KR20220151574A - Method, server and computer program for uploading product sale contents using website crawling based on artificial intelligence - Google Patents

Abstract

Provided are a method, a server and a computer program for uploading product sale content using website crawling based on artificial intelligence. The method for uploading product sale content using website crawling based on artificial intelligence according to various embodiments of the present invention, which is performed by a computing device, comprises the steps of: collecting the URL of a first website on which first product sale content is uploaded; generating crawling result data by performing crawling on the first product sale content using the collected URL; and uploading second product sale content on a second website using the generated crawling result data. Therefore, provided are a method, a server and a computer program for uploading product sale content using website crawling based on artificial intelligence, wherein product sale content can be imported and uploaded more quickly and conveniently.

Description

Product sales content posting method, server and computer program using artificial intelligence-based website crawling

Various embodiments of the present invention relate to a method, server, and computer program for posting product sales content using artificial intelligence-based website crawling.

BACKGROUND OF THE INVENTION [0002] Recently, with the development of technologies such as the Internet, electronic commerce through the Internet has become a new commerce method in addition to conventional real-time transactions. This e-commerce method is generally characterized in that a transaction between a seller who has opened a shopping mall website on a global communication network called the Internet and a buyer who wants to purchase a product by entering the shopping mall is performed by communication using a computer.

This Internet shopping mall method utilizes the latest rapidly developed communication protocol, image processing technology, software, multimedia, and multi-task technology to the maximum, and does not require wide stores and management personnel like conventional market transactions, and the desired product without the effort of the buyer. It is easy to find, and the publicity effect is greater than the publicity effort, so it is in the limelight of many business operators.

In accordance with this trend, Internet shopping malls will provide various types of services and functions, such as a product sales service that sells products to general customers online, as well as a service that recommends products suitable for users' characteristics and inclinations or provides promotions. However, there is a problem that active development is not being made for functions that improve the convenience of business operators who operate online shopping malls.

Conventional Internet shopping mall operators, especially Internet clothing shopping mall operators, sometimes sell products that they have manufactured and produced themselves through their own Internet shopping malls, but in most cases, they directly visit domestic and foreign markets to sell products. The purchase method was used. However, in the case of the method of buying a product by selling products directly, there is a problem in that considerable time and effort are required until the product is viewed, purchased, and actually received.

In order to solve these problems, the method is gradually changing from the method of selling products directly to the method of ordering products online through websites operated in domestic and foreign wholesale and retail markets. Since products are ordered through websites, in the case of the conventional method, it is difficult to order and manage products through multiple websites, and it takes a lot of time.

In particular, when using overseas sites, there is a problem that more time and effort are required due to language problems as well as these difficulties, and various problems such as incorrect ordering due to failure to grasp information properly.

The problem to be solved by the present invention is to crawl product sales content posted on a first website to be crawled to generate crawl result data, and sell the product on a second website to post product sales content through this. By posting content, it is to provide a method for posting product sales content using artificial intelligence-based website crawling, a server, and a computer program that can bring and post product sales content more quickly and conveniently.

Another problem to be solved by the present invention is to bring a plurality of product sales contents posted on a plurality of websites and integrate and provide them through one website, thereby registering a plurality of websites as one website is used. It is to provide a method for posting product sales content using artificial intelligence-based website crawling, a server, and a computer program that can check and order products.

The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

In order to solve the above problems, a method for posting product sales content using artificial intelligence-based website crawling according to an embodiment of the present invention is a method performed by a computing device, wherein a first product sales content is posted. Collecting the URL of the website, generating crawl result data by performing a crawl on the first product sales content using the collected URL, and using the generated crawl result data to create a second website. It may include a step of posting second product sales content.

In various embodiments, the generating of the crawling result data may include, when the first website is set as a crawling target by a user, on the first website from the time the first website is set as the crawling target. A step of automatically performing crawling on new product sales content that is posted may be included.

In various embodiments, the generating of the crawling result data may include receiving a crawling condition set by a user and at least one content that satisfies the set crawling condition among a plurality of first product sales contents posted on the first website. A step of selectively crawling only the first product sales content may be included.

In various embodiments, the step of generating the crawling result data may include a first image analysis model - the first image analysis model is pre-learned using a plurality of product images labeled with product information as learning data. If it is determined that two or more products are included in the product image by analyzing the product image included in the first product sales content through the model, crawling information on each of the two or more products results in two or more crawling results It may include the step of generating.

In various embodiments, the posting of the second product sales content may include using two or more crawling result data generated corresponding to two or more products included in the first product sales content, respectively, on the second website. When two or more second product sales contents are posted, but the two or more second product sales contents are interconnected and the user selects one of the two or more second product sales contents, the selected second product sales contents are selected. A step of providing the remaining second product sales content connected with any one second product sales content may also be included.

In various embodiments, the generating of the crawling result data may include performing crawling by automatically translating the one or more foreign languages into a preset language by a user when one or more foreign languages are included in the first product sales content. can include

In various embodiments, the generating of the crawling result data may include generating the first crawling result data by performing crawling on the first product sales content and a predetermined time from the time point when the first crawling result data is generated. and generating second crawling result data by re-crawling the first product sales content at intervals, wherein the step of posting the second product sales content includes the generated first crawl result data and the second crawl result data. Comparing the generated second crawling result data, and correcting posted second product sales content corresponding to the generated first crawling result data.

In various embodiments, the posting of the second product sales content may include a second image analysis model - the second image analysis model uses a plurality of product images labeled with product attribute information as training data. A pre-learned model - Attribute information on a product included in the product image by analyzing the product image included in the first product sales content through - The attribute information includes at least one of a category and a style of the product - Extracting and posting the second product sales content in an area corresponding to the attribute information on the extracted product on the second website.

In various embodiments, posting the second product sales content in an area corresponding to the attribute information on the extracted product on the second website may include the attribute information on the extracted product and the one or more product images. or calculating the accuracy of attribute information on the extracted product using feedback information on the product included in the one or more product images, and if the calculated accuracy is equal to or greater than a preset value, the second product sales content is displayed as Posting on a second website, and requesting confirmation of attribute information on the extracted product when the calculated accuracy is less than the predetermined value.

In various embodiments, posting the content for sale of the second product in an area corresponding to the attribute information on the extracted product on the second website may include selling the first product using the second image analysis model. Calculating the accuracy of the attribute information on the extracted product by comparing N pieces of attribute information extracted by analyzing the product image included in the content N times, and if the calculated accuracy is greater than or equal to a preset value, the second product Posting content for sale on the second website, and requesting confirmation of attribute information of the extracted product when the calculated accuracy is less than the preset value.

In various embodiments, the posting of the second product sales content may include determining whether one or more pieces of information included in the generated crawling result data are abnormal, and if it is determined that the one or more pieces of information are normal. The method may include posting the merchandise sales content and requesting confirmation of the one or more pieces of information when it is determined that the one or more pieces of information are abnormal.

In various embodiments, the step of posting the second product sales content may include comparing one or more pieces of information included in the generated crawl result data with one or more pieces of information included in each of a plurality of pre-stored crawl result data to generate the generated content. Calculating a similarity between crawling result data and a plurality of pre-stored crawling result data, and posting the second product sales content using the generated crawling result data only when the maximum value of the calculated similarity is less than a preset value steps may be included.

