KR20230031643A - Apparatus for extracting pros and cons keyword based on product review - Google Patents

Abstract

The present invention relates to an apparatus for extracting advantage and disadvantage keywords based on a product review and, more specifically, to an apparatus capable of automatically extracting advantage keywords and disadvantage keywords by using a product review of an online shopping mall. An objective of the present invention is to allow a user to intuitively recognize product information. According to the present invention, the apparatus for extracting advantage and disadvantage keywords based on a product review comprises: an input part receiving review data for a product to be analyzed from an online sale page; a memory storing a program generating advantage keyword information and disadvantage keyword information for the product to be analyzed by using the review data; and a processor executing the program. The processor analyzes the review data to classify the review data into positive review information and negative review information, analyzes the positive review information to generate the advantage keyword information, and analyzes the negative review information to generate the disadvantage keyword information.

Description

Advantages and disadvantages keyword extraction device based on product reviews {APPARATUS FOR EXTRACTING PROS AND CONS KEYWORD BASED ON PRODUCT REVIEW}

The present invention relates to an apparatus for extracting pros and cons keywords based on product reviews, and more particularly, to an apparatus for automatically extracting keywords of merits and demerits using product reviews of an online shopping mall.

Due to the Corona 19 virus, as the size of the online commerce market has grown rapidly, the demand to sell products online is rapidly increasing.

In 2020, online product transactions increased by about 20% compared to 2019. e-Commerce transactions amounted to 161.4 trillion won, followed by Naver with 26.8 trillion won, Coupang with 21.0 trillion won, and eBay Korea with 20.0 trillion won. There are trillions of platforms, such as Alibaba, Ali-express, Taobao, 1688, and TMall.

Consumers have a need to be curious about the price of their favorite product on other platforms while shopping on the Internet, because they want to buy the product they like at the lowest price.

However, consumers have to search for similar products by visiting different shopping platforms one by one by inputting titles and model numbers, or by capturing and uploading images, making it difficult to find the same product on multiple platforms.

Even for the same product, there will always be a price difference between platforms. If you register the same product on platform A and buy and sell it on platform B, which is cheaper than this when an order comes in, you will make a profit.

However, the costs involved in distributing (shipping) products purchased on Platform B to buyers on Platform A (shopping platform fees, local shipping fees, international shipping fees, customs clearance fees, etc.) must be smaller than the price difference (margin) between platforms to be actual profit. Connected.

In the case of Naver Smart Store sellers, according to the 2021 Q4 performance announcement, sales grew 44.6% compared to the same quarter last year, and the store (seller) that sells things online is about 410,000.

However, if the seller is not a seller who has secured a delivery factory through an existing transaction relationship, even if he or she finds a product he or she wants to sell on the Internet, the delivery (manufacturing) factory of the product meets the conditions satisfactory to the seller in terms of quality and delivery price. There is a problem that is difficult to find.

In addition, in order to search for overseas products similar to the products they wish to sell found on the Internet, Korean sellers access multiple overseas wholesale platforms one by one and translate them into languages supported by the overseas wholesale platforms (eg English, Chinese, etc.). It is necessary to search by inputting keywords, but in this process, a lot of work time is consumed, and it is difficult to find a product that the seller is satisfied with.

Recently, as online shopping malls are popular, the number of products sold online is very large, and accordingly, the number of cases in which buyers or intermediate distributors analyze the corresponding products through reviews is increasing.

As the number of products and consequently the number of reviews becomes very large, a device that extracts only keywords from reviews is required. In the case of extracting advantages and disadvantages from information, there is a problem in that it is difficult to grasp information about the overall purchase.

Prior art Korean Patent Registration No. 10-2135674 discloses the content of automatically extracting keywords of the product using product reviews. Specifically, if 10 reviews on a specific product A are registered, each review Contents of extracting a total of 30 keywords, 3 of each related keyword, and extracting information such as information of a user who wrote a corresponding review and an evaluation score given by the user are disclosed. According to the prior art, keywords for each review are extracted, making it difficult to grasp overall keywords for the product, and only keywords are simply extracted without distinction between advantages and disadvantages. There is a problem that is difficult to figure out whether it is related to .

Prior art Korean Patent Registration No. 10-1903717 discloses a configuration for extracting advantages and disadvantages of a specific product from information uploaded by users to web documents such as blogs and community sites about the product, which is based on reviews Thus, the advantages and disadvantages of a specific product are provided in the form of a sentence, and the weight of the sentence is also provided together. Since this review information is extracted from web documents rather than reviews left by users on the product page, there is a problem that there is a high possibility that it contains only information about the product itself and no information about the overall purchase such as seller and delivery, Since advantages and disadvantages are provided in the form of sentences, users who need to grasp a lot of information in a short time may feel that unnecessary information is provided.

The present invention has been proposed to solve the above problems, by using review information obtained from a sales page of a specific product to extract merit keywords and demerit keywords for the product, and to provide analysis results for overall product purchase. The purpose is to provide a product review-based pros and cons keyword extraction device that can do this.

The present invention relates to an apparatus for extracting pros and cons keywords based on product reviews, and more particularly, to an apparatus for automatically extracting keywords of merits and demerits using product reviews of an online shopping mall.

An apparatus for extracting pros and cons keywords based on product review according to the present invention includes an input unit for receiving review data for a product to be analyzed on an online sales page, and a program for generating merit keyword information and demerit keyword information for the product to be analyzed using the review data. The processor includes a processor that executes the stored memory and program, the processor analyzes the review data, classifies it into positive review information and negative review information, analyzes the positive review information to generate the merit keyword information, and analyzes the negative review information. to generate the disadvantage keyword information.

The processor analyzes the review data and classifies it into positive review information and negative review information based on predetermined review score criteria.

The processor extracts keywords by analyzing appearance frequencies and relationships between words in review data composed of sentences, and generates strength keyword information and weakness keyword information.

The processor decomposes the extracted keyword into morpheme units to generate a basic keyword, and generates strength keyword information and weakness keyword information based thereon.

The processor generates and provides the strength keyword information and the weakness keyword information in words, and provides them in a predetermined number of words.

According to the present invention, it is possible to search for the lowest price product on multiple platforms with a single click during internet shopping or surfing.

According to the present invention, by registering a product on the platform without inventory and purchasing and delivering it from the lowest price platform when an order occurs, that is, through an inventory-free automation system using artificial intelligence, without risk of capital loss, through cross-platform product arbitrage trading. There is an effect that can realize a profit.

According to the present invention, it is possible to significantly reduce labor costs by automating simple repetitive tasks by manpower required in existing purchasing agencies, and to sell products to consumers at reasonable prices through competition at the lowest price compared to existing products.

According to the present invention, it is possible to perform a search using an input image (an image or image link input by a user to search for a desired product) for product information listed on an overseas wholesale platform. When a user inputs an image or image link, it is possible to simultaneously call (API, crawl) input data to a plurality of overseas wholesale platforms and provide search results for product information of each platform.

By converting and providing product information such as product name and price to correspond to the user's country setting, there is an effect of supporting the user to intuitively recognize product information.

Filtering is performed on search results using factory information and product information, thereby increasing user satisfaction with search results.

Since a specified number of keywords are extracted for a specific product, it is easy to grasp the general keywords for the product, and there is an effect of enabling intuitive identification of advantages and disadvantages of the product based on the keywords.

By extracting and analyzing keywords for pros and cons only for reviews registered on the sales page of a specific product, it is possible to collectively grasp reviews of not only the product itself but also the seller and delivery.

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

1 shows a product search device using image information according to an embodiment of the present invention.
2 illustrates a product search method using image information according to an embodiment of the present invention.
3 illustrates a process of calling a foreign wholesale platform according to an embodiment of the present invention.
4 illustrates a process of converting searched information and providing search results according to an embodiment of the present invention.
5 illustrates a filtering and sorting process for search results according to an embodiment of the present invention.
6 illustrates product search results using image information according to an embodiment of the present invention.
7 illustrates a product search apparatus using keyword information according to another embodiment of the present invention.
8 illustrates a product search method using keyword information according to another embodiment of the present invention.
9 illustrates a process of converting searched information and providing search results according to another embodiment of the present invention.
10 illustrates a filtering and sorting process for search results according to another embodiment of the present invention.
11 illustrates an apparatus for analyzing supply costs of overseas wholesale products according to another embodiment of the present invention.
12 illustrates a supply cost analysis method for overseas wholesale products according to another embodiment of the present invention.
13 shows a final supply cost analysis providing screen for each option for a product selected from an overseas wholesale platform according to another embodiment of the present invention.
14 illustrates an open market seller service providing apparatus according to another embodiment of the present invention.
15 illustrates an open market seller service providing method according to another embodiment of the present invention.
16 illustrates an apparatus for predicting time-series demand for open market products according to another embodiment of the present invention.
17 illustrates a method for predicting time-series demand for open market products according to another embodiment of the present invention.
18 illustrates an overseas exhibition product information providing device according to another embodiment of the present invention.
19 illustrates a method of providing product information at an overseas exhibition according to another embodiment of the present invention.
20 illustrates a process of providing product information at an overseas exhibition according to another embodiment of the present invention.
21 illustrates an apparatus for constructing an open market database and providing services according to another embodiment of the present invention.
22 illustrates a method for constructing an open market database and providing a service according to another embodiment of the present invention.
23 illustrates the relationship between products, keywords, and stores according to another embodiment of the present invention.
24 to 26 illustrate complementary entity creation and update processes between product, store, and keyword DBs.
27 shows a product analysis information providing screen according to another embodiment of the present invention.
28 illustrates a search word analysis information providing screen according to another embodiment of the present invention.
29 and 30 illustrate an apparatus for extracting pros and cons keywords based on product reviews according to another embodiment of the present invention.
31 illustrates keyword extraction according to another embodiment of the present invention.
32 illustrates a method for extracting pros and cons keywords based on product reviews according to another embodiment of the present invention.
33a and 33b illustrate keywords extracted by the apparatus for extracting pros and cons keywords based on product reviews according to another embodiment of the present invention.
34 shows an artificial intelligence-based product transaction automation system according to another embodiment of the present invention.
35 illustrates an artificial intelligence-based product transaction automation method according to another embodiment of the present invention.
36 illustrates a multi-platform cheapest product search assistance system according to another embodiment of the present invention.

