560 with a user query 510 as input. Relevance scores 532 , 542 , 552 , 562 may be probability values that user query 510 is associated with each of a plurality of product categories 530 , 540 , 550 , 560 . For example, when 'knit dress' is input as the user query 510, the probability value associated with the product category 530 of 'fashion' included in category 1, which is the major category, is '0.4'. Possibly, the probability value associated with the product category 540 of 'maternity/child care' may be '0.1'. In addition, the probability value associated with the product category 550 of 'ladies' fashion' included in category 2 in which 'knit dress' is a middle class may be '0.5', and the probability value associated with the product category 560 of 'men's fashion' may be '0.5'. The probability value to do may be '0.1'.

Relevance scores 532 , 542 , 552 , 562 for each product category 530 , 540 , 550 , 560 can be calculated by statistically based classification model 520 . The statistical-based classification model 520 can learn to calculate probability values associated with multiple product categories from the input user query 510 . In this case, the product category extraction server or user terminal can receive learning data for learning the statistical-based classification model 520 from the search log database (eg, 340 in FIG. 3). For example, the learning data received from the search log database may include a pair of the user query 510 and product information (eg, category information of a specific product) selected by the user from the product search results corresponding to the user query 510. . Statistical-based classification model 520 may also be a machine learning or artificial neural network model for determining the relevance of multiple product categories to user query 510 .

In FIG. 5, one statistically based classification model 520 is used to calculate the related scores 532, 542, 552, 562 of the four product categories 530, 540, 550, 560 for the user query 510. It is not limited to this. According to various implementations, the statistically-based classification model 520 can be configured to calculate the relevance scores of less than four or more than five product categories for the user query 510 . Also, depending on how the user query 510 is pre-processed or analyzed, more than one statistically-based classification model can be used. In one embodiment, the user query 510 is split into multiple smaller units of data (e.g., morpheme and/or syllable units of data), and the split data is input to corresponding statistically-based classification models. can be In this case, the final relevance score can be calculated by summing or combining relevance scores calculated from a plurality of statistical-based classification models.

FIG. 6 is a diagram illustrating a configuration for calculating relevance scores 632, 642, 652, and 662 using a preprocessor 610 and a classifier 620 including a statistical-based classification model according to an embodiment of the present disclosure. Preprocessor 610 and classifier 620 can be generated, stored, or executed by product category extraction server 330 or user terminal 310 shown in FIGS. As shown in FIG. 6, the product category extraction server or user terminal includes a preprocessor 610 that analyzes a user query 510 to generate input data for statistical-based classification models 622 and 624, and and a classifier 620 that uses the input data to calculate relevance scores for multiple product categories.

The pre-processor 610 may include a syllable unit divider 612 that divides the user query 510 into syllable units, and a morpheme unit divider 614 that divides the user query 510 into morpheme units. In one embodiment, the syllable unit divider 612 and the morpheme unit divider 614 can transform the text of the user query 510 into syllable embedding vectors and morpheme embedding vectors, respectively. Each of the syllable unit segmenter 612 and the morpheme unit segmenter 614 can be implemented with an artificial neural network trained to embed the user query 510 in a continuous vector space.

In addition, the classifier 620 is a first statistical-based classification model 622 that receives the user query 510 divided into syllable units, and a second statistical-based classification model 622 that receives the user query 510 divided into morpheme units. Can include a classification model 624 and a combiner 626 that sums or combines the relevance scores calculated from each of the statistically based classification models 622, 624 to calculate final relevance scores 632, 642, 652, 662. .

In one embodiment, the product category extraction server or user terminal inputs the first statistical-based classification model 622 that receives the user query 510 divided into syllable units and the user query 510 divided into morpheme units. Relevance scores 632, 642, 652, and 662 can be calculated using a second statistically-based classification model 624 as follows.