In various embodiments, the step of receiving an image including one or more products from a user through the second website, a second image analysis model - the second image analysis model is labeling property information about the product Extracting attribute information about one or more products included in the input image through a pre-learned model using a plurality of product images as learning data and a plurality of second products posted on the second website The method may further include selecting one or more second product sales contents corresponding to attribute information on the extracted one or more products from among sales contents, and providing the selected one or more second product sales contents to the user as a search result. can

In various embodiments, collecting social data for a plurality of products, extracting fashion trend information by analyzing the collected social data, and posting on the second website based on the extracted fashion trend information The method may further include recommending and providing one or more second product sales contents among a plurality of second product sales contents to the user.

A product sales content posting server using artificial intelligence-based website crawling according to another embodiment of the present invention for solving the above problems is loaded into a processor, a network interface, a memory, and the memory, and is operated by the processor. A computer program that is executed, wherein the computer program includes instructions for collecting a URL of a first website on which first product sales content is posted, and information about the first product sales content using the collected URL. It may include an instruction for generating crawl result data by performing crawling, and an instruction for posting second product sales content on a second website using the generated crawl result data.

A computer program recorded on a computer-readable recording medium according to another embodiment of the present invention for solving the above problems includes the steps of collecting a URL of a first website where a first product sales content is posted; Generating crawl result data by performing crawling on the first product sales content using the collected URLs, and posting second product sales content on a second website using the generated crawl result data. It can be stored in a computer-readable recording medium in order to execute.

Other specific details of the invention are included in the detailed description and drawings.

According to various embodiments of the present invention, product sales content posted on a first website to be crawled is crawled to generate crawl result data, and through this, the product sales content is posted on a second website to be posted. By posting sales content, it is possible to import and post product sales content more quickly and conveniently.

In addition, by bringing a plurality of product sales contents posted on multiple websites and integrating and providing them through one website, it is possible to check and order products registered on multiple websites as one website is used. .

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

1 is a diagram illustrating a system for posting product sales content using artificial intelligence-based website crawling according to an embodiment of the present invention.
2 is a hardware configuration diagram of a product sales content publishing server using artificial intelligence-based website crawling according to another embodiment of the present invention.
3 is a flowchart of a method of posting merchandise sales content using artificial intelligence-based website crawling according to another embodiment of the present invention.
4 is a diagram illustrating a process of generating crawl result data by crawling product sales content posted on a first domestic website, in various embodiments.
5 is a diagram illustrating a process of generating crawling result data by crawling product sales content posted on a first foreign website, according to various embodiments.
6 is a diagram illustrating a process of dividing each of a plurality of products when a plurality of products are included in one product image, in various embodiments.
7 is a diagram illustrating a process of extracting attribute information for each of a plurality of products in various embodiments.
8 is a diagram illustrating a process of posting product sales content on a second website using crawling data, in various embodiments.
9 is a diagram illustrating a process of searching product sales content through an image search function of a second website, in various embodiments.
10 and 11 are diagrams illustrating a process of recommending and providing product sales content that sells a product similar to product sales content selected by a user, in various embodiments.
12 is a diagram illustrating a personalized recommendation service provided by a product sales content posting server using artificial intelligence-based website crawling in various embodiments.
13 is a diagram illustrating a marketing automation service provided by a product sales content publishing server using artificial intelligence-based website crawling in various embodiments.
14 and 15 are diagrams illustrating category data of products applicable in various embodiments.
16 is a diagram illustrating style data of products applicable in various embodiments.
17 is a diagram illustrating color data of products applicable in various embodiments.
18 is a diagram illustrating size data of products applicable to various embodiments.
19 is a diagram illustrating tariff code (HS CODE) data of products applicable to various embodiments.

Advantages and features of the present invention, and methods of achieving them, will become clear with reference to the detailed description of the following embodiments taken in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, only these embodiments are intended to complete the disclosure of the present invention, and are common in the art to which the present invention belongs. It is provided to fully inform the person skilled in the art of the scope of the invention, and the invention is only defined by the scope of the claims.

Terminology used herein is for describing the embodiments and is not intended to limit the present invention. In this specification, singular forms also include plural forms unless specifically stated otherwise in a phrase. As used herein, "comprises" and/or "comprising" does not exclude the presence or addition of one or more other elements other than the recited elements. Like reference numerals throughout the specification refer to like elements, and “and/or” includes each and every combination of one or more of the recited elements. Although "first", "second", etc. are used to describe various components, these components are not limited by these terms, of course. These terms are only used to distinguish one component from another. Accordingly, it goes without saying that the first element mentioned below may also be the second element within the technical spirit of the present invention.

Unless otherwise defined, all terms (including technical and scientific terms) used in this specification may be used with meanings commonly understood by those skilled in the art to which the present invention belongs. In addition, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly specifically defined.

The term "unit" or "module" used in the specification means a hardware component such as software, FPGA or ASIC, and "unit" or "module" performs certain roles. However, "unit" or "module" is not meant to be limited to software or hardware. A “unit” or “module” may be configured to reside in an addressable storage medium and may be configured to reproduce one or more processors. Thus, as an example, a “unit” or “module” may refer to components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays and variables. Functions provided within components and "units" or "modules" may be combined into smaller numbers of components and "units" or "modules" or may be combined into additional components and "units" or "modules". can be further separated.

The spatially relative terms "below", "beneath", "lower", "above", "upper", etc. It can be used to easily describe a component's correlation with other components. Spatially relative terms should be understood as including different orientations of elements in use or operation in addition to the orientations shown in the drawings. For example, if you flip a component that is shown in a drawing, a component described as "below" or "beneath" another component will be placed "above" the other component. can Thus, the exemplary term “below” may include directions of both below and above. Components may also be oriented in other orientations, and thus spatially relative terms may be interpreted according to orientation.

In this specification, a computer means any kind of hardware device including at least one processor, and may be understood as encompassing a software configuration operating in a corresponding hardware device according to an embodiment. For example, a computer may be understood as including a smartphone, a tablet PC, a desktop computer, a laptop computer, and user clients and applications running on each device, but is not limited thereto.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Although each step described in this specification is described as being performed by a computer, the subject of each step is not limited thereto, and at least a part of each step may be performed in different devices according to embodiments.

In addition, in the method of posting product sales contents using artificial intelligence-based website crawling described in this specification, an online shopping mall operator that sells clothes online to general customers can view products through websites in a number of domestic and foreign wholesale and retail markets. It is described as providing "one-stop platform and website for shopping mall owners" that not only crawls and integrates product sales contents so that they can conveniently check and order, but also manages orders. However, this is just one example. However, it is not limited thereto, and the same can be applied to online shopping malls for general customers or products other than clothes. Hereinafter, a method of providing a one-stop platform and website for shopping mall owners will be described with reference to FIGS. 1 to 19 .

1 is a diagram illustrating a system for posting product sales content using artificial intelligence-based website crawling according to an embodiment of the present invention.