The foregoing and other objects, advantages and characteristics of the present invention, and a method of achieving them will become clear with reference to the detailed embodiments described below in conjunction with the accompanying drawings.

However, the present invention is not limited to the embodiments disclosed below and may be implemented in various different forms, and only the following embodiments provide the purpose of the invention, As only provided to easily inform the configuration and effect, the scope of the present invention is defined by the description of the claims.

Meanwhile, terms used in this specification are for describing the embodiments and are 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” means the presence of one or more other components, steps, operations, and/or elements in which a stated component, step, operation, and/or element is present. or added.

1. Product search using image information

Due to the Corona 19 virus, as the size of the online commerce market has grown rapidly, the demand to sell products online is rapidly increasing.

In the case of Naver Smart Store sellers, according to the 2021 Q4 performance announcement, sales grew 44.6% compared to the same quarter last year, and the store (seller) that sells things online is about 410,000.

However, if the seller is not a seller who has secured a delivery factory through an existing transaction relationship, even if he or she finds a product he or she wants to sell on the Internet, the delivery (manufacturing) factory of the product meets the conditions satisfactory to the seller in terms of quality and delivery price. There is a problem that is difficult to find.

In addition, in order to search for overseas products similar to the products they wish to sell found on the Internet, Korean sellers access multiple overseas wholesale platforms one by one and translate them into languages supported by the overseas wholesale platforms (eg English, Chinese, etc.). It is necessary to search by inputting keywords, but in this process, a lot of work time is consumed, and it is difficult to find a product that the seller is satisfied with.

1 shows a product search device using image information according to an embodiment of the present invention.

An apparatus for searching products using image information according to an embodiment of the present invention includes an input unit 110 that receives input image information, a memory 120 storing a program for searching similar products in overseas wholesale platforms using the input image information, and It includes a processor 130 that executes a program, and the processor 130 searches similar products determined to be similar to the input image using input image information including at least one of an image link and an image file, and sets the search Provide similar product search results considering the information.

The input unit 110 receives input image information including at least one of an image link and an image file.

The processor 130 simultaneously calls a plurality of overseas wholesale platforms using at least one of input image information and keyword information obtained by converting the input image information, and provides similar product search results.

The processor 130 considers the setting information corresponding to whether the category is restricted or not, and provides similar recommended products by including them in the search results.

The processor 130 recognizes the category of an image when searching for an image (when calling an API), and provides a product search result by limiting the category to the recognized category in the overseas wholesale platform.

At this time, factors for limiting search results include categories, input product images, and tags associated with keywords.

The processor 130 provides similar product search results by translating the main information (eg, product name) of similar products into the language of a preset country, and considers the exchange rate for the price information of similar products, and the currency of the country set by the user. Provide similar product search results including information converted into units.

The processor 130 performs at least one of filtering and sorting of similar product search results using search setting information including at least one of factory information and product information.

2 shows a product search method using image information according to an embodiment of the present invention, FIG. 3 shows a procedure for calling an overseas wholesale platform according to an embodiment of the present invention, and FIG. 4 shows a process according to an embodiment of the present invention. A process of converting searched information and providing search results is shown, and FIG. 5 illustrates a process of filtering and sorting search results according to an embodiment of the present invention.

A product search method using image information according to an embodiment of the present invention includes receiving image information (S210), making a call from an overseas wholesale platform using the image information (S220), and searching for information according to the call. Converting and providing search results (S230) and performing filtering and sorting on the search results (S240).

In step S210, an image file or image link directly uploaded from the user device is received.

As another example, a service provided image (an open market product image, an overseas insight image, an overseas online exhibition image, etc.) is received or an image in a browser selected through a chrome extension is received.

In step S220, the image information received in step S210 is stored in the input data server, and similar products are searched for in overseas wholesale platforms.

In this case, the overseas wholesale platform to be searched may be directly selected by the user or selected by the product search device.

Step S221 checks whether an image search function exists in the overseas wholesale platform.

If it is confirmed that the image search function exists in step S221, the first overseas wholesale platform having the image search function (as a plurality of overseas wholesale platforms, the first overseas wholesale platform supporting the image search function to help those skilled in the art understand) As an overseas wholesale platform, an overseas wholesale platform that does not support an image search function is defined as a second overseas wholesale platform).

If it is confirmed that the image search function does not exist in step S221, the image information is converted into keywords, the keywords are translated into a language supported by the corresponding overseas wholesale platform, and the converted and translated keywords are used in the second overseas wholesale platform. make a call

According to an embodiment of the present invention, in the process of searching for similar products on overseas wholesale platforms using image information (image files or image links) or converting image information into keywords to search for similar products on overseas wholesale platforms, According to the user's settings, it is possible to provide not only similar products but also related similar recommended products as search results.

For example, it is assumed that a user wants to secure basketball shoes of the Air Jordan 11 model, which are difficult to obtain in Korea, through an overseas wholesale platform and sell them.

As the user inputs image information about the Air Jordan 11 model, a product is searched on an overseas wholesale platform.

At this time, the user may wish to search only for basketball shoes, or may wish to search not only for basketball shoes but also for other products to which the design of the Air Jordan 11 model is applied (similar recommended products, for example, key rings, cell phone cases, etc.).

If the user wants to search only for basketball shoes, if other products (key rings, cell phone cases) are included in the search results, user satisfaction will decrease. It will be a means to increase user satisfaction by including and providing products of unexpected categories as various search results.

For similar recommended products, the order in which they are displayed in the search results is determined in consideration of the history of other users actually clicking on the search results and confirming the product information.

For example, the user requests to provide similar recommended products as a search result, and confirms an Air Jordan interior wall clock as a search result. As for the Air Jordan model, Air Jordan 1 to Air Jordan 35 were released, and the product is a product in which the characters of the Air Jordan basketball shoe model corresponding to the numbers 1 to 12 are placed at each number position (1 to 12) on the wall clock.

Through the search result, the user can check similar recommended products as well as products similar to the input image, check product information through information conversion described later for the product, and search results based on factory condition information and product condition information. it is possible to check

In addition, when the user selects a function to inquire a supplier of the searched product (Air Jordan interior wall clock), a product manufacturing request request including image information may be transmitted. For example, with respect to the Air Jordan interior wall clock, it is possible to request a quote by sending a production request for a clock expressed as an Air Jordan model version character whose month/day information corresponds to each number.

At this time, the user requests a quotation through the website of the product search service provider using image information, with the contact information of the person in charge of the supplier of the selected similar product or similar recommended product (eg, the corresponding information is e-mail information obtained through crawling) It is possible to transmit (including desired product specifications and supply request quantity information).

For a result of making a call in the first and second overseas wholesale platforms, search information is collected.

In the aforementioned calling process, if an API exists, an API call is performed, and if an API does not exist, a browser result provided by direct input from the corresponding platform is parsed.

Step S230 converts information retrieved as a result of calling the overseas wholesale platform and provides search results.

In step S231, the collected product information is translated according to the user's country setting information. For example, the product information, which is the result of a search on the Chinese wholesale platform, is translated into Korean according to the user's country setting information (Korea), and in step S232 provides search results for product information for which translation has been performed.

In step S231, it is checked whether price information is included in the searched information, and if the price information is included, the price information is converted by applying a real-time exchange rate according to the user's country setting information.

In step S231, numbers to be translated are distinguished from numbers that do not require translation, and conversion is not performed on numbers that do not require translation.

In step S240, at least one of filtering and sorting is performed on the product list provided in step S230 according to a factory whose quality has been verified or a specific condition (eg, price) according to a user's request condition.

Step S241 receives at least one of factory condition information and product condition selection information, and step S242 provides filtered or sorted search results.

Factory condition information includes in-platform supply level information, supply volume information, product description matching information, response speed information, delivery speed information, refund rate information, and dispute rate information.

Using the company profile in the overseas wholesaler platform, it is possible to use factory condition information by performing filtering based on preset values for product description matching information, delivery speed information, refund rate information, and dispute rate information.

For example, through automatic filtering (eg, selecting only good factories), it is possible to provide related information by filtering for factories that satisfy a condition for a certain reference value compared to the average.

Product condition information includes specific keyword information, category information, price information, shipping cost information, and MOQ (minimum quantity) information.

6 illustrates product search results using image information according to an embodiment of the present invention.

Referring to FIG. 6, when a user directly inputs an image file or image link, or inputs an overseas fair or open market product image as a search image, the language set by the user for similar products searched for on overseas wholesale platforms (Korean ), the product name and price information are included in the search results and provided.

According to the prior art, there is a problem in that it is difficult to check various information on products of an overseas wholesale platform through a search engine algorithm, and image search provided by an overseas wholesale platform has a limitation in that only products of the corresponding platform can be searched.

According to an embodiment of the present invention, when a user inputs an image file or image link information, the input data is simultaneously called (API, crawled) to a plurality of overseas wholesale platforms, and search results including product information of each platform are provided. There is an effect that can be done.

In the case of calling, as described above, if an image search API exists, the API is used, and if the API is not provided, an image search is performed directly on the corresponding platform, and the search result is provided by parsing the browser result.

If the image search function itself is not supported by the overseas wholesale platform, the image analysis engine is used to convert the input image into keywords to perform a search, and the search results from the overseas wholesale platform are provided through API or crawling.