Statistical-based classification models 622 and 624 included in the classifier 620 receive the user query 510 divided into units of syllables and morphemes, and obtain relevance scores 632, 642, 632, 642, 632, 642, 652, 662 can be calculated. Each of the statistically-based classification models 622, 624 uses learning data extracted from the search log database to obtain relevance scores 632, 642, 652, 662 of a plurality of product categories 630, 640, 650, 660 for the user query 510. You can learn to calculate. The learning data extracted from the search log database may include pairs of user queries and product information (for example, category information of specific products) selected by the user among product search results corresponding to the user queries.

Statistical-based classification models 622 and 624 do not use the text (for example, sentences, phrases, words, etc.) that make up the user query 510 as they are, but divide them into units of syllables or morphemes. , or the user query 510 does not appear in the training data and is unfamiliar. For example, if the user query 510 is 'knit dress', 'ni', 'to', 'one', 'pi', and 'su' obtained by dividing 'knit dress' into units of syllables are classified into the statistical-based classification model 622 . , and if the user query 510 contains typographical errors such as 'nit-op-piece', it is divided into syllable units 'ni', 'to', 'o', 'pee', It is the same as or similar to the relevance score calculated by inputting 'su' into the statistical-based classification model 622 . As another example, if the user query 510 is 'knit dress', the relevance score calculated by inputting 'knit' and 'dress' into the morpheme unit of the 'knit dress' into the statistical-based classification model 624 and , When the order of morphemes is changed such as 'one piece knit' in the user query 510, the relevance score calculated by inputting 'one piece' and 'knit' divided into morpheme units into the statistical-based classification model 622 is , identical or similar.

The relevance scores calculated by the statistically-based classification models 622 and 624 can be summed or combined by a combiner 626 . As a result, the final relevance scores 632, 642, 652, 662 can be calculated with higher accuracy than when using a single statistically based classification model. With the above configuration, it is possible to maximize the matching coverage of the learning data used by the statistical-based classification models 622 and 624, and improve the accuracy of calculating user queries and relevance scores of multiple product categories.

FIG. 7 is a diagram illustrating an example of generating learning data of a statistical-based classification model for product category extraction from various search results according to an embodiment of the present disclosure. Training data used for training the statistically-based classification model can be extracted from the search log database 780 . The search log database 780 stores, for example, user queries input by the user while using the shopping search service (e.g. sentences, phrases, words, etc. containing the name of the product to be searched), shopping search results (e.g. , search results including product names, product categories, etc.), and user responses to shopping search results (e.g., user clicks on specific product names, product categories, etc., or the number of clicks) can be saved. The learning data of the statistical-based classification model may include pairs of user queries extracted from the search log database 780 and product categories associated with the user queries.

As shown in the figure, for example, when a user inputs a user query 710 into a search window of a shopping search site or an integrated search site, search results related to the input user query 710 may be output. The search results 130 include a search window containing the user query 710, a user interface 720 for filtering product categories, search products 730, 740, product categories 732, 734, etc. associated with the search products 730, 740. can include

In one embodiment, when a user who has entered a user query 710 clicks a product category 722 included in the user interface 720, the learning data 750 of the user query 710 and clicked product category 722 pair is stored in a search log database. 780 can be saved. The processor of the product category extraction server can extract learning data including pairs of user queries 710 and product categories from the search log database 780 storing such user queries 710 and click log information. For example, when the user inputs 'knit dress' as the user query 710 and clicks 'fashion' among product categories through the user interface 720 to generate search log data, 'knit dress-fashion' is generated. Paired learning data 750 can be generated.

As another example, when a user who has entered a user query 710 clicks on products 730 and 740 output as search results 130, product categories 732 and 734 related to the user query 710 and the products 730 and 740 are displayed. Paired learning data 750 , 760 , 770 may be stored in search log database 780 . For example, if the user clicks on products 730 and 740 in the search results 130 after inputting 'knit dress' as the user query 710, the product category associated with the clicked products 730 and 740 is '. Learning data 750 for pairs of fashion' and user queries 710, learning data 760 for pairs of intermediate product category 'women's fashion' and user queries 710, and small product category 'dress' and user queries A total of three training data may be generated, including 710 pairs of training data 770 . Additionally or alternatively, if the user who entered the user query 710 clicks a product category 732, 734 associated with the product, all product categories included in the clicked product category 732, 734 and the user query 710 pair. of training data 750, 760, 770 may be generated.