Referring to FIG. 1, a product sales content posting system using artificial intelligence-based website crawling according to an embodiment of the present invention includes a product sales content posting server 100, a user terminal 200, and an external server 300. can do.

Here, the product sales content posting system using artificial intelligence-based website crawling shown in FIG. 1 is according to an embodiment, and its components are not limited to the embodiment shown in FIG. 1, and additional, if necessary, may be changed or deleted.

In one embodiment, the product sales content publishing server 100 (hereinafter referred to as "server 100") may provide a one-stop platform and website for shopping mall owners.

Here, the one-stop platform and website for shopping mall owners provided by the server 100 allows business operators to purchase products they want to sell through their shopping mall without having to search multiple wholesale and retail websites through a single website. This may mean that not only information on corresponding products can be integrated and checked, but also orders for products registered on a plurality of websites can be placed through the corresponding website.

For example, the server 100 collects information by crawling product sales contents from websites (eg, a first website) operated in domestic and foreign wholesale markets where users (eg, shopping mall owners) order products, and , The collected information can be automatically posted on a website operated by the server 100 (eg, a second website).

In addition, the server 100 may process orders and payments for products through the second website, outputs an order form for products ordered and paid for, and processes the order on the first website where the product is registered. By proceeding, the order can be acted as an agent.

In various embodiments, the server 100 may be connected to the user terminal 200 through the network 400, and may provide a connection link (eg, URL) of the second website to the user terminal 200, It may be implemented in a form in which the user can access the second website operated by the server 100 through the URL. Here, the method of posting product sales content using artificial intelligence-based website crawling provided by the server 100 is described as being implemented in the form of a web and provided to users, but is not limited thereto, and is provided in the form of an application. can be implemented as

Here, the network 400 may refer to a connection structure capable of exchanging information between nodes such as a plurality of terminals and servers. For example, the network 400 includes a local area network (LAN), a wide area network (WAN), a world wide web (WWW), a wired and wireless data communication network, a telephone network, a wired and wireless television communication network, and the like. do.

In addition, here, the wireless data communication networks are 3G, 4G, 5G, 3GPP (3rd Generation Partnership Project), 5GPP (5th Generation Partnership Project), LTE (Long Term Evolution), WIMAX (World Interoperability for Microwave Access), Wi-Fi (Wi-Fi) Fi), Internet, LAN (Local Area Network), Wireless LAN (Wireless Local Area Network), WAN (Wide Area Network), PAN (Personal Area Network), RF (Radio Frequency), Bluetooth network, A Near-Field Communication (NFC) network, a satellite broadcasting network, an analog broadcasting network, a Digital Multimedia Broadcasting (DMB) network, and the like are included, but are not limited thereto.

In addition, here, the user terminal 200 is a wireless communication device that ensures portability and mobility, and includes navigation, PCS (Personal Communication System), GSM (Global System for Mobile communications), PDC (Personal Digital Cellular), PHS (Personal Handyphone) System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), Wibro (Wireless Broadband Internet) terminal, smart All types of handheld-based wireless communication devices such as smartphones, smart pads, tablet PCs, etc. may be included, but are not limited thereto.

In various embodiments, the external server 300 may be connected to the server 100 through the network 400, and generated as the server 100 performs a product sales content posting process using artificial intelligence-based website crawling. Various information and data (eg, crawl result data) can be stored and managed. For example, the external server 300 may be a storage server provided separately outside the server 100, but is not limited thereto. Hereinafter, with reference to FIG. 2 , the hardware configuration of the server 100 providing a method of posting merchandise sales content using artificial intelligence-based website crawling will be described.

2 is a hardware configuration diagram of a product sales content publishing server using artificial intelligence-based website crawling according to another embodiment of the present invention.

Referring to FIG. 2 , in various embodiments, the server 100 includes one or more processors 110, a memory 120 that loads a computer program 151 executed by the processor 110, and a bus 130. ), a communication interface 140 and a storage 150 for storing the computer program 151. Here, in FIG. 2, only components related to the embodiment of the present invention are shown. Therefore, those skilled in the art to which the present invention pertains can know that other general-purpose components may be further included in addition to the components shown in FIG. 2 .

The processor 110 controls the overall operation of each component of the server 100. The processor 110 includes a Central Processing Unit (CPU), a Micro Processor Unit (MPU), a Micro Controller Unit (MCU), a Graphic Processing Unit (GPU), or any type of processor well known in the art of the present invention. It can be.

Also, the processor 110 may perform an operation for at least one application or program for executing a method according to embodiments of the present invention, and the server 100 may include one or more processors.

In various embodiments, the processor 110 may temporarily and/or permanently store signals (or data) processed in the processor 110 (RAM: Random Access Memory, not shown) and ROM (ROM: Read -Only Memory, not shown) may be further included. In addition, the processor 110 may be implemented in the form of a system on chip (SoC) including at least one of a graphics processing unit, RAM, and ROM.

Memory 120 stores various data, commands and/or information. Memory 120 may load computer program 151 from storage 150 to execute methods/operations according to various embodiments of the present invention. When the computer program 151 is loaded into the memory 120, the processor 110 may perform the method/operation by executing one or more instructions constituting the computer program 151. The memory 120 may be implemented as a volatile memory such as RAM, but the technical scope of the present disclosure is not limited thereto.

The bus 130 provides a communication function between components of the server 100 . The bus 130 may be implemented in various types of buses such as an address bus, a data bus, and a control bus.

The communication interface 140 supports wired and wireless Internet communication of the server 100 . Also, the communication interface 140 may support various communication methods other than internet communication. To this end, the communication interface 140 may include a communication module well known in the art. In some embodiments, communication interface 140 may be omitted.

The storage 150 may non-temporarily store the computer program 151 . When product sales content posting process using artificial intelligence-based website crawling is performed through the server 100, the storage 150 stores various information necessary to provide the product sales content posting process using artificial intelligence-based website crawling. can be saved

The storage 150 may be a non-volatile memory such as read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, or the like, a hard disk, a removable disk, or a device well known in the art. It may be configured to include any known type of computer-readable recording medium.

Computer program 151 may include one or more instructions that when loaded into memory 120 cause processor 110 to perform methods/operations in accordance with various embodiments of the invention. That is, the processor 110 may perform the method/operation according to various embodiments of the present disclosure by executing the one or more instructions.

In one embodiment, the computer program 151 collects the URL of the first website on which the first product sales content is posted, performs a crawl on the first product sales content using the collected URL, and results in crawling data. At least one method for performing a product sales content posting method using artificial intelligence-based website crawling, which includes the step of generating and posting second product sales content on a second website using the generated crawling result data. may contain instructions.

Steps of a method or algorithm described in connection with an embodiment of the present invention may be implemented directly in hardware, implemented in a software module executed by hardware, or implemented by a combination thereof. A software module may include random access memory (RAM), read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, hard disk, removable disk, CD-ROM, or It may reside in any form of computer readable recording medium well known in the art to which the present invention pertains.