If the overseas wholesale platform search results are vast, it takes a lot of time for users to find the product they want, and even when they find the product they want, they have anxiety about the product quality. According to an embodiment of the present invention, it is possible to provide reliable search results by filtering or sorting searched product information under specific conditions.

According to an embodiment of the present invention, search results are filtered with factory information and product information.

At this time, it is possible to check whether the factory has the ability to manufacture and supply high-quality products by using the factory information. Information on the dispute rate and dispute rate is included, and it is possible to perform filtering or sorting based on each item.

Product information includes specific keywords, categories, prices, shipping costs, and MOQ (minimum quantity) information, and similarly, it is possible to perform filtering or sorting based on each item.

According to the present invention, it is possible to perform a search using an input image (an image or image link input by a user to search for a desired product) for product information listed on an overseas wholesale platform. When a user inputs an image or image link, it is possible to simultaneously call (API, crawl) input data to a plurality of overseas wholesale platforms and provide search results for product information of each platform.

By converting and providing product information such as product name and price to correspond to the user's country setting, there is an effect of supporting the user to intuitively recognize product information.

Filtering is performed on search results using factory information and product information, thereby increasing user satisfaction with search results.

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

Meanwhile, the product search method using image information according to an embodiment of the present invention may be implemented in a computer system or recorded on a recording medium. A computer system may include at least one processor, a memory, a user input device, a data communication bus, a user output device, and a storage. Each of the aforementioned components communicates data through a data communication bus.

The computer system may further include a network interface coupled to the network. The processor may be a central processing unit (CPU) or a semiconductor device that processes instructions stored in memory and/or storage.

The memory and storage may include various types of volatile or non-volatile storage media. For example, memory may include ROM and RAM.

Accordingly, the product search method using image information according to an embodiment of the present invention may be implemented as a method executable on a computer. When the product search method using image information according to an embodiment of the present invention is performed in a computer device, computer readable instructions may perform the product search method using image information according to the present invention.

On the other hand, the product search method using image information according to the present invention described above can be implemented as computer readable code on a computer readable recording medium. Computer-readable recording media includes all types of recording media in which data that can be decoded by a computer system is stored. For example, there may be read only memory (ROM), random access memory (RAM), a magnetic tape, a magnetic disk, a flash memory, an optical data storage device, and the like. In addition, the computer-readable recording medium may be distributed in computer systems connected through a computer communication network, and stored and executed as readable codes in a distributed manner.

2. Product search using keyword information

Due to the Corona 19 virus, as the size of the online commerce market has grown rapidly, the demand to sell products online is rapidly increasing.

In the case of Naver Smart Store sellers, according to the 2021 Q4 performance announcement, sales grew 44.6% compared to the same quarter last year, and the store (seller) that sells things online is about 410,000.

However, if the seller is not a seller who has secured a delivery factory through an existing transaction relationship, even if he or she finds a product he or she wants to sell on the Internet, the delivery (manufacturing) factory of the product meets the conditions satisfactory to the seller in terms of quality and delivery price. There is a problem that is difficult to find.

In addition, in order to search for overseas products similar to the products they wish to sell found on the Internet, Korean sellers access multiple overseas wholesale platforms one by one and translate them into languages supported by the overseas wholesale platforms (eg English, Chinese, etc.). It is necessary to search by inputting keywords, but in this process, a lot of work time is consumed, and it is difficult to find a product that the seller is satisfied with.

7 illustrates a product search apparatus using keyword information according to another embodiment of the present invention.

An apparatus for searching products using keyword information according to an embodiment of the present invention includes an input unit 710 receiving input keyword information, a memory 720 storing a program for searching products in an overseas wholesale platform using the input keyword information, and a program and a processor 730 that executes, and the processor 730 translates the input keyword into a language supported by each overseas wholesale platform, calls the overseas wholesale platform, and provides product search results.

The processor 730 simultaneously calls a plurality of overseas wholesale platforms and provides product search results.

The processor 730 considers setting information corresponding to whether the category is restricted or not, and provides similar recommended products by including them in the search results.

The processor 730 recognizes the category of the image when searching for an image (when calling an API), and provides a product search result by limiting the category to the category recognized in the overseas wholesale platform.

At this time, factors for limiting search results include categories, input product images, and tags associated with keywords.

The processor 730 provides a product search result by translating the main information of the product according to the search result into the language of a preset country, and converts the price information of the product into the currency unit of the country set by the user in consideration of the exchange rate. Provide product search results including information.

The processor 730 performs at least one of filtering and sorting on product search results using search setting information including at least one of factory information and product information.

8 shows a product search method using keyword information according to another embodiment of the present invention, FIG. 9 shows a process of converting searched information and providing search results according to another embodiment of the present invention, and FIG. A process of filtering and sorting search results according to another embodiment of the present invention is shown.

A product search method using keyword information according to an embodiment of the present invention includes receiving keyword information (S810), making a call in an overseas wholesale platform using the keyword information (S820), Converting and providing search results (S830) and performing filtering and sorting on the search results (S840).

In step S810, keyword information is received from the user device.

At this time, it is possible for the user to directly input keywords or service provision keywords, and domestic hot keywords, related keywords, autocomplete keywords, etc. are provided as service provision keywords.

In step S820, the keyword information received in step S810 is stored in the input data server, and similar products are searched for in overseas wholesale platforms using the keyword information.

Step S820 translates the entered keyword into a language supported by each overseas wholesale platform, and calls the overseas wholesale platform using the translated keyword.

When an API exists, an API call is performed, and when there is no API, a browser result provided by direct input to the corresponding platform is parsed.

In this case, the overseas wholesale platform to be searched may be directly selected by the user or selected by the product search device.

According to an embodiment of the present invention, in the process of searching for similar products in overseas wholesale platforms using keyword information as described above, it is important to provide not only similar products but also related similar recommended products as search results according to user settings. possible.

For example, it is assumed that a user wants to secure basketball shoes of the Air Jordan 11 model, which are difficult to obtain in Korea, through an overseas wholesale platform and sell them.

As the user inputs keyword information as "Air Jordan 11", the keyword is translated into a language (English, Chinese, etc.) supported by the overseas wholesale platform, and the product is searched through the overseas wholesale platform.

At this time, the user may wish to search only for basketball shoes, or may wish to search not only for basketball shoes but also for other products to which the design of the Air Jordan 11 model is applied (similar recommended products, for example, key rings, cell phone cases, etc.).

If the user wants to search only for basketball shoes, if other products (key rings, cell phone cases) are included in the search results, user satisfaction will decrease. It will be a means to increase user satisfaction by including and providing products of unexpected categories as various search results.

At this time, depending on whether the user inputs the keyword "Air Jordan 11" or "Air Jordan 11 basketball shoes" as the keyword, it is determined whether similar recommended products are additionally included in the search results, and the search results are determined accordingly. to provide.

In addition, even when the user inputs “Air Jordan 11 basketball shoes” as a keyword, if a search result for similar recommended products is requested according to a separate user setting, search results including similar recommended products are provided.

For similar recommended products, the order in which they are displayed in the search results is determined in consideration of the history of other users actually clicking on the search results and confirming the product information.

For example, the user requests to provide similar recommended products as a search result, and confirms an Air Jordan interior wall clock as a search result. As for the Air Jordan model, Air Jordan 1 to Air Jordan 35 were released, and the product is a product in which the characters of the Air Jordan basketball shoe model version corresponding to numbers 1 to 12 are placed at each number position (1 to 12) on the wall clock.

The user can check similar recommended products for the keyword "Air Jordan 11" entered through the search results, check product information through information conversion described later for the product, and based on factory condition information and product condition information It is possible to check the search results.

In addition, when the user selects a function to inquire a supplier of the searched product (Air Jordan interior wall clock), a product manufacturing request request including detailed request information may be transmitted. For example, with respect to the Air Jordan interior wall clock, it is possible to request a quote by sending a production request for a clock expressed as an Air Jordan model version character whose month/day information corresponds to each number.

At this time, if the user inputs the production request and quotation request in their native language, they are translated into the language of the producer listed on the corresponding overseas platform, and finally the production request and quotation request information is generated and transmitted.

At this time, the user requests a quotation through the website of the product search service provider using keyword information with the contact information of the person in charge of the supplier of the selected similar product or similar recommended product (eg, the corresponding information is e-mail information obtained through crawling) It is possible to transmit requested information translated into the language of the supplier while transmitting the desired product specification and supply request quantity information.

In step S830, search information is collected for the result of the call from the overseas wholesale platform, conversion is performed on the information retrieved as a result of the call to the overseas wholesale platform, and the search result is provided.

In step S831, the collected product information is translated according to the user's country setting information. For example, the product information, which is the result of a search on the Chinese wholesale platform, is translated into Korean according to the user's country setting information (Korea), and in step S232 provides search results for product information for which translation has been performed.

In step S831, it is checked whether price information is included in the retrieved information, and if price information is included, price information is converted by applying a real-time exchange rate according to the user's country setting information.

Step S831 distinguishes numbers to be translated from numbers that do not require translation, and does not perform conversion on numbers that do not require translation.

In step S840, at least one of filtering and sorting is performed on the product list provided in step S830 according to a factory whose quality has been verified or a specific condition (eg, price) according to a user's request condition.

Step S841 receives at least one of factory condition information and product condition selection information, and step S842 provides filtered or sorted search results.

Factory condition information includes in-platform supply level information, supply volume information, product description matching information, response speed information, delivery speed information, refund rate information, and dispute rate information.

Using the company profile in the overseas wholesaler platform, it is possible to use factory condition information by performing filtering based on preset values for product description matching information, delivery speed information, refund rate information, and dispute rate information.

For example, through automatic filtering (eg, selecting only good factories), it is possible to provide related information by filtering for factories that satisfy a condition for a certain reference value compared to the average.

Product condition information includes specific keyword information, category information, price information, shipping cost information, and MOQ (minimum quantity) information.