FIG. 8 is a flowchart illustrating a product category extraction method 800 for shopping searches performed by a product category extraction server according to one embodiment of the present disclosure. In one embodiment, the product category extraction method 800 for shopping searches can begin upon receipt of a user query (S810). In one embodiment, a user query can be received from a search window of a shopping site or an integrated search site displayed on a user terminal and forwarded to a processor of a product category extraction server.

Based on the received user query, a product search result corresponding to the user query is extracted from the product database (S820). In one embodiment, the processor of the product category extraction server can extract product search results for the user query from the product database. At this time, the product search result may include product search ranking information determined based on user queries and information such as lexical similarity of product names and user popularity of products stored in the product database. For example, if the user query is 'knit dress', among the products with a product name that includes at least a part of 'knit dress', products with a high purchase rate by users or products with a high click rate by users are It can be extracted as a product search result with a high rank.

The user query can be divided into units of syllables and/or morphemes (S830). In one embodiment, the processor of the product category extraction server can divide the user query into units of syllables and/or morphemes to input into the statistically-based classification model. For example, a user query can be converted into a syllable embedding vector and/or a morpheme embedding vector and input to a statistically based classification model.

In addition, the statistical-based classification model can calculate the relevance score of a plurality of product categories with respect to the user query (S840). The product category may include multiple category groups divided into a predetermined number of steps, and the processor may calculate relevance scores for the user query and product categories included in each category group. When the user query is divided into syllables or morphemes, the statistically based classification model can calculate a probability value that the user query divided into syllables or morphemes is associated with multiple product categories.

In another embodiment, the processor of the product category extraction server can divide the user query into units of syllables and morphemes, respectively. The processor calculates, using a first statistically-based classification model, a first probability that the user query divided into syllables is associated with multiple product categories, and using a second statistically-based classification model, A second probability that the morpheme-divided user query is associated with the product category can be calculated. The processor can then sum or combine the first probability and the second probability to calculate a final product category relevance score. As described above, by dividing the user query into syllable and/or morpheme units and calculating the relevance score for the product category using the statistical-based classification model, even if the user query is mistyped or not included in the learning data, It is possible to improve the accuracy of calculating the related score of the product category.

Based on the calculated relevance score, the product search ranking according to the product search result for the user query may be adjusted (S850). In one embodiment, the processor of the product category extraction server applies the calculated at least one related score as a weighted value of the product search results belonging to the corresponding product category among the product search results, and ranks the product search. Adjustable. As a result, products including product categories with high relevance scores can be displayed at the top of the search results. By displaying products having product categories related to the user query at the top of the search results, it is possible to increase the click-through rate and purchase rate of the relevant products.

FIG. 9 is a flowchart illustrating a product search ranking adjustment method 900 for shopping searches according to one embodiment of the present disclosure. In one embodiment, the product search ranking adjustment method 900 may begin by calculating relevance scores of multiple product categories to a user query (S910). A method of calculating the relevance score of a plurality of product categories with respect to a user query may be performed in the same manner as the method described with reference to FIG.

Next, a product search ranking can be adjusted by applying the related score of the product category as a weighted value of the product search results belonging to the corresponding product category among the product search results (S920). As an alternative, one or more related scores with a related score greater than or equal to a predetermined threshold are extracted from the calculated related scores, and one or more top related scores are added to the corresponding product category from the product search results. It can be applied as a weighted value of the search result of the belonging product to adjust the product search ranking.