Components of the present invention may be implemented as a program (or application) to be executed in combination with a computer, which is hardware, and stored in a medium. Components of the present invention may be implemented as software programming or software elements, and similarly, embodiments may include various algorithms implemented as data structures, processes, routines, or combinations of other programming constructs, such as C, C++ , Java (Java), can be implemented in a programming or scripting language such as assembler (assembler). Functional aspects may be implemented in an algorithm running on one or more processors. Hereinafter, with reference to FIGS. 3 to 8 and 14 to 19 , a description will be given of a method of posting merchandise sales content using artificial intelligence-based website crawling performed by the server 100 .

3 is a flowchart of a method of posting merchandise sales content using artificial intelligence-based website crawling according to another embodiment of the present invention.

Referring to FIG. 3 , in step S110 , the server 100 may collect the URL (Uniform Resource Locator) of the first website (eg, 10 in FIGS. 4 and 5 ) on which the first product sales content is posted.

In various embodiments, the server 100 may provide a user interface (UI) to the user terminal 200 of an administrator (eg, an administrator operating the second website), and the user interface (UI) to be crawled through the UI. A URL for a page of the first website on which product sales content is posted may be directly input.

In various embodiments, when a specific website is set as a crawling target by an administrator, the server 100 may automatically collect URLs for each page of a plurality of product sales content posted on the specific website.

In addition, the server 100 may automatically collect the URL of the page where the new product sales content is posted whenever new product sales content is posted on a website set as a crawling target. However, the server 100 is not limited thereto, and when a specific website is set as a target for crawling, the server 100 may check whether new product sales content is posted on the specific website at predetermined intervals, and when the new product sales content is posted. URLs of pages on which new product sales contents are posted may be collected, that is, URLs may be collected at predetermined intervals.

In various embodiments, the server 100 may receive input of product sales content itself, rather than the URL of a page on which the product sales content to be crawled is posted. For example, the server 100 may receive a captured image generated by capturing specific product sales content itself, rather than the URL of a page on which the specific product sales content is posted, on the first website. However, it is not limited thereto.

In step S120, the server 100 performs crawling on the first product sales content using the URL collected in step S110 (or an image captured directly uploaded product sales content, etc.) to generate crawl result data. can do.

First, referring to FIG. 4 , when the server 100 receives the URL 11 of the page on which the first product sales content posted on the first website 10 is posted, the first through the URL 11 All information included in the first product sales content (eg, product image 12, product information 13 ) and product sales store information 14, etc.) may be collected.

Thereafter, the server 100 may generate crawl result data 20 by processing information collected by performing crawling on the first product sales content. Here, the crawling result data 20 shown in FIG. 4 is implemented in the form of Excel data, but is not limited thereto.

Here, the product code (store number) included in the crawling result data (20), English product name (purchased product name), supplier product name (store information), model name (URL), product summary description (store information), product detailed description ( Mixed fusion and image), supply price (purchase price), option input (color and size), color setting, image registration (detail, list, small list and reduction), and manufacturer information are crawled for the first product sales content. It may be information collected by performing, but is not limited thereto.

In addition, the product classification number included in the crawling result data 20 may be information determined according to the property of the product determined by analyzing the product image included in the first product sales content, but is not limited thereto. Here, the method of determining the property of the product by the server 100 analyzing the product image included in the first product sales content may be performed using a pre-learned image analysis model, and a detailed method will be described later.

In addition, the product name included in the crawling result data 20 is crawled for the attribute (eg, category) of the product determined according to the above method (eg, a method using a pre-learned image analysis model) and the first product sales content. It may be a combination of purchased product names (English product names) collected by performing, but is not limited thereto.

In addition, the search term setting item included in the crawl result data 20 may be directly input by a manager who wants to crawl the first product sales content on the first website 10, but is not limited thereto, and is not limited thereto. : analyzing the first product sales content, extracting a plurality of keywords, setting top n keywords as search words according to the number of extracted keywords, etc.) may be applied.

In addition, display status, sales status, option use, item composition method, option display method, option style, supplier and brand information items included in the crawl result data 20 are set in advance by the manager who wants to crawl the first product sales content. It may be a fixed value, but is not limited thereto.

In various embodiments, when one or more foreign languages are included in the content for selling the first product, the server 100 automatically translates the one or more foreign languages into a preset language from the user and performs crawling, thereby generating crawling result data. .

For example, as shown in FIG. 5 , the server 100 inputs the URL 11 of a page where the first product sales content posted on the first website 10 in a foreign country (eg, China) is posted. In this case, access the page where the first product sales content is posted through the URL 11 and perform a crawl on the first product sales content, but among the information collected as a result of the crawl, the information written in a foreign language (Chinese) is sent to the administrator It can be translated into a preset language (eg, Korean), and crawling result data can be generated using the translated information.

At this time, in order to automatically translate the foreign language included in the first product sales content into a preset language, the server 100 converts data converted into a preset language for each of a plurality of pieces of information included in the first product sales content in advance. can be recorded and stored (e.g. FIGS. 17 and 18)

At this time, when the first product sales content to be crawled is the foreign first website 10, the server 100 performs overseas customs clearance based on the property (eg, category) of the product included in the first product sales content. HS CODE can be extracted, and the extracted HS CODE can be added to the crawling result data (eg, FIG. 19).

In various embodiments, the server 100 may determine a crawling method according to the type of website on which the first product selling content to be crawled is posted. For example, as shown in FIGS. 4 and 5, a form in which product sales content is posted is determined on a website, and the form is different for each website, so the same crawling method is applied to different websites. It can be difficult to do.

In consideration of this, the server 100 performs a crawling method (e.g., an area to be crawled (e.g., 12, 13, and 14 in FIGS. 4 and 5) for each website on which product sales content to be crawled is posted. Whether information needs to be translated, etc.) can be set in advance, and crawling results can be generated more quickly by performing crawling by applying the crawling method set in response to the URL entered by the administrator.

Here, the crawling method set for each website may be a value set by an administrator, but is not limited thereto, and may be automatically set based on a crawling history of product sales content posted on each website.

In various embodiments, when the server 100 sets the first website 10 as a crawling target, the first website 10 is posted on the first website 10 from the time it is set as a crawling target. Crawling for new product sales content can be performed automatically. At this time, the server 100 may crawl the new product sales content in real time in response to the new product sales content being posted on the first website 10 set as a crawling target, but is not limited thereto, and at a preset period. It is possible to automatically crawl one or more new product sales content posted on the first website 10 set as a crawling target every time.

In various embodiments, when the server 100 is set as a crawling target for the first website 10, the server 100 automatically crawls the plurality of first product sales content posted on the first website 10, , It is possible to crawl only one or more first product sales contents that satisfy a predetermined crawling condition from the user. For example, the server 100 automatically crawls a plurality of first product sales content posted on the first website 10 set as a crawling target, but when receiving a condition of “for women” from the user, a plurality of Among the first product sales content, only first product sales content for women's products or unisex products may be crawled.

In various embodiments, the server 100 analyzes the product images included in the first product sales content, and when two or more products are included in the product image, that is, the first product sales content sells products that sell two or more products. In the case of content, two or more crawling result data may be generated by classifying information collected as content for sale of the first product is crawled into two or more products, respectively.

First, the server 100 may identify a product included in the product image by analyzing the product image using the first image analysis model 30 (eg, based on the SSD model) as shown in FIG. 6, It may be determined whether the number of identified products is two or more.