According to the present invention, by translating a keyword input by a user into a supported language of an overseas wholesale platform, simultaneously calling a plurality of overseas wholesale platforms, and collecting and providing search results, the user can individually obtain product information from each overseas wholesale platform. It has the effect of being able to search for products without having to search and organize fragmented information.

By converting and providing product information such as product name and price to correspond to the user's country setting, there is an effect of supporting the user to intuitively recognize product information.

Filtering is performed on search results using factory information and product information, thereby increasing user satisfaction with search results.

Meanwhile, a product search method using keyword information according to an embodiment of the present invention may be implemented in a computer system or recorded on a recording medium. A computer system may include at least one processor, a memory, a user input device, a data communication bus, a user output device, and a storage. Each of the aforementioned components communicates data through a data communication bus.

The computer system may further include a network interface coupled to the network. The processor may be a central processing unit (CPU) or a semiconductor device that processes instructions stored in memory and/or storage.

The memory and storage may include various types of volatile or non-volatile storage media. For example, memory may include ROM and RAM.

Therefore, the product search method using keyword information according to an embodiment of the present invention may be implemented as a method executable on a computer. When the product search method using keyword information according to an embodiment of the present invention is executed in a computer device, computer readable instructions may perform the product search method using keyword information according to the present invention.

Meanwhile, the product search method using keyword information according to the present invention described above can be implemented as computer readable codes on a computer readable recording medium. Computer-readable recording media includes all types of recording media in which data that can be decoded by a computer system is stored. For example, there may be read only memory (ROM), random access memory (RAM), a magnetic tape, a magnetic disk, a flash memory, an optical data storage device, and the like. In addition, the computer-readable recording medium may be distributed in computer systems connected through a computer communication network, and stored and executed as readable codes in a distributed manner.

3. Supply cost analysis of overseas wholesale products

Due to Corona 19, the online commerce market is growing more rapidly, and the number of sellers who want to sell products online is also rapidly increasing.

In the case of Naver Smart Store sellers, according to the 2021 Q4 performance announcement, sales grew 44.6% compared to the same quarter last year, and the store (seller) that sells things online is about 410,000.

In order to receive good products at low prices, sellers choose overseas item sourcing as their main product supply method.

However, even when a seller finds a desired product on an overseas wholesale platform, it is difficult to calculate the supply cost for actually importing the product.

In other words, in order to calculate the supply cost, it is necessary to consider tariffs, customs duties, local shipping charges, customs fees, domestic shipping charges, and costs required for various certifications according to FTA status. As a result, you can only estimate 1.3 to 1.5 times the wholesale price as the cost of supply (actual import price).

When using an agency service (purchase agency, etc.), the seller must pay a high commission and use the agency service, even though the process and each item (actual purchase amount, local delivery fee, exchange rate, etc.) are not transparently disclosed. There is a problem with

11 illustrates an apparatus for analyzing supply costs of overseas wholesale products according to another embodiment of the present invention.

The service according to the present invention comprises an input unit 1110 for receiving a request for supply costs for overseas wholesale products, a memory 1120 storing a program for analyzing supply costs for overseas wholesale products, and a processor 1130 for executing the program. Including, the processor 1130 analyzes the supply cost including all costs for overseas wholesale products according to the laws and regulations of the country where the user is located.

The input unit 1110 receives a request for a supply cost for a specific overseas wholesale product from a list of product search results from a plurality of overseas wholesale platforms using at least one of image information and keyword information.

The processor 1130 parses information on overseas wholesale products that are subject to supply cost analysis.

The processor 1130 analyzes the supply cost by considering customs duties according to the HS code, customs duties, local shipping fees, customs clearance fees, domestic shipping costs, and certification fees.

Processor 1130 converts the supply cost into the currency of the country in which the user is located to calculate the total expected cost.

The processor 1130 presents recommended products by parsing information in the search list as well as products for which supply cost analysis is requested in the search list according to membership settings.

For example, the processor 1130 receives product information from a plurality of overseas wholesale platforms from an input through image information or keyword information, and provides a list including product photos, product names, and price information per piece.

At this time, it is assumed that when the user requests a supply cost analysis for product A, a supply cost analysis for a quantity of 100 items is requested.

In receiving product information from a plurality of overseas wholesale platforms, the processor 130 builds a DB by accumulating supply cost information per product according to the supply quantity, and analyzes supply cost by the user changing the quantity of product A to 200. When requesting, supply cost analysis of product A, as well as supply cost information for product B, in which the cost of supply is lowered by more than a predetermined value (eg 10%) when the quantity is changed to 200, is provided to the user. , to support user decision-making.

At this time, the supply cost analysis and recommendation for product B related to product A is performed based on the image analysis displayed in the list or the product relevance value in the built DB, and whether or not to recommend is determined according to the external similarity above the preset value. It is decided.

The processor 1130 provides notification of supply cost analysis and recommendation information at a preset cycle through SNS of a customer account.

At this time, the processor 1130 may transmit supply cost analysis information or product recommendation information in the field of interest of the user or a field that meets a specific condition by using the user's setting information, the user's search pattern information, or the user's purchase pattern information. do.

12 illustrates a supply cost analysis method for overseas wholesale products according to another embodiment of the present invention.

The supply cost analysis method of an overseas wholesale product according to an embodiment of the present invention includes receiving information on a product subject to supply cost analysis from an online overseas wholesale platform (S1210), analyzing the supply cost (S1230), and analyzing the analysis result. and providing (S1240).

In step S1210, a list of products inquired in a preset method using a plurality of online overseas wholesale platforms is displayed, and a product subject to supply cost analysis is determined according to the user's selection.

At this time, the product list is a list of products searched by inputting image information (image file, image link, etc.) or a list of products searched by inputting keyword information in the UI provided through the service providing server.

As another example, it is possible to determine a product subject to supply cost analysis by inputting product information or link information registered in an overseas wholesale platform.

When an overseas wholesale platform product is input, a step (S1220) of parsing information on a product subject to supply cost analysis is performed.

Step S1230 analyzes the supply cost by considering the product price, tariff according to the HS code, customs duty, local shipping fee, customs clearance fee, domestic shipping fee, and certification fee (e.g. radio wave certification, etc.) Calculate.

At this time, it is possible to perform supply cost analysis using category information other than the HS code.

For example, in the dining table category, a tariff of 8 per cent applies.

In addition, it is possible to analyze the seller's search history and final product sourcing history information to provide recommended product information that satisfies predetermined standards for customs duties, local delivery fees, customs clearance fees, domestic delivery costs, and certification fees.

Step S1240 provides the user with an analysis result of the supply cost, and when the user requests quantity adjustment, the supply cost is recalculated according to the adjusted quantity, and the result is provided.

13 shows a final supply cost analysis providing screen for each option for a product selected from an overseas wholesale platform according to another embodiment of the present invention.

When a product search is performed on a plurality of overseas wholesale platforms on the service provider's webpage, and supply cost analysis for a specific product is requested from the searched product list, the final supply cost for each option of the product is analyzed, and the analysis result is displayed. to provide.

At this time, analysis is performed including customs duties, customs duties, customs clearance fees, and domestic shipping fees.

Depending on the options for each product, the CIF price, import duty, import VAT, shipping cost (including warehouse cost), customs clearance cost (including warehouse cost), supply cost per product, and total estimated cost based on the order quantity are calculated.

According to the present invention, when a user requests a supply cost inquiry for a specific product on an online overseas wholesale platform, the final supply cost according to the commercial law of the user's country and details of the supply cost are converted into the monetary unit of the user's country. By providing this information, it has the effect of helping to clearly recognize the details of supply cost calculation.

Users do not have to calculate the product's exchange rate applied price, customs duties, customs duties, overseas shipping costs, customs clearance fees, domestic shipping costs, and various certification fees, but supply through online overseas wholesale platforms by requesting a supply cost inquiry for the product of interest. It is possible to check the cost intuitively.

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

Meanwhile, the supply cost analysis method of overseas wholesale products according to an embodiment of the present invention may be implemented in a computer system or recorded on a recording medium. A computer system may include at least one processor, a memory, a user input device, a data communication bus, a user output device, and a storage. Each of the aforementioned components communicates data through a data communication bus.

The computer system may further include a network interface coupled to the network. The processor may be a central processing unit (CPU) or a semiconductor device that processes instructions stored in memory and/or storage.

The memory and storage may include various types of volatile or non-volatile storage media. For example, memory may include ROM and RAM.

Therefore, the supply cost analysis method of overseas wholesale products according to an embodiment of the present invention can be implemented in a computer-executable method. When the supply cost analysis method of overseas wholesale products according to an embodiment of the present invention is performed in a computer device, computer readable instructions may perform the supply cost analysis method of overseas wholesale products according to the present invention.

On the other hand, the above-described method for analyzing supply costs of overseas wholesale products according to the present invention can be implemented as computer readable codes on a computer readable recording medium. Computer-readable recording media includes all types of recording media in which data that can be decoded by a computer system is stored. For example, there may be read only memory (ROM), random access memory (RAM), a magnetic tape, a magnetic disk, a flash memory, an optical data storage device, and the like. In addition, the computer-readable recording medium may be distributed in computer systems connected through a computer communication network, and stored and executed as readable codes in a distributed manner.

4. Provide open market seller service

As the online commerce market has rapidly increased due to Corona 19, the demand to sell products online is also rapidly increasing.

In the case of Naver Smart Store sellers, according to the 2021 Q4 performance announcement, sales grew 44.6% compared to the same quarter last year, and there are about 410,000 stores (sellers) that sell things online.

In order to sell a product through an open market, there are an item selection stage to decide which products to sell, an item sourcing stage to purchase inexpensive and high-quality products related to the selected item, and an item sales stage to register and sell the sourcing items on the open market. need.