A product category weight that determines the importance of a plurality of product categories can be determined (S930). The product category weight value may indicate the degree of influence of the extracted product category on adjustment of the search ranking of related products. In one embodiment, the product category weight value may be a default value for each product category. Alternatively, the product category weights can vary dynamically depending on user responses to product search results. Specifically, the processor may output a search result screen by adjusting the product search order, and receive the number of clicks on product information on the output product search result screen. If the number of clicks on product information is greater than or equal to a preset threshold, the weighting value of the product category can be changed.

A product search ranking may be adjusted based on product category weights and related scores that determine weights of a plurality of product categories (S940). As a result of adjusting the product search ranking, products related to the extracted product category may be displayed higher than other products not related to the extracted product category. Thereby, it is possible to output to the user terminal a search result reflecting the product category intended by the user based on the user query.

FIG. 10 is a flowchart illustrating a product category extraction method 1000 for shopping search performed by a user terminal according to one embodiment of the present disclosure. In one embodiment, the product category extraction method 1000 for shopping searches performed by a user terminal can be initiated by a processor of the user terminal receiving a user query (S1010). A user query can be received from a search window of a shopping site or an integrated search site displayed on the user terminal.

The processor of the user terminal can receive the user query and the relevance scores of multiple product categories according to the statistical-based classification model (S1020). A product category can include multiple categories divided into a predetermined number of steps. At this time, the statistical-based classification model can be learned in advance to calculate the relevance score of a plurality of product categories with respect to the user query. In addition, in order to calculate a relevance score with high accuracy, the statistical-based classification model receives user queries divided into units of syllables and/or morphemes and uses two or more statistics for calculating relevance scores of product categories. A base classification model can be included.

The processor of the user terminal may adjust the product search order for the user query based on the calculated relevance score (S1030). Specifically, the processor may apply the relevance score of the product category as a weighted value of the product search results belonging to the corresponding product category among the product search results, thereby adjusting the product search order. Additionally or alternatively, the processor adjusts the product search order by applying the relevance score of the product category belonging to each of the plurality of category groups as a weighted value of the product search results belonging to the corresponding product category among the product search results. can.

FIG. 11 is a flowchart illustrating a method 1100 of training a statistically-based classification model for product category extraction according to one embodiment of the present disclosure. As shown, the learning method 1100 includes a processor generating first training data including a plurality of first user queries and a plurality of first product categories associated with the plurality of first user queries. (S1110) can start. The processor can extract a plurality of first product categories from a search log database associated with a plurality of first user queries. The search log database may include product category information selected or clicked by the user from user queries and product search results extracted thereby. For example, the processor can extract product categories associated with a plurality of user queries from at least one of a federated search log database, a shopping or product search log database. With the above configuration, the learning data is generated by the user's actions (e.g., product category filtering, selection or clicking of products or product categories, etc.) on user queries entered by the user and product search results extracted by the user queries. Since it is obtained from the search log database, it is possible to efficiently secure training data for the statistically based classification model.

The processor may learn a statistically-based classification model to calculate relevance scores of one or more product categories corresponding to the second user query based on the plurality of first learning data (S1120). In one embodiment, the processor can divide the plurality of first user queries into units of morphemes to learn a statistically based classification model. Specifically, the processor can learn a statistically based classification model to calculate relevance scores of one or more product categories corresponding to user queries based on a plurality of first user queries divided into units of morphemes. . Additionally or alternatively, the processor can divide the plurality of first user queries into units of syllables to learn a statistically based classification model. The processor can learn a statistically-based classification model to calculate relevance scores for one or more product categories corresponding to user queries based on a plurality of first user queries divided into syllable units.

In another embodiment, the processor can divide the plurality of first user queries into units of syllables and morphemes to learn a statistically based classification model. Specifically, the processor can divide the plurality of first user queries into units of morphemes and syllables. Then, the processor learns a first statistical-based classification model to calculate relevance scores of one or more product categories corresponding to the user queries based on the plurality of first user queries divided into morpheme units. can. Also, the processor learns a second statistically based classification model to calculate relevance scores of one or more product categories corresponding to the user queries based on the plurality of first user queries divided into syllable units. can. As described above, the processor can more accurately calculate the relevance score of the user query and product category by using two or more statistically-based classification models according to the method of dividing the user query.