Here, the first image analysis model 30 may be a supervised model using a plurality of product images labeled with product information (eg, product name, product type, wearing position, etc.) as learning data, , It may be a model that receives a product image and extracts a result of identifying a product included in the product image.

Here, the first image analysis model 30 is composed of one or more network functions, and the one or more network functions may be composed of a set of interconnected calculation units that may be generally referred to as 'nodes'. These 'nodes' may also be referred to as 'neurons'. One or more network functions include at least one or more nodes. Nodes (or neurons) that make up one or more network functions can be interconnected by one or more 'links'.

Within the first image analysis model 30 , one or more nodes connected through links may form a relative relationship of an input node and an output node. The concept of an input node and an output node is relative, and any node in an output node relationship with one node may have an input node relationship with another node, and vice versa. As described above, the input node to output node relationship can be created around the link. More than one output node can be connected to one input node through a link, and vice versa.

In a relationship between an input node and an output node connected through one link, the value of the output node may be determined based on data input to the input node. Here, a node interconnecting an input node and an output node may have a weight. The weight may be variable, and may be variable by a user or an algorithm in order to perform a function desired by the first image analysis model 30 . For example, when one or more input nodes are interconnected by respective links to one output node, the output node is set to a link corresponding to values input to input nodes connected to the output node and respective input nodes. An output node value may be determined based on the weight.

As described above, in the first image analysis model 30 , one or more nodes are interconnected through one or more links to form an input node and output node relationship in the first image analysis model 30 . The characteristics of the first image analysis model 30 are determined according to the number of nodes and links in the first image analysis model 30, the relationship between the nodes and links, and the value of the weight assigned to each link. can be determined For example, when there are two first image analysis models 30 having the same number of nodes and links and different weight values between the links, the two first image analysis models 30 are considered to be different from each other. can be recognized.

Some of the nodes constituting the first image analysis model 30 may form one layer based on distances from the first input node. For example, a set of nodes having a distance of n from the first input node may constitute n layers. The distance from the first input node may be defined by the minimum number of links that must be passed through to reach the corresponding node from the first input node. However, the definition of such a layer is arbitrary for explanation, and the order of a layer in the first image analysis model 30 may be defined in a different way from the above-described method. For example, a layer of nodes may be defined by a distance from a final output node.

The first input node may refer to one or more nodes to which data is directly input without going through a link in relation to other nodes among nodes in the first image analysis model 30 . Alternatively, in the relationship between nodes based on a link in the network of the first image analysis model 30, it may mean nodes that do not have other input nodes connected by a link. Similarly, the final output node may refer to one or more nodes that do not have an output node in relation to other nodes among nodes in the first image analysis model 30 . Also, the hidden nodes may refer to nodes constituting the first image analysis model 30 other than the first input node and the last output node. In the first image analysis model 30 according to an embodiment of the present invention, the number of nodes of the input layer may be greater than that of the hidden layer close to the output layer, and the number of nodes decreases as the process progresses from the input layer to the hidden layer. It may be the first image analysis model 30 of

The first image analysis model 30 may include one or more hidden layers. A hidden node of a hidden layer may use outputs of previous layers and outputs of neighboring hidden nodes as inputs. The number of hidden nodes for each hidden layer may be the same or different. The number of nodes of the input layer may be determined based on the number of data fields of the input data and may be the same as or different from the number of hidden nodes. Input data input to the input layer may be operated by a hidden node of the hidden layer and may be output by a fully connected layer (FCL) that is an output layer.

In addition, supervised learning is a method of generating learning data by labeling specific data and information related to the specific data, and learning using the same. In this method, learning data is generated by labeling two data having a causal relationship, and the generated learning is performed. It means learning from data.

More specifically, the server 100 may perform learning on one or more network functions constituting an artificial intelligence model using labeled training data. For example, the server 100 inputs each of the learning input data to one or more network functions, and compares each of the output data calculated by the one or more network functions with each of the learning output data corresponding to the label of each of the learning input data to make an error. can be derived. That is, in training of an artificial intelligence model, learning input data may be input to an input layer of one or more network functions, and learning output data may be compared with outputs of one or more network functions.

The server 100 may train an artificial intelligence model based on an error between an operation result of one or more network functions for learning input data and a learning output data (label).

Also, the server 100 may adjust weights of one or more network functions based on the error in a backpropagation method. That is, the server 100 may adjust the weight so that the output of the one or more network functions approaches the learning output data based on the error between the calculation result of the one or more network functions for the learning input data and the learning output data.

When learning of one or more network functions is performed for a predetermined epoch or more, the server 100 may determine whether to stop learning using verification data. The predetermined epochs may be part of an overall learning target epoch.

Verification data may consist of at least some of the labeled training data. That is, the server 100 performs learning of the artificial intelligence model through the learning data, and after the learning of the artificial intelligence model is repeated at least a predetermined epoch, the learning effect of the artificial intelligence model exceeds a predetermined level using the verification data. It can be judged whether or not For example, when the server 100 performs learning with a target repetition learning number of 10 using 100 learning data, after performing repetition learning 10 times, which is a predetermined epoch, using 10 verification data, 3 By performing repeated learning, if the change in the output of the artificial intelligence model during the three iterative learning is less than a predetermined level, it is determined that further learning is meaningless and the learning may be terminated.

That is, the verification data can be used to determine the completion of learning based on whether the effect of learning per epoch is greater than or less than a certain level in repeated learning of the artificial intelligence model. The above-described number of learning data, verification data, and number of repetitions are only examples, and the present invention is not limited thereto.

The server 100 may generate an artificial intelligence model by testing performance of one or more network functions using test data and determining whether to activate one or more network functions. The test data may be used to verify the performance of the artificial intelligence model and may be composed of at least a part of training data. For example, 70% of the training data can be used for training of an AI model (i.e., learning to adjust weights to output similar results to labels), and 30% of the training data can be used to verify the performance of an AI model. It can be used as test data for The server 100 may determine whether to activate the artificial intelligence model according to whether or not the performance exceeds a predetermined level by inputting test data to the artificial intelligence model for which learning has been completed and measuring an error.

The server 100 verifies the performance of the trained AI model using test data for the AI model that has been trained, and uses the AI model in other applications if the performance of the AI model that has been trained is above a predetermined standard. can be activated to do so.

In addition, the server 100 may deactivate and discard the AI model when the performance of the AI model that has been trained is below a predetermined standard. For example, the server 100 may determine the performance of the generated artificial intelligence model based on factors such as accuracy, precision, and recall. The performance evaluation criteria described above are only examples and the present invention is not limited thereto. According to an embodiment of the present invention, the server 100 may independently train each artificial intelligence model to generate a plurality of artificial intelligence models, evaluate performance, and use only artificial intelligence models having a certain performance or higher. However, it is not limited thereto.

Next, as a result of analyzing product images using the first image analysis model 30, the server 100 identifies three products, a short-sleeved T-shirt, a bag, and shoes as shown in FIG. By performing crawling on product sales content, the collected information can be classified into three products to generate crawl result data about short-sleeved T-shirts, crawl result data about bags, and crawl result data about shoes.