Sellers rely on intuition to select items that are expected to sell well, find and source items individually from domestic or overseas factories, and register the sourced items on an open market platform one by one to sell products.

According to the prior art, a tool for automatically registering sourcing items on an open market platform has recently been proposed, and various marketing tools have been proposed, but the entire process for selling in the open market is fragmented, so sellers can There are problems in that a lot of time is consumed and omissions and human errors occur as different services are used for each process.

14 illustrates an open market seller service providing apparatus according to another embodiment of the present invention.

An open market seller service providing apparatus according to an embodiment of the present invention includes an input unit 1410 for receiving item-related information, a memory 1420 storing a program for selecting, purchasing, and selling an item, and a program for executing the program. It includes a processor 1430, and the processor 1430 provides item selection, item purchase, and item sales services through a single platform.

The processor 1430 provides at least one of product comparison and analysis information using keywords, product discovery information, overseas insight search information, overseas exhibition site search information, and SNS data analysis information as reference data for item selection.

The processor 1430 provides information on the search volume of the open market platform and the number of pre-registered products as reference data for item selection.

The processor 1430 provides a supply cost analysis result as reference data for item selection.

The processor 1430 performs automatic order processing in response to a purchase request for an overseas wholesale platform item.

The processor 1430 automatically registers the purchase item in the open market platform.

15 illustrates an open market seller service providing method according to another embodiment of the present invention.

The open market seller service providing method according to an embodiment of the present invention includes the steps of analyzing a big data market (S1510), selecting an item (S1520), purchasing an item (S1530), and an open market platform. It includes registering a product in (S1540) and performing marketing (S1550).

Step S1510 performs product comparative analysis using keywords, product discovery, overseas insight search, overseas exhibition site search, and SNS data analysis.

In step S1520, an item is selected based on a search result from a plurality of domestic and overseas wholesale platforms based on the analysis result in step S1510.

At this time, step S1520 provides information on demand (search amount) and supply (number of registered products) of the open market platform as the analysis result of step S1510.

In step S1530, the cost of supply of the product selected by the seller is analyzed and provided, and a purchasing process for the item is performed.

At this time, the supply cost is analyzed considering the tariff, customs tax, local delivery fee, customs clearance fee, domestic delivery fee, and certification fee according to the HS code, and the supply cost analysis result according to the purchase quantity and option is provided.

At this time, it is possible to perform supply cost analysis using category information other than the HS code.

For example, in the dining table category, a tariff of 8 per cent applies.

In addition, it is possible to analyze the seller's search history and final product sourcing history information to provide recommended product information that satisfies predetermined standards for customs duties, local delivery fees, customs clearance fees, domestic delivery costs, and certification fees.

Step S1540 automatically registers the product selected by the seller on the open market platform.

Step S1550 performs marketing for products registered on the open market platform, and performs SNS marketing, open market advertisement marketing, search advertisement marketing, experience group advertisement marketing, and blog marketing.

According to an embodiment of the present invention, it is possible to sequentially perform the above steps S1510 to S1550, and as another example, not only can each step be individually performed, but also an item purchase step (S1530) or registration for an open market platform. It is also possible to perform from step S1540.

According to the present invention, the seller can manage item selection, purchase, and sale on a single platform.

In the process of selecting an item, it is possible to check the demand (search volume) and supply (number of registered products) of the open market platform, and to check various insights as data, rather than simply selecting an item based on intuition.

According to the present invention, it is possible to directly purchase products listed on the overseas wholesale sourcing platform, which has a difficult purchasing procedure (in the case of 1688, which is currently a Chinese domestic wholesale platform, purchases cannot be made without a Chinese bankbook and card), and the seller can purchase items If is selected, the cost of supply is analyzed and provided in currency units of the country where the seller is located, and there is an effect of enabling automatic ordering through the system.

When product order processing is completed, local delivery, inspection, and customs clearance procedures are automatically carried out, and products are delivered to the address requested by the seller and automatically registered on the open market platform. It is possible to provide inventory and product registration services together.

A data-based marketing service is provided to increase the sales volume of products registered for sale, and it is possible to induce consumer inflow through SNS advertisements, search advertisements, and the like.

That is, after the seller carefully selects and places an order, the seller's convenience is increased by providing a fulfillment service through the platform for the remaining procedures so that only CS and issue response are performed.

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

Meanwhile, the open market seller service providing method according to an embodiment of the present invention may be implemented in a computer system or recorded on a recording medium. A computer system may include at least one processor, a memory, a user input device, a data communication bus, a user output device, and a storage. Each of the aforementioned components communicates data through a data communication bus.

The computer system may further include a network interface coupled to the network. The processor may be a central processing unit (CPU) or a semiconductor device that processes instructions stored in memory and/or storage.

The memory and storage may include various types of volatile or non-volatile storage media. For example, memory may include ROM and RAM.

Accordingly, the open market seller service providing method according to an embodiment of the present invention may be implemented as a method executable on a computer. When the open market seller service providing method according to an embodiment of the present invention is executed in a computer device, computer readable instructions may perform the open market seller service providing method according to the present invention.

Meanwhile, the above-described open market seller service providing method according to the present invention can be implemented as computer readable code on a computer readable recording medium. Computer-readable recording media includes all types of recording media in which data that can be decoded by a computer system is stored. For example, there may be read only memory (ROM), random access memory (RAM), a magnetic tape, a magnetic disk, a flash memory, an optical data storage device, and the like. In addition, the computer-readable recording medium may be distributed in computer systems connected through a computer communication network, and stored and executed as readable codes in a distributed manner.

5. Forecasting Time Series Demand for Open Market Products

Procurement of an appropriate quantity is very important when selling products in the open market, as inventory issues occur when products are procured in excess of demand, and products cannot be sold to consumers due to lack of inventory when products are procured less than demand. This is an important issue.

SMEs (Small and Medium Enterprises), which need to operate with a small number of people in the process of item selection, sourcing (purchasing), and sales, refer to the quantity sold in the previous month or estimate additional quantities by simply predicting demand based on intuition. There is a control problem.

16 illustrates an apparatus for predicting time-series demand for open market products according to another embodiment of the present invention.

Prediction of an appropriate quantity for sales is performed using demand (open market search volume), supply (number of products registered in open market), and sales (number of open market product purchases) information.

For example, in the case of vases, if you look at the trend of search volume for a preset period (eg, 2 years), you can predict demand by selling well in March and April each year and gradually decreasing sales volume from May, and the number of registered products (supply) information (If there were 500 competing products in March 2020, but if there were 1,000 competing products in March 2021, the competition doubled).

For products sold in the open market, information on the number of purchases is provided to help consumers make purchasing decisions.

According to an embodiment of the present invention, when a seller intends to sell a corresponding product or a similar product, the seller calculates demand by considering demand (increase in search volume) and supply (increase in the number of registered products) based on sales volume information.

If the monthly sales volume is 100 units, next month's demand is expected to decrease by 10% compared to this month, the competing product is expected to increase by 10%, and the sales volume of the product is decreasing by 5%, the sales demand for the product in the next month is calculated as 77 units. You can, and you can adjust your inventory accordingly.

Calculation of the appropriate quantity according to an embodiment of the present invention is performed by [Equation 1] below.

[Equation 1]

For example, if the monthly sales volume is 100 units, if the demand increase/decrease value decreases by 10%, the supply increase/decrease value increases by 20%, and the sales volume increase/decrease value decreases by 5% through regression analysis, the appropriate quantity is 100×0.9×1.2× It becomes 0.95.

An apparatus for predicting time-series demand for open market products according to an embodiment of the present invention includes an input unit 1610 that receives information for predicting demand, and a memory 1620 storing a program for deriving an appropriate quantity using the information for predicting demand. and a processor 1630 that executes the program, and the processor 1630 derives appropriate quantity information by using information on increase/decrease in demand, increase/decrease in supply, and increase/decrease in sales volume for products listed on the open market platform.

The processor 1630 derives appropriate quantity information using demand increase/decrease numerical information through search volume trend analysis in the open market.

The processor 1630 derives appropriate quantity information using supply increase/decrease numerical information through analysis of the number of products registered in the open market.

The processor 1630 derives appropriate quantity information using numerical information of sales volume increase and decrease through analysis of the number of product purchases in the open market.

The processor 1630 provides an alarm to the user who has derived the appropriate quantity information using the open market monitoring information. For example, when it is determined that proper quantity information needs to be changed using open market monitoring information to a seller whose sourcing process has not yet been completed after deriving proper quantity information, an alarm is provided so that it can be referenced for sourcing.

The processor 1630 proposes a sales pricing strategy for the corresponding item in addition to appropriate quantity derivation information derived through the increase/decrease in demand, increase/decrease in supply, and increase/decrease in sales volume. For example, when the daily increase in supply is monitored to be greater than or equal to a preset value, it is determined that the competitor has risen sharply, and the competitor's pricing status is analyzed. Considering the results of analysis of competitors' pricing status and referring to figures of changes in demand, recommend pricing strategies for products currently being sold by sellers, and provide expected profit information for the recommended pricing strategies. From this, sellers are supported to select a strategy that can ultimately obtain a larger expected profit by gaining an edge in terms of sales volume even if the selling price per unit is discounted when competitors surge and demand soars. .

17 illustrates a method for predicting time-series demand for open market products according to another embodiment of the present invention.

The method for predicting time-series demand for open market products according to an embodiment of the present invention includes calling item information (S1710) and providing a derivation result for an appropriate quantity (S1720).

In step S1710, current sales volume, demand increase/decrease value, supply increase/decrease value, and sales volume increase/decrease value information of the corresponding item are called.

For the increase and decrease in demand, increase and decrease in supply, and increase and decrease in sales, the information is called from the ensemble of three regression analysis models, and analysis using time series data is performed to analyze the increase and decrease of the three factors.