The product category extraction method for shopping search and the product category extraction learning method described above can also be embodied as computer-readable codes on a computer-readable recording medium. The recording medium may be of continuous storage or temporary storage for execution or download of the computer-executable program. In addition, the medium can be a variety of recording means or storage means in the form of a single or multiple pieces of hardware combined, but is not limited to a medium directly connected to a computer system, and is distributed over a network. can be anything. Examples of media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical recording media such as CD-ROMs and DVDs, and magneto-optical media such as floptical disks. medium), ROM, RAM, flash memory, etc., which are configured to store program instructions. Examples of other media include recording media and storage media managed by application stores that distribute applications, sites that supply and distribute various software, and servers.

The methods, acts, or techniques of the disclosure may be embodied by various means. For example, such techniques may be implemented in hardware, firmware, software, or any combination thereof. A person of ordinary skill in the art can appreciate that the various exemplary logical blocks, modules, circuits and algorithmic steps described in the present disclosure can be embodied in electronic hardware, computer software, or a combination of both. should be. To clearly illustrate such mutual alternatives of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software will vary depending on the particular application and design requirements imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementations should not be interpreted as departing from the scope of the present disclosure. .

In hardware implementations, the processing units utilized in performing the techniques may be one or more of ASICs, DSPs, digital signal processing devices (DSPDs), programmable logic devices (PLDs). ), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, electronic devices, and other electronic devices designed to perform the functions described in this disclosure. It can also be embodied within a unit, a computer, or a combination thereof.

Thus, the various exemplary logic blocks, modules and circuits described by this disclosure may be general purpose processors, DSPs, ASICs, FPGAs or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or , may be embodied or performed in any combination, including those designed to perform the functions described herein. A general-purpose processor can be a microprocessor, but, in the alternative, the processor can be any conventional processor, controller, microcontroller, or state machine. A processor may also be embodied in a combination of computing devices, such as a DSP and a microprocessor, multiple microprocessors, one or more microprocessors associated with a DSP core, or any other combination of configurations. can.

In firmware and/or software implementations, techniques may include random access memory (RAM), read-only memory (ROM), non-volatile random access memory (NVRAM), programmable read-only memory (PROM), erasable memory (EPROM). It can be embodied in instructions stored on computer readable media such as programmable read-only memory), electrically erasable PROM (EEPROM), flash memory, compact disc (CD), magnetic or optical data storage devices, and the like. The instructions may be executable by one or more processors and may cause the processors to perform certain aspects of the functions described in this disclosure.

When implemented in software, the techniques can be stored on or transferred over a computer-readable medium as one or more instructions or code. Computer-readable media includes both computer storage media and communication media that facilitates transfer of a computer program from one place to another. A storage media may be any available media that can be accessed by a computer. By way of non-limiting example, such computer readable media may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any desired program code. It can also include any other computer-accessible medium that can be used to transport or store instructions or data structures in a form. Also, any connection is properly termed a computer-readable medium.

For example, software may be transferred from a website, server or other remote source using coaxial cable, fiber optic cable, lead wire, digital subscriber line (DSL), or wireless technologies such as infrared, radio and microwave. Coaxial cable, fiber optic cable, lead wire, digital subscriber line, or wireless technologies such as infrared, radio and microwave, where applicable, are included within the definition of medium. As used herein, disk and disc include CDs, laser discs, optical discs, DVDs (digital versatile discs), floppy discs and Blu-ray discs, where disk generally refers to a magnetic A disc reproduces data optically using a laser, whereas a disc reproduces data optically. Combinations and the like should also be included within the scope of computer-readable media and the like.