In step S130 , the server 100 may post second product sales content on a second website (eg, 50 in FIG. 8 ) using the crawling result data generated in step S120 .

Here, the second product sales content is processed by processing the information on the first product sales content posted on the first website 10 according to the upload method and template of the second website 50, and the second website 50 It may refer to product sales content posted on the platform.

For example, when the crawling result data is generated through step S120, the server 100 processes the information included in the crawling result data according to the form of the second website 50, so that the second website 50 Second product sales content may be automatically posted on the screen. However, the server 100 is not limited thereto, and the server 100 may receive at least one crawl result data selected from the manager among the plurality of crawl result data generated through step S120, and use only the selected crawl result data to visit the second website ( 50) Second product sales content may be posted on the screen.

In various embodiments, the server 100 may analyze product images included in the crawl result data to extract product attributes (eg, categories and styles) included in the product images, and attributes of the extracted products. Accordingly, second product sales content corresponding to the crawling result data may be posted.

First, the server 100 analyzes product images using the second image analysis model 40 (eg, based on SE-Resnext101) as shown in FIG. Jackets, zip-ups, jumpers, jean jackets, cardigans, etc.), knits, shirts and blouses (basic shirts, patterned shirts, blouses, etc.), tops (man-to-man, hoodies, long sleeves, short sleeves, vests, etc.), bottoms (cotton pants, jeans, slacks) , wide, shorts, overalls, etc.), one-piece (check, flower, long, mini, etc.), set (two-piece, jumpsuit, sweatpants, etc.) and style (e.g. simple basic, lovely, kitsch, unique, romantic) , sexy glam, street, sporty, office look, missy, home wear, beach wear, etc.) information (e.g., FIGS. 14 and 15) can be extracted.

Here, the second image analysis model 40 may be a supervised model using a plurality of product images labeled with product category and style information as training data.

Here, the method of extracting the attribute information on the product is to generate a second image analysis model 40 different from the first image analysis model 30, and use the second image analysis model 40 to image the product. Although it is described as extracting attribute information about a product by analyzing , this is only one example and is not limited thereto, and the function of the first image analysis model 30 and the second image using one image analysis model The function of the analysis model 40 can be performed simultaneously.

Thereafter, the server 100 may post second product sales content in an area corresponding to attribute information on the extracted product on the second website 50 . For example, as shown in FIG. 8 , when the product category included in the product image is a dress, the server 100 displays a second page corresponding to the corresponding crawling result data in the dress category area of the second website 50 . Product sales content can be automatically uploaded.

In various embodiments, the server 100 analyzes the product image and posts the second product sales content corresponding to the crawl result data including the corresponding product image based on the attribute information of the extracted product, but to the extracted product. Based on the accuracy of the attribute information for, it may be determined whether to post the second product sales content.

First, the server 100 may calculate the accuracy of attribute information on the extracted product by analyzing the product image.

For example, the server 100 may obtain feedback information on a product image or a product included in the product image, by analyzing the product image and comparing attribute information and feedback information on the extracted product, by analyzing the product image It is possible to calculate the accuracy of the attribute information for the extracted product (eg, 90% accuracy when 9 pieces of feedback information out of 10 pieces of feedback information are the same as the extracted attribute information).

Here, the feedback information may be input of the category and style of the product included in the product image in response to providing the product image to a plurality of users (preset users or random users). It is not limited, and may be product category and style information extracted by analyzing comments and reviews left by a plurality of users who have purchased the corresponding product (eg, keyword analysis).

As another example, the server 100 may extract N pieces of attribute information about a product by analyzing the same product image or a plurality of images included in the same product sales content N times using the second image analysis model, and extracting N pieces of attribute information about the product. Accuracy can be calculated by comparing the N pieces of attribute information (eg, when 9 pieces of attribute information out of 10 pieces of attribute information are the same, the accuracy is 90%).

Thereafter, when the calculated accuracy is greater than or equal to a predetermined value (eg, 90% or more), the server 100 may post second product sales content corresponding to the crawl result data including the corresponding product image.

Meanwhile, when the calculated accuracy is less than a predetermined value, the server 100 may analyze the product image and provide the extracted product attribute information to the manager so that the manager checks and corrects the product attribute information. . At this time, the server 100 may generate learning data by labeling the attribute information confirmed and corrected by the manager on the corresponding product image, and by using the generated learning data to train the second image analysis model 40, The accuracy of the second image analysis model 40 may be continuously increased.

In various embodiments, the server 100 may determine whether or not the second product sales content corresponding to the crawling result data is posted by determining whether one or more pieces of information included in the crawling result data are abnormal.

First, the server 100 may determine whether or not information necessary for posting second product sales content is included in the crawl result data. For example, the server 100 requires first to fifth information to post second product sales content based on a posting form of second product sales content previously set on the second website 50 . In this case, it is possible to determine whether the first to fifth information is present in the crawling result data.

Thereafter, the server 100 may determine whether one or more pieces of information included in the crawl result data are abnormal based on pre-stored data (eg, FIGS. 14 to 19 ). For example, the server 100 analyzes whether the color information field of the product included in the crawl result data includes content related to color or content not related to color, and determines whether there is an abnormality in the color information. and when it is determined that the information included in the color information area is information not related to color, it may be determined that there is an abnormality in the color information.

Here, the method of determining the abnormality of the color information may be determined by analyzing the color information included in the crawling result data and determining whether the extracted keyword matches the previously stored color data (eg, FIG. 17), but is limited thereto. It doesn't work.

Thereafter, when it is determined that one or more pieces of information included in the crawling result data are normal, the server 100 may post second product sales content corresponding to the corresponding crawling result data.

Meanwhile, when it is determined that one or more pieces of information included in the crawling result data are abnormal, the server 100 requests the manager to confirm information included in the crawling result data (in particular, information determined to be abnormal). can At this time, if it is determined that the color information included in the crawling result data is abnormal, the server 100 may grant a correction right to directly modify the color information of the crawling result data from the manager, and the color information corrected by the manager. If the information is information that is not recorded in the pre-stored color data, the color data may be updated by reflecting it.

In various embodiments, the server 100 compares one or more pieces of information included in the crawl result data generated while crawling the first product sales content with one or more pieces of information included in each of a plurality of pre-stored crawl result data. The similarity may be calculated, and the second product sales content corresponding to the corresponding crawling result data may be posted only when the maximum value of the calculated similarity is less than a preset value.

In general, since the same product sales content is not posted on one website, there is no probability that the same crawl result data will be generated. There is a possibility that data will be generated as a result of crawling, and if it is posted on the second website 50 without filtering, there is a problem that the same second product sales content is posted, causing confusion among buyers.

Considering this point, the server 100 compares one or more pieces of information included in the crawling result data generated through step S120 with one or more pieces of information included in each of a plurality of pre-stored crawling result data, and the similarity calculated is set in advance. The above problem can be solved by posting the second product sales content using the corresponding crawl result data only when there is no product sales content less than the value, that is, the same or excessively similar.

In various embodiments, the server 100 performs a crawl on one first product sales content and when two or more crawling result data are generated (one first product sales content includes information on two or more products). When two or more crawling result data are generated by dividing it), two or more second product sales contents may be posted using each of the two or more crawl result data, but the two or more second product sale contents may be matched with each other.