Step S1720 derives a numerical value for an appropriate quantity using the current sales volume, demand increase/decrease value, supply increase/decrease value, and sales volume increase/decrease value information, and provides the derived result.

Calculation of the appropriate quantity is as shown in [Equation 1].

According to the present invention, there is an effect that it is possible to adjust the quantity of products for sale by referring to demand data for which systematic forecasting has been performed.

It is possible to reduce logistics costs such as storage by efficiently performing inventory management by appropriately performing the quantity of products for sale.

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

Meanwhile, the method for predicting time-series demand for open market products according to an embodiment of the present invention may be implemented in a computer system or recorded in a recording medium. A computer system may include at least one processor, a memory, a user input device, a data communication bus, a user output device, and a storage. Each of the aforementioned components communicates data through a data communication bus.

The computer system may further include a network interface coupled to the network. The processor may be a central processing unit (CPU) or a semiconductor device that processes instructions stored in memory and/or storage.

The memory and storage may include various types of volatile or non-volatile storage media. For example, memory may include ROM and RAM.

Accordingly, the method for predicting time-series demand for open market products according to an embodiment of the present invention may be implemented in a computer-executable method. When the method for predicting time-series demand for open market products according to an embodiment of the present invention is executed in a computer device, computer readable instructions may perform the method for predicting time-series demand for open market products according to the present invention.

Meanwhile, the above-described method for predicting time-series demand for open market products according to the present invention can be implemented as computer-readable code on a computer-readable recording medium. Computer-readable recording media includes all types of recording media in which data that can be decoded by a computer system is stored. For example, there may be read only memory (ROM), random access memory (RAM), a magnetic tape, a magnetic disk, a flash memory, an optical data storage device, and the like. In addition, the computer-readable recording medium may be distributed in computer systems connected through a computer communication network, and stored and executed as readable codes in a distributed manner.

6. Provision of overseas exhibition product information

Sellers use trade fairs to buy goods in the distribution market, but due to Corona 19, trade fairs around the world have been canceled or postponed, and most have been replaced by online fairs.

However, there are many distributors who are unaware of this fact, and even if they want to use online fairs, it is difficult to access and check the fair sites individually.

Even when a seller visits an online exhibition site to find a new product and finds a product that the seller wants to sell, there is a problem in that it takes a considerable amount of time to organize product information because each exhibition has a different format.

18 illustrates an overseas exhibition product information providing device according to another embodiment of the present invention.

According to an embodiment of the present invention, product information and trend information of online fairs scattered in each site are collected and provided to the seller, and classification and index refinement are performed to retrieve product and brand information of the fair.

For each fair, the classification of categories is different. This is because no specific standards for product classification are applied. For example, in the case of "lighting", the category may be different from lighting, interior lighting, lighting, etc. for different overseas exhibitions.

Therefore, in order to provide the category classification of each overseas exhibition collected by the crawler (or manual) in the service according to the embodiment of the present invention, redefinition is performed as an integrated category classification.

In general, a method of mapping and swapping categories from each overseas exhibition to an integrated category is used. If there are 2,000 categories in one overseas exhibition, mapping to the integrated category should be performed 2,000 times. In this case, human error may occur due to repetitive work by manpower.

According to an embodiment of the present invention, through word embedding, automatic mapping is performed to an integrated category having the shortest distance on a vector from a category of a specific overseas exhibition. Automatic mapping/conversion to categories is performed, so it is possible to accumulate in the DB without the need for additional work.

The unified category classification may be created based on a category of a predetermined target platform.

An overseas exhibition product information providing device according to an embodiment of the present invention includes an input unit 1810 for receiving a product search request for an overseas online exhibition, and a program for searching products at a plurality of overseas online exhibition sites according to the product search request is stored. It includes a memory 1820 and a processor 1830 that executes a program, and the processor 1830 selects a product corresponding to a product inquiry request in a plurality of overseas online fairs and provides a selected list.

The input unit 1810 receives a product search request including at least one of keyword information, fair information, and category information.

The processor 1830 generates and provides exhibition product information including brand name, brand description, product name, product image, product description, and contact information.

When the seller does not know detailed information about the fair, the processor 1830 selects and provides a product list using any one of input keyword information and category information and user setting information. Product Overseas product information is provided by referring to the information set to select and provide the list. The processor 1830 selects and provides a list of fairs or products that are expected to be of interest to the seller in consideration of setting information as well as keyword information and category information, but displays the fair or Adjust the order in which products are displayed on the list. In addition, the processor 1830 generates exhibition characteristic information using history information of each overseas exhibition, and selects and provides recommended exhibition information in consideration of a user's search intention from setting information.

19 illustrates a method of providing product information at an overseas exhibition according to another embodiment of the present invention.

A method for providing product information at overseas fairs according to an embodiment of the present invention includes receiving at least one of keyword information, fair information, and category information (S1910), selecting and providing fair product information (S1920), and fair product information. and performing at least one of filtering and sorting on (S1930).

Step S1920 selects product information listed in a plurality of overseas online fairs according to at least one of keyword information, fair information, and category information. At this time, the fair product information includes brand name, brand description, product name, Includes product images, product descriptions, and contact information.

Step S1930 searches similar products from a plurality of overseas wholesale platforms using image information (image file, image link) or keyword information for product information at overseas fairs, and filtering or sorting in consideration of factory condition information and product condition selection information provides results.

Factory condition information includes in-platform supply level information, supply volume information, product description matching information, response speed information, delivery speed information, refund rate information, and dispute rate information.

Product condition information includes specific keyword information, category information, price information, shipping cost information, and MOQ (minimum quantity) information.

20 illustrates a process of providing product information at an overseas exhibition according to another embodiment of the present invention.

When an overseas exhibition is selected on the page of an overseas exhibition product information providing service provider, it is possible to select and display the result of the corresponding exhibition.

According to the exemplary embodiment of the present invention, there is an effect that allows a seller who searches for products to source products to check product information for each fair, category, and keyword.

Even if the seller does not know the exact name of the exhibition, it is possible to check products of famous overseas exhibitions by using categories or keywords, so that complementary search results using search criteria can be provided.

By processing multiple overseas trade fair information into one format (brand name, brand description, product image, product description, contact information) and providing it to the seller, you can clearly see the products of interest without having to go to the fair or arrange them yourself. It is possible to support to check.

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

On the other hand, the overseas fair product information providing method according to an embodiment of the present invention can be implemented in a computer system or recorded on a recording medium. A computer system may include at least one processor, a memory, a user input device, a data communication bus, a user output device, and a storage. Each of the aforementioned components communicates data through a data communication bus.

The computer system may further include a network interface coupled to the network. The processor may be a central processing unit (CPU) or a semiconductor device that processes instructions stored in memory and/or storage.

The memory and storage may include various types of volatile or non-volatile storage media. For example, memory may include ROM and RAM.

Accordingly, the method for providing product information at overseas exhibitions according to an embodiment of the present invention may be implemented as a method executable on a computer. When the method for providing product information at overseas exhibitions according to an embodiment of the present invention is executed in a computer device, computer-readable instructions may perform the method for providing product information at overseas exhibitions according to the present invention.

Meanwhile, the above-described method for providing product information at overseas fairs according to the present invention can be implemented as computer-readable codes on a computer-readable recording medium. Computer-readable recording media includes all types of recording media in which data that can be decoded by a computer system is stored. For example, there may be read only memory (ROM), random access memory (RAM), a magnetic tape, a magnetic disk, a flash memory, an optical data storage device, and the like. In addition, the computer-readable recording medium may be distributed in computer systems connected through a computer communication network, and stored and executed as readable codes in a distributed manner.

7. Open market database construction and service provision

As the e-commerce market expands, various open market platforms are being created, and various data are being produced on the open market platforms.

However, the open market information is scattered from the buyer's point of view, and it only provides information about the seller's products and stores (store), and the seller information in the open market is on the internal page (similar to the admin page). Available in a separate form.

21 illustrates an apparatus for constructing an open market database and providing services according to another embodiment of the present invention.

According to an embodiment of the present invention, open market information is reconstructed and provided from the seller's point of view.

In the open market, products are searched and found, so 'search word' becomes 'demand' and 'number of registered products' becomes 'supply'.

That is, it is advantageous for the seller to select a market with a large search volume and a small number of registered products.

By the way, search volume and number of registered products are important, but information on various factors (age ratio, gender ratio, whether a brand is monopolized, existence of a power seller, etc.) is necessary for market analysis from multiple angles.

In addition, it is possible to estimate sales by checking data (number of products sold, sales volume) of stores promoting in the market selected by the seller.

When selecting an item, a seller considers whether to sell a product that is currently selling well or whether to sell a product that will sell well in the future. Users who supply, predict and sell 'goods that will sell well' can develop new markets or satisfy new needs of buyers.

It is possible to confirm that a product is selling well or a product with a high sales volume through product data of a store, and it is possible to confirm that a product that is selling well is a product that corresponds to a search word whose monthly search volume (demand) is increasing.

Through the above factors and data, it is possible to understand and analyze the market even if the level of knowledge about the market is not high.

In other words, it is possible to analyze the market and select items based on data, rather than relying on intuition to predict the market and select items.

According to the prior art, open market data services from the seller's point of view have been proposed, but they provide information by scraping (crawling) in real time, providing the same type of information as the method (eg, keyword input) provided in the open market. It is possible to provide it, but there is a limit to not being able to conduct a multilateral market analysis through reconstruction.

That is, it is possible to check a list of currently sold products with the search word 'dumbbell' and check the sales volume of each product list, but it is impossible to search for products with monthly sales of 1,000 or more without a search word.

In addition, when a failure occurs in the open market platform or a schema is changed, there is a problem in that immediate service response is not possible.

According to an embodiment of the present invention, a service that reconstructs and builds open market data based on products, keywords, and stores, and provides information through a web or app using this is proposed.