A software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium may be coupled to the processor, such that the processor reads information from, and writes information to, the storage medium. Alternatively, the storage medium may be integrated into the processor. The processor and storage medium can also reside within an ASIC. The ASIC may also reside in a user terminal. Alternatively, the processor and storage medium may exist as separate components in the user terminal.

Although the foregoing embodiments are described as utilizing aspects of the presently disclosed subject matter in one or more stand-alone computer systems, the disclosure is not so limited, and may be used in networked or distributed computing environments, such as. It can be embodied by any computing environment. Further, aspects of the subject matter of this disclosure may be embodied in multiple processing chips or devices, and storage may be similarly affected across multiple devices. Such devices may also include PCs, network servers and handheld devices.

In the present specification, the present disclosure has been described with some embodiments, but various modifications and changes can be made without departing from the present disclosure that can be understood by a person of ordinary skill in the technical field to which the disclosed invention belongs. It is possible. Also, such variations and modifications should be understood to fall within the scope of the claims appended hereto.

110 User Query 120 Product Category 122 Category 1
124 Category 2
126 Category 3
130 Search Result 310 User Terminal 320 Network 330 Product Category Extraction Server 340 Search Log Database 520 Statistical Base Classification Model 610 Preprocessor 620 Classifier

Claims (20)