Here, mutual matching may mean that when a buyer selects any one of two or more matched second product sales contents, the rest of the second product sales contents that are not selected are also provided. have. For example, as shown in FIG. 6 , when a single product image includes a short-sleeved T-shirt, a bag, and shoes to generate a total of three crawling result data, the server 100 uses the three crawling result data to generate short-sleeved T-shirts, bags, and shoes. T-shirt sales content, bag sales content, and shoe sales content can be posted, and short-sleeved T-shirt sales content, bag sales content, and shoe sales content can be matched with each other. Thereafter, when the user selects the short-sleeved T-shirt sales content, the server 100 may provide not only the short-sleeved T-shirt sales content, but also matching bag sales content and shoe sales content.

In various embodiments, the server 100 performs crawling on the first product sales content to generate first crawl result data, and uses it to continuously search for the first product sales content even after posting the second product sales content. By performing crawling on the content, the changed content may be reflected in the second product sales content.

For example, the server 100 may generate first crawl result data by performing crawling on first product sales content, and post second product sales content using the result data. Thereafter, the server 100 may generate second crawl result data by re-crawling the first product sales content at predetermined intervals from the time when the first crawl result data is generated, and may generate second crawl result data and Second product sales content posted in correspondence to the first crawl result data may be corrected by comparing the second crawl result data.

When collecting information by crawling the first product sales content posted on the first webpage and posting the second product sales content through it, the second product sales content is registered using only the information collected initially, There is a problem in that changes in the content of the original product sales content (eg, a product of a specific color is out of stock, a price change, etc.) cannot be continuously reflected.

In consideration of this, the server 100 continuously crawls the first product sales content even after posting the second product sales content by using the crawl result data generated by crawling the first product sales content, so that the content is changed. It is possible to continuously check whether or not the contents have been changed, and if the contents are changed, the second product sales contents may be corrected accordingly.

The above-mentioned method of posting product sales content using artificial intelligence-based web site crawling has been described with reference to the flow chart shown in the drawings. For a brief explanation, the method of posting product sales content using artificial intelligence-based website crawling has been illustrated and described as a series of blocks, but the present invention is not limited to the order of the blocks, and some blocks are shown and operated in this specification. may be performed in a different order or concurrently. In addition, new blocks not described in the present specification and drawings may be added, or some blocks may be deleted or changed. Hereinafter, various services provided by the server 100 through the second website 50 will be described with reference to FIGS. 9 to 13 .

9 is a diagram illustrating a process of searching product sales content through an image search function of a second website, in various embodiments.

Referring to FIG. 9 , in various embodiments, the server 100 receives an input of an image including one or more products from a user (eg, a business operator operating an online shopping mall or a general customer) through the second website 50. And, by analyzing the image using a second image analysis model (eg, 40 of FIG. 7 ), attribute information (eg, category, style, color, shape, etc.) of one or more products included in the image can be extracted. Among the plurality of second product sales contents posted on the second website 50, one or more second product sales contents corresponding to the extracted attribute information may be provided as a search result.

At this time, if there are two or more products included in the image input from the user, the server 100 may individually extract attribute information for each of the two or more products, and each of the two or more products based on the individually extracted attribute information. You can provide search results for .

Here, the server 100, when there is no one or more second product sales contents corresponding to the attribute information of one or more products included in the image, the product included in the image input by the user based on the attribute information of the one or more products. It is possible to recommend and provide second product sales content that sells similar products. Hereinafter, it will be described with reference to FIGS. 10 and 11.

10 and 11 are diagrams illustrating a process of recommending and providing product sales content that sells a product similar to product sales content selected by a user, in various embodiments.

10 and 11 , in various embodiments, the server 100, when there is no second product sales content corresponding to the product property information included in the image input by the user, based on the product property information Second product sales content having attribute information similar to the corresponding product may be selected as recommended product sales content and provided to the user.

Here, the second product sales content having similar attribute information may refer to second product sales content in which the same attribute information as the product attribute information extracted by analyzing the image is included in a predetermined ratio or more, but is not limited thereto. don't

In various embodiments, when the server 100 receives a request for providing second product sales content for a product similar to a specific product from the user, in response to the request, the server 100 provides a product having attribute information similar to the specific product being viewed by the user. 2 Product sales content can be provided.

For example, the server 100 provides a page that outputs detailed information on the selected second product sales content in response to the user selecting specific second product sales content, but displays a "similar product" displayed on the corresponding page. When the "More" button is selected, other product sales content having properties similar to those of the second product sales content selected by the user may be recommended and provided.

12 is a diagram illustrating a personalized recommendation service provided by a product sales content posting server using artificial intelligence-based website crawling in various embodiments.

Referring to FIG. 12 , in various embodiments, the server 100 may receive a search word input from a user (online shopping mall operator) through a second website 50, and a plurality of search words posted on the second website 50 may be input. A list of one or more second product sales contents corresponding to an input search word among the second product sales contents of the user may be provided.

In this case, the server 100 may provide content for selling recommended products to the user based on a search word input by the user and search history data of a plurality of users. For example, as shown in FIG. 12 , when receiving a “check” from the user, the server 100 may provide one or more second product sales contents including attributes of the check.

In addition, the server 100 selects N second product sales content in the order of the highest number of selections among the plurality of second product sales content selected by users who input "check" based on search history data of a plurality of users. and provide recommendations to users.

In addition, the server 100 recommends the top N search words to the user from the most input search word among a plurality of search words input simultaneously with "check" based on search history data of a plurality of users, or selects the top N search words. The used search results may be recommended to the user.

In addition, the server 100 may select recommended second product sales content for the user using behavioral data and profiling data (eg, search word input history, second product sales content selection history, order history, etc.) of a specific user. In addition, selected recommended second product sales content may be listed and provided to the user. For example, the server 100 extracts attribute information for each of the products requested by the user based on the user's order history, and uses the extracted attribute information to provide second product sales content to be recommended to the user. By selecting, product sales content with a high probability that the user will order may be recommended and provided.

In various embodiments, the server 100 may collect social data (eg, SNS data) for a plurality of products, extract fashion trend information by analyzing the collected social data, and extract the fashion trend information. Based on this, one or more second product sales contents among a plurality of second product sales contents posted on the second website 50 may be recommended and provided to the user.

For example, the server 100 may analyze social data for a plurality of products, extract keywords for each of the plurality of products, and attribute the extracted keywords (eg, total number of keywords, positive keywords). number of keywords and negative keywords) can be determined, and based on the properties of the determined keywords, a consumer real-time trend trend analysis report including information on popular products and information on unpopular products can be generated. can

Thereafter, the server 100 may recommend and provide second product sales content corresponding to a popular product or a product whose popularity is rapidly increasing to the user based on the consumer real-time trend trend analysis report.

In various embodiments, the server 100 may analyze sales data of a plurality of online shopping mall sites (eg, e-commerce sites) to select top N products in order of highest sales volume, and attribute information of the selected products. may be extracted, and second product sales content corresponding to the extracted product attribute information may be selected and recommended to the user.