An apparatus for constructing an open market database and providing services according to an embodiment of the present invention includes an input unit 2110 for receiving search word information, store information, and product information, and a program for constructing an open market database and providing an information screen according to a user call. It includes a stored memory 2120 and a processor 2130 that executes a program, and the processor 2130 builds a product, store, and keyword database and performs complementary updates to the database.

In constructing a product database, the processor 2130 collects a unique ID of a specific open market platform in relation to a product and stores the entire database as a dump on a periodic basis.

In constructing the store database, the processor 2130 collects a unique ID of a specific open market platform in relation to the store, and stores the entire database as a dump on a periodic basis.

In building a search word database, the processor 2130 checks whether the database exists by category and product name, generates an ID of a preset number of digits, and stores the entire database as a dump for each cycle.

The processor 2130 generates a related search word registered to a product, creates and updates a list of products mapped to the search word, creates store information registered to the product, and creates and updates a list of products registered to the store, update information

The processor 2130 creates a list of products registered in a search word, creates and updates store information registered in a product, creates and updates search word information registered in a product, and updates the search word information.

The processor 2130 generates store information, creates and updates a product list registered in the store, and updates the store information.

22 illustrates a method for constructing an open market database and providing a service according to another embodiment of the present invention.

A method for constructing an open market database and providing services according to an embodiment of the present invention includes collecting and constructing an initial database of products, stores, and keywords (S2210), and performing complementary updates to each database (S2220). and calling the constructed information to provide on-screen information (S2230).

23 illustrates the relationship between products, keywords, and stores according to another embodiment of the present invention.

A product belongs to one store, a search term is mapped to a product, and a store is composed of products.

A product has a related search word through product name, tag, and information information.

In constructing a product database according to an embodiment of the present invention, a unique ID of a specific open market platform is collected in relation to a product, and the entire database is dumped for each period.

According to the prior art, when a user searches for a keyword, information in the DB is updated, and in the case of a keyword that is not frequently searched for, there is a problem of providing old information. For example, if the last user inputs "bachelor kimchi" in March 2020, if the user inputs "bachelor kimchi" in March 21, then March 2020 information is provided and updated with March 21 data. carry out

This is to display data quickly, but the user has to perform a search twice consecutively with an interval in order to view the latest information, which is inconvenient.

That is, in the case of keywords that are not frequently entered, there is a limit to ensuring data up-to-date until two consecutive searches are performed, and when keyword information is updated, the tag or form of the crawled source (e.g., Naver) is changed. If so, immediate update is not possible, and old data may be displayed.

According to an embodiment of the present invention, all input keywords are updated to the latest for a predetermined period of time (eg, one week), and a batch operation of dumping them is performed.

According to an embodiment of the present invention, a process of updating when a user inputs a keyword within a service is also performed in parallel with the above process.

Through this, there is an effect in that the user can be provided with the latest information that has been updated at least before a predetermined period (eg, 1 week) regardless of changes in the platform.

Product information is parsed through open platform menus, keyword search, and user input (keywords, etc.), new creation or update is performed by checking whether a product exists in the database, time series information (price, sales volume, etc.) and snap Save the shot information to the database.

Product database fields are shown in [Table 1] below.

id rank url imageurl adId productTitle priceUnit openDate category1Id category2Id category3Id category4Id categoryPlusID category1Name category2Name category3Name category4Name category4Name category4NameOrg categoryLevel adcrUrl mobileLowPrice lowPrice price deliveryFeeContent reviewCountSum purchaseCnt mallCount zzim scoreInfo spider purchaseCntAverage reviewRefresahedAt salesAverage reviews mallId mallInfoCache mallIntroduction mallName mallNameOrg mallNo mallPcUrl mallProdMblUrl mallProductId mallProductUrl brand BrandNo maker makerNo relatedKeyword Tag review_summary OverseaTp

In building a store database according to an embodiment of the present invention, a unique ID of a specific open market platform is collected in relation to a store, and the entire database is dumped and stored for each period.

Using the open platform menu and store information of the product, open platform API calls are made and search term information is parsed to check whether the database exists, create or update a new store, and perform time series information (monthly sales, number of products, etc.) and snapshot information are stored in the database.

Store database fields are shown in [Table 2] below.

mallId mallIntroduction mallName mallNameOrg mallNo mallPcUrl mallProdMblUrl mallProductId mallProductUrl mallProductTotal mallProductList grade csGrade relatedKeyword salesAverage salesTotalPrice openDate storeZzim mallId mallIntroduction mallName mallNameOrg mallNo mallPcUrl

In the construction of a search word database according to an embodiment of the present invention, since the search word does not have a unique ID of a specific open market platform, it is determined whether the DB exists by category (eg, life / health), name (eg, dumbbell), An 8-digit ID is created, and the entire database is dumped for each cycle.

According to product search word information and search word (keyword) search, open platform API is called and search word information is parsed. New creation or update is performed by checking whether the search term exists in the database, and time series information (monthly search volume, number of registered products, etc.) and snapshot information are stored in the database.

Search word database fields are shown in [Table 3] below.

keywordId keywordName rank cId level createdAt source searchCount degree totalItem searchCnt clickAvg sallesSum avgPrice adFee AvgSalesTotal isSeason reviewCountSum recentReveiwSum scoreInfo powerSum keywordProductList source relatedKeyword

24 to 26 illustrate complementary entity creation and update processes between product, store, and keyword DBs.

Referring to FIG. 24 , product information is created (S2410), a related search word registered to the product is created and updated (S2420), and a product list mapped to the search word is created and updated (S2430). In addition, store information registered in the product is created and updated (S2440), and a product list registered in the store is created and updated (S2450).

Referring to FIG. 25 , search word information is generated (S2510), a product list registered to the search word is created and updated (S2520), and store information registered to the product is created and updated (S2530). In addition, search word information registered in the product is created and updated (S2540).

Referring to FIG. 26 , store information is created (S2610), and a product list registered in the store is created and updated (S2620).

According to an embodiment of the present invention, when a user query is input, a built database is called and provided on the screen. At this time, new data (product, store, search term) is created and updated.

27 illustrates a product analysis information providing screen according to another embodiment of the present invention, and FIG. 28 illustrates a search word analysis information providing screen according to another embodiment of the present invention.

Product analysis information includes information on shipping cost, monthly average sales volume, advertisement status, product registration date, number of sellers, number of reviews, rating, number of favorites, expected profit, and price.

Search word analysis information includes the number of registered products, monthly search volume, average number of monthly ad clicks, average monthly sales volume, average price, advertising cost, seasonal product, total number of reviews, star rating, number of power sellers, competitive strength, proportion by device, and gender proportion , share by age, and monthly search trend information are included.

According to an embodiment of the present invention, a seller (user) can check information by product, store, and search word, and can check combined information according to a request (e.g., a combination of keyword and product information). information).

Product, store, and search term databases are updated in real time in a complementary manner to secure the latest data, and dump storage is performed periodically in preparation for system failure handling or schema change, so that normal data can be reliably checked.

According to the prior art, open market data is scraped or crawled in real time, and the data is organized and provided so that the call speed performance is low and it is difficult to prepare for changes in the open market. By constructing and creating and updating complementary databases, improving user query call speed and dumping each database and the entire database to prepare for changes in the open market, market analysis and item selection are always stable. It is possible to provide related services.

Meanwhile, the open market database construction and service providing method according to an embodiment of the present invention may be implemented in a computer system or recorded in a recording medium. A computer system may include at least one processor, a memory, a user input device, a data communication bus, a user output device, and a storage. Each of the aforementioned components communicates data through a data communication bus.

The computer system may further include a network interface coupled to the network. The processor may be a central processing unit (CPU) or a semiconductor device that processes instructions stored in memory and/or storage.

The memory and storage may include various types of volatile or non-volatile storage media. For example, memory may include ROM and RAM.

Accordingly, the method for constructing an open market database and providing a service according to an embodiment of the present invention may be implemented as a method executable on a computer. When the open market database construction and service provision method according to an embodiment of the present invention is executed in a computer device, computer readable instructions may perform the open market database construction and service provision method according to the present invention.

Meanwhile, the above-described method for constructing an open market database and providing a service according to the present invention can be implemented as computer readable codes on a computer readable recording medium. Computer-readable recording media includes all types of recording media in which data that can be decoded by a computer system is stored. For example, there may be read only memory (ROM), random access memory (RAM), a magnetic tape, a magnetic disk, a flash memory, an optical data storage device, and the like. In addition, the computer-readable recording medium may be distributed in computer systems connected through a computer communication network, and stored and executed as readable codes in a distributed manner.

29 and 30 illustrate an apparatus for extracting pros and cons keywords based on product reviews according to another embodiment of the present invention.

Referring to FIG. 29, an apparatus for extracting pros and cons keywords based on product reviews according to another embodiment of the present invention is an analysis target using an input unit 2910 for receiving review data on a product to be analyzed on an online sales page, and the review data. A memory 2920 storing a program for generating merit keyword information and demerit keyword information about a product, and a processor 2930 executing the program, the processor 2930 analyzes the review data to obtain positive review information and negative review information. , analyzes positive review information to generate the merit keyword information, and analyzes negative review information to generate the demerit keyword information.

The processor 2930 analyzes the review data and classifies it into positive review information and negative review information based on predetermined review score criteria.

The processor 2930 extracts keywords by analyzing appearance frequencies and relationships between words in review data composed of sentences, and generates strength keyword information and weakness keyword information.

The processor 2930 decomposes the extracted keyword into morpheme units to generate a basic keyword, and generates strength keyword information and weakness keyword information based on the decomposition of the extracted keyword.

The processor 2930 generates and provides the strength keyword information and the weakness keyword information in words, and provides them in a predetermined number of words.