In a product category extraction method for shopping search performed on a computer system,
receiving a user query;
extracting product search results for the user query from a product database;
calculating a relevance score of the user query and a plurality of product categories according to a statistically-based classification model; and
adjusting the product search ranking according to the product search results based on the calculated relevance score;
The step of calculating the relevance score of the user query and the plurality of product categories by the statistically-based classification model,
dividing the user query into units of syllables; and
calculating a first probability that the syllable-divided user query is associated with the plurality of product categories according to a first statistically-based classification model;
Product category extraction method for shopping search. The step of adjusting the product search ranking includes:
2. The method according to claim 1, further comprising adjusting the product search order by applying the related scores of the plurality of product categories as a weighted value of product search results belonging to the corresponding product category among the product search results. Product category extraction method for shopping search. 2. The product category extraction method for shopping search of claim 1, wherein the plurality of product categories includes a plurality of category groups divided into a predetermined number of steps. The step of adjusting the product search ranking according to the product search results based on the calculated relevance score,
adjusting the product search order by applying the related score of the product category belonging to each of the plurality of category groups as a weighted value of the product search results belonging to the corresponding product category among the product search results; The product category extraction method for shopping search according to claim 3. The step of adjusting the product search ranking according to the product search results based on the calculated relevance score,
a step of extracting one or more top-level related scores that are equal to or greater than a predetermined threshold from the calculated related scores; and
5. The method according to any one of claims 1 to 4, further comprising adjusting the product search order by applying the one or more top related scores as a weighted value to the product search results belonging to the corresponding product category among the product search results. A product category extraction method for shopping search according to any one of the items. The step of calculating the relevance score of the user query and the plurality of product categories by the statistically-based classification model,
dividing the user query into units of morphemes ;
calculating a second probability that the user query divided into morpheme units is associated with the plurality of product categories by a second statistically-based classification model; and
6. The shopping search method according to any one of claims 1 to 5, further comprising calculating a probability as a related score of said plurality of product categories based on said first probability and said second probability. Product category extraction method for The step of adjusting the product search ranking according to the product search results based on the calculated relevance score,
7. The method for shopping search according to any one of claims 1 to 6 , further comprising: adjusting the product search order based on the product category weighting value for determining importance of the plurality of product categories and the relevance score. product category extraction method. A step of outputting a search result screen by adjusting the product search order;
receiving the number of clicks on product information on the output search result screen; and
8. The product category extraction method for shopping search as claimed in claim 7 , further comprising changing the product category weight value when the number of clicks is greater than or equal to a predetermined threshold. In a product category extraction method for shopping search executed on a user terminal,
receiving a user query;
calculating a relevance score of the user query and a plurality of product categories according to a statistically-based classification model; and
adjusting a product search ranking for the user query based on the calculated relevance score;
The step of calculating the relevance score of the user query and the plurality of product categories by the statistically-based classification model,
dividing the user query into units of syllables; and
calculating a first probability that the syllable-divided user query is associated with the plurality of product categories according to a first statistically-based classification model;
Product category extraction method for shopping search. adjusting the product search ranking for the user query based on the calculated relevance score,
10. The shopping according to claim 9 , further comprising adjusting the product search order by applying the relevance score of the plurality of product categories as a weighted value of product search results belonging to the corresponding product category among product search results. Product category extraction method for searching. 10. The product category extraction method for shopping search according to claim 9 , wherein the plurality of product categories includes a plurality of category groups divided into a predetermined number of steps. adjusting the product search ranking for the user query based on the calculated relevance score,
adjusting the product search order by applying the relevance score of the product category belonging to each of the plurality of category groups as a weighted value of the product search results belonging to the corresponding product category among the product search results. Item 12. A product category extraction method for shopping search according to Item 11 . The step of calculating the relevance score of the user query and the plurality of product categories by the statistically-based classification model,
dividing the user query into units of morphemes;
calculating a second probability that the user query divided into morpheme units is associated with the plurality of product categories by a second statistically-based classification model; and
13. For shopping search according to any one of claims 9 to 12, further comprising calculating a probability as a relevance score of said plurality of product categories based on said first probability and said second probability. product category extraction method. A computer program for executing the product category extraction method for shopping search according to any one of claims 1 to 13 on a computer. In a learning method of product category extraction for shopping search performed by a computer system,
generating a plurality of first learning data including a plurality of first user queries and a plurality of first product categories associated with the plurality of first user queries; and
learning a statistically based classification model to calculate relevance scores for one or more product categories corresponding to a second user query based on the plurality of first learning data ;
The step of learning the statistically-based classification model includes:
dividing the plurality of first user queries into units of syllables; and
Based on the plurality of first user queries divided into syllable units, learning a statistically-based classification model to calculate the relevance score of one or more product categories corresponding to the user query,
How to learn product category extraction for shopping search. The step of generating the plurality of first learning data includes:
16. The shopping search of claim 15, comprising extracting the plurality of first product categories from at least one of integrated search click information, shopping click information, or a product database associated with the plurality of first user queries. How to learn product category extraction for The step of learning the statistically-based classification model includes:
dividing the plurality of first user queries into units of morphemes; and
Based on the plurality of first user queries divided into units of the morpheme, learning the statistical-based classification model to calculate the relevance score of one or more product categories corresponding to the user query, A learning method for product category extraction for shopping search according to claim 15 or 16. A computer program for executing the learning method for product category extraction for shopping search according to any one of claims 15 to 17 on a computer. In the product category extraction system for shopping search,
a communication module for receiving user queries;
memory; and
at least one processor coupled to the memory and configured to execute computer readable instructions contained in the memory;
The at least one processor
Extracting product search results for the user query from a product database, calculating the relevance score of the user query and a plurality of product categories using a statistically based classification model, and based on the calculated relevance score, the product search result is configured to adjust the product search ranking by
Calculating the related scores of the user query and the plurality of product categories by the statistical-based classification model includes:
dividing the user query into units of syllables; and
calculating a first probability that the user query segmented into syllables is associated with the plurality of product categories according to a first statistically-based classification model;
Product category extraction system for shopping search. Calculating the related scores of the user query and the plurality of product categories by the statistical-based classification model includes:
dividing the user query into units of morphemes;
calculating a second probability that the user query divided into morpheme units is associated with the plurality of product categories by a second statistically-based classification model; and
20. The product category extraction system for shopping search of claim 19, further comprising: calculating a probability as a relevance score of said plurality of product categories based on said first probability and said second probability. .

Images