13 is a diagram illustrating a marketing automation service provided by a product sales content publishing server using artificial intelligence-based website crawling in various embodiments.

Referring to FIG. 13 , in various embodiments, the server 100 may expose user-customized content to prevent users from leaving. For example, if the first user puts a first product in the shopping cart but the payment is not performed until a preset time elapses, the server 100 performs a promotion for the product the first user put in the shopping cart (eg, discount, additional accumulation, etc.), the first user may be induced to purchase the corresponding product.

In addition, the server 100 may provide a user-customized message or promotion according to the purpose based on the user's purpose of using the second website 50 (eg, discovery of new products, reordering of previously ordered products, etc.). For example, when the user intends to reorder a previously ordered product, a guide message guiding additional accumulation or issuance of a discount coupon for the previously ordered product may be provided. In addition, when the user visits the second website 50 for the purpose of finding out new products, the server 100 is a popular product among products for which the user has no order history (eg, a product with a large order quantity or a high rating). A guide message for guiding information about may be provided.

In addition, the server 100 may analyze and manage elements requiring improvement using behavioral data of each of a plurality of users. In addition, the server 100 may support a retargeting campaign to optimize retention and purchase conversion.

In various embodiments, the server 100 may provide an inventory management service (inventory AI bot) that manages the user's inventory based on the user's order history data. For example, the server 100 provides information on the first product on the online shopping mall operated by the user (e.g., number of views, click-through rate, shopping cart) for the first product purchased by the user through the second website 50. It is possible to calculate an inventory list to be held in advance for each of a plurality of products by analyzing the number of items put in the shopping cart, the number of days in the shopping cart, the purchase conversion rate, and the sales volume data.

For example, the server 100 learns information about each of a plurality of products (e.g., number of views, click-through rate, number of times put in a shopping cart, days in the shopping cart, purchase conversion rate, and sales data) and information about the amount of inventory to be retained accordingly. An inventory list to be held in advance for each of a plurality of products may be calculated using a pre-learned model based on data.

At this time, the server 100 may correct the calculated inventory amount by considering various external factors. For example, the server 100 provides a plurality of products based on the season (interlocking data from the Korea Meteorological Administration), other shopping malls (changes in price and sales volume of other shopping malls), shopping platform exhibitions and events, and corona (corona stage and group prohibition, etc.) For each, the amount of inventory to be held in advance can be corrected.

Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art to which the present invention pertains can be implemented in other specific forms without changing the technical spirit or essential features of the present invention. you will be able to understand Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive.

100: Product sales content posting server
200: user terminal
300: external server
400: Network

Claims (14)

In a method performed by a computing device,
collecting a URL of a first website on which content for selling a first product is posted;
generating crawl result data by performing a crawl on the first product sales content using the collected URL; and
Posting second product sales content on a second website using the generated crawling result data,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 1,
The step of generating the crawling result data,
When the first website is set as a crawling target by the user, automatically crawling new product sales content posted on the first website from the time the first website is set as the crawling target. including,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 1,
The step of generating the crawling result data,
Receiving crawling conditions set by the user; and
Selectively crawling only one or more first product sales contents that satisfy the set crawling condition among a plurality of first product sales contents posted on the first website,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 1,
The step of generating the crawling result data,
A product included in the first product sales content through a first image analysis model - the first image analysis model is a model pre-learned by using a plurality of product images labeled with product information as learning data - When it is determined that two or more products are included in the product image by analyzing the image, generating two or more crawling result data by crawling information on each of the two or more products,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 4,
The step of posting the second product sales content,
Two or more second product sales contents are posted on the second website using two or more crawling result data generated corresponding to each of the two or more products included in the first product sales contents, but the two or more second product sales contents are posted. When any one of the two or more second product sales contents is selected from the user by interconnecting the product sales contents, the remaining second product sales contents connected to the selected one of the second product sales contents are also displayed. Including the steps of providing together,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 1,
The step of generating the crawling result data,
When one or more foreign languages are included in the first product sales content, performing crawling by automatically translating the one or more foreign languages into a language preset by the user,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 1,
The step of generating the crawling result data,
generating first crawl result data by performing a crawl on the first product sales content; and
Generating second crawl result data by re-crawling the first product sales content at predetermined intervals from the time when the first crawl result data is generated;
The step of posting the second product sales content,
Comparing the generated first crawl result data and the generated second crawl result data, and correcting the posted second product sales content corresponding to the generated first crawl result data,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 1,
The step of posting the second product sales content,
A second image analysis model, wherein the second image analysis model is a model pre-learned by using a plurality of product images labeled with product attribute information as learning data, through which the first product sales content is included. extracting attribute information on a product included in the product image by analyzing the product image, wherein the attribute information includes at least one of a category and a style of the product; and
Posting the second product sales content in an area corresponding to attribute information on the extracted product on the second website,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 8,
The step of posting the second product sales content in an area corresponding to the attribute information on the extracted product on the second website,
Calculating accuracy of attribute information on the extracted product using attribute information on the extracted product and the one or more product images or feedback information on the product included in the one or more product images; and
When the calculated accuracy is greater than or equal to a preset value, the second product sales content is posted on the second website, and when the calculated accuracy is less than the preset value, confirmation of attribute information of the extracted product. Including the steps of requesting,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 8,
The step of posting the second product sales content in an area corresponding to the attribute information on the extracted product on the second website,
calculating accuracy of the attribute information of the extracted product by comparing N pieces of attribute information extracted by analyzing product images included in the first product sales content N times using the second image analysis model; and
When the calculated accuracy is greater than or equal to a preset value, the second product sales content is posted on the second website, and when the calculated accuracy is less than the preset value, confirmation of attribute information of the extracted product. Including the steps of requesting,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 1,
The step of posting the second product sales content,
Determining whether one or more information included in the generated crawling result data is abnormal; and
Posting the merchandise sales content when it is determined that there is nothing wrong with the one or more information, and requesting confirmation of the one or more information when it is determined that there is something wrong with the one or more information,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 1,
The step of posting the second product sales content,
Comparing one or more information included in the generated crawling result data with one or more information included in each of a plurality of pre-stored crawling result data to calculate a similarity between the generated crawling result data and a plurality of pre-stored crawling result data ; and
Posting the second product sales content using the generated crawling result data only when the calculated maximum value of the similarity is less than a preset value,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 1,
receiving an image including one or more products from a user through the second website;
One or more images included in the input image through a second image analysis model - the second image analysis model is a model pre-learned by using a plurality of product images labeled with product attribute information as learning data Extracting attribute information about a product; and
Among the plurality of second product sales contents posted on the second website, one or more second product sales contents corresponding to the attribute information on the extracted one or more products are selected, and the selected one or more second products are provided to the user. Further comprising providing sales content as a search result,
A method of posting product sales content using artificial intelligence-based website crawling. According to claim 1,
Collecting social data for a plurality of products;
extracting fashion trend information by analyzing the collected social data; and
Further comprising recommending and providing one or more second product sales contents among a plurality of second product sales contents posted on the second website to a user based on the extracted fashion trend information.
A method of posting product sales content using artificial intelligence-based website crawling.

Images