Referring to FIG. 30 , an apparatus for extracting pros and cons keywords based on product reviews according to another embodiment of the present invention includes a review extractor 3100, an advantage keyword extractor 3210, a disadvantage keyword extractor 3220, and an analysis result integrator. (3300).

The review extractor 3100 receives product data 3001 of an online shopping mall, receives product data 3001 including review data from a sales page of a target product from which keywords are to be extracted, and distinguishes between positive reviews and negative reviews. do.

At this time, the product data 3001 is product data registered on the sales site, which necessarily includes the name of the product, reviews and star ratings, and category information of the product, as well as other information (e.g., price, source, etc.) ) may also be included.

The review extractor 3100 transmits positive review information to the merit keyword extractor 3210 and transmits negative review information to the demerit keyword extractor 3220 .

The review extractor 3100 classifies each review into positive review information or negative review information based on a score registered with the review (eg, a star score or a numerical score).

For example, reviews with 1 or 2 points are classified as negative review information, and reviews with 4 or 5 points are classified as positive review information, according to preset conditions in reviews with a perfect score of 5 stars.

The merit keyword extractor 3210 and the demerit keyword extractor 3220 extract merit and demerit keywords from the positive and negative review information classified by the review extractor 3100 .

The analysis result integrator 3300 integrates the merit keyword information extracted by the merit keyword extractor 3210 and the demerit keyword information extracted by the demerit keyword extractor 3220, and outputs a product pros and cons keyword 3002. .

The analysis result integrator 3300 integrates the pros and cons keyword information and outputs product pros and cons keywords 3002. Contains keywords and is provided in word form.

31 illustrates keyword extraction according to another embodiment of the present invention.

In the keyword extraction process, merit keywords and demerit keywords are extracted, respectively. If the input review data 3003 is positive review information, a merit keyword is extracted, and a demerit keyword is extracted if it is negative review information.

The keyword ranking extractor 3211 receives the review data 3003 and extracts keywords included in the reviews, and extracts keyword information including ranks and scores.

The keyword ranking extractor 3211 extracts keywords having rankings and scores by analyzing appearance frequencies and relationships between words in review data composed of several sentences.

The morpheme analyzer 3212 decomposes the extracted keyword into morphological units.

The morpheme analyzer 3213 decomposes the extracted keywords into morpheme units to generate keywords in a basic form.

For example, 'pretty', 'pretty', 'pretty', etc. have the same meaning but are expressed in different forms, so each has a low score and can be included in the extraction result in a redundant state. It is possible to avoid and establish true scores and rankings.

The post-processing unit 3213 prevents duplication of results, performs processing to exclude specific words, and outputs the keyword 3004 as a result.

The post-processing unit 3213 excludes English, numbers, and redundant words from the result, and excludes the merit keyword when extracting the demerit keyword to generate a more accurate result.

There are cases in which meaningless reviews such as "asdfgasdsasdff" are left in order to obtain benefits such as points awarded according to leaving a review of a predetermined standard (eg, 10 or more characters).

That is, words left in languages other than the official language of the region where the service is provided (eg, Korea) are excluded when extracting keywords.

As another example, it is determined whether a word included in the review data is a word included in the dictionary, and if the word is not included in the dictionary (in the case of "asdfgasdsasdff" described above), it is excluded from the keyword extraction process and included in the dictionary. In the case of a word (eg, great, good, etc.), it is possible to determine it as a meaningful review and include it when extracting keywords.

In addition, there are cases in which it is difficult to grasp the meaning of product evaluation with words included in numbers.

For example, in the case of "I ordered on the 3rd and it arrived on the 6th", words from "3rd and 6th" are included in the dataset as sentences are divided by word, and these words are included when selecting keywords. Since there is room for interpretation as a keyword unrelated to product evaluation, the post-processing unit 3213 excludes numbers from the result.

The keyword 3004, which is a result output through the above-described processes, includes keywords for strengths and weaknesses of products to be analyzed extracted according to the types of input reviews.

32 illustrates a method for extracting pros and cons keywords based on product reviews according to another embodiment of the present invention.

A method for extracting pros and cons keywords based on product reviews according to another embodiment of the present invention includes extracting review data using product data (S3201), extracting keyword rankings (S3203), performing morphological analysis and post-processing. A step S3203 and a step S3204 of providing keyword analysis results are included.

In step S3201, product data on a target product to be analyzed that is sold on the online product sales webpage is acquired, and review data is extracted.

At this time, in step S3201, the data is classified into positive review data and negative review data based on preset review scores.

Step S3202 extracts keywords included in the review data, and extracts keyword information including rankings and scores.

In step S3203, morpheme analysis is performed on the extracted keywords, keywords are generated in a basic form, and post-processing is performed.

In step S3204, merit keyword information and demerit keyword information for the product to be analyzed are provided as an analysis result.

When a user leaves a positive review because of a high level of satisfaction with the corresponding product, the actual rating may be given as 1 point out of 5 points due to the user's mistake.

In this case, as another embodiment, analysis of the review sentence is performed first, and it is confirmed whether the assigned review score (star score) matches the positive/negative classification of the actual review.

If the review score given as described above is 1 point, but as a result of the analysis of the review sentence, there is no negative opinion in the content (that is, if it is confirmed that there is no negative word use through the keyword extraction and morphological analysis techniques described above) case), the review is treated as positive review information without considering the assigned star rating.

In addition, if it is not a mode of leaving each review score by category (e.g., if it is set to give a star rating for the product itself and a star rating for delivery separately), if the product itself is satisfied but dissatisfied with the delivery, or The reverse may be the case.

In this case, as another embodiment, it is preferable to generate merit keyword information and demerit keyword information for each item based on the analysis result of the review data rather than simply determining a positive/negative review based on the review score itself.

At this time, it is possible to identify relationships such as proximity between each word (for example, in the case of review data such as "I like the product, but delivery is too slow", or to identify the relationship between product-like-like only for the product itself) It is determined that it is a positive review, the relationship between delivery-too-slow is identified, and the delivery is judged to be a negative review), and whether or not connection adverbs (but, but, but, but, etc.) are used are identified, and each item (product , delivery, counseling, etc.) generate merit keyword information and demerit keyword information.

As described above, by identifying the relationship between each word included in the review data, if the same “fast” is extracted as the advantage keyword information, the user can determine whether the delivery is fast or the operation of the product (e.g., laptop computer) is fast. It is possible to include and provide additional keyword information (delivery, operation, etc.) in keywords such as “quick (delivery)” or “quick (action)” so as to be smoothly checked.

33a and 33b illustrate keywords extracted by the apparatus for extracting pros and cons keywords based on product reviews according to another embodiment of the present invention.

The number of positive reviews for a specific product name, the number of negative reviews, examples of positive reviews, examples of negative reviews, keywords for positive reviews of products and keywords for negative reviews ( The corresponding keyword can be provided as a keyword corresponding to the top n ranks in terms of frequency of occurrence according to a preset setting).

34 shows an artificial intelligence-based product transaction automation system according to another embodiment of the present invention.

An artificial intelligence-based product transaction automation system according to another embodiment of the present invention collects price information of the same product uploaded to a plurality of platforms and extracts a list of target products with a possibility of profit realization. 3402), an automatic product registration unit 3403 for registering the target product on the target platform, and determining whether or not to expose the target product on the target platform using the lowest price monitoring information on the target product, and using the lowest price monitoring information Product exposure maintenance and profit determination unit 3404 that determines whether expected profit meets a preset margin standard and when a customer's order for a target product is received, the target product is purchased on a preset platform, and the target product is and a product purchase and delivery control unit 3405 that controls delivery to the customer.

The product list extraction unit 3402 uses the product big data 3401 to extract a list in consideration of price differences and distribution costs of the same product uploaded to a plurality of platforms.

For example, it is assumed that product A is sold for 6,372 won on a Chinese platform called B. And, it is assumed that the same product A is sold at 18,800 won, 19,000 won, and 19,100 won by other sellers on a Korean platform called C, respectively.

If the transportation cost required to purchase product A on platform B and ship it to Korea is KRW 6,000, and the eCP fee is KRW 2,190, purchase product A on platform B and complete distribution costs (KRW 8,190) until delivery to Korea. ), the total cost is 14,562 won.

In other words, even for the same product A, if the selling price is sold at KRW 18,800 using platform C as the product sales platform, and the product is purchased and delivered to the customer using platform B as the product purchase platform, the profit margin is KRW 4,238, which is 23%. of revenue is calculated.

That is, the product list extraction unit 3402 determines that the value obtained by subtracting distribution costs from the price difference between platforms of the same product A exceeds a preset margin (eg, 0 won or 1000 won). In this case, the corresponding product A is determined as a target product capable of realizing profits.

The automatic product registration unit 3403 can register the target product on the C platform, and can register on the D and E platforms other than the C platform in the above example.

If the product exposure maintenance and profit determination unit 3404 determines that the margin margin collapses when the price of the target product is updated to the lowest price using the lowest price monitoring information, the sale of the target product is stopped.

For example, it is assumed that another competitor newly registers product A on platform C and sells it at 15,000 won.

Assuming that the selling price of product A on platform B and overall costs for distribution are the same as in the previous example, the price of product A on platform C is adjusted from 18,800 won to 15,000 won, taking into account the selling price of competitors. In this case, the profit is 438 won.

In this case, if the preset margin margin line is 400 won, since the margin is greater than the preset margin margin line, the product exposure maintenance and profit determination unit 3404 automatically adjusts the price of product A to 15,000 won and registers it. Control information is transmitted to the registration unit 3404.

As another example, when the preset margin margin line is 500 won, since the margin is smaller than the preset margin margin line, the product exposure maintenance and profit determination unit 3404 determines whether or not to sell product A. The automatic product registration unit 3404 ) to transmit control information.

As another example, if the preset margin margin line is 500 won, it is determined that the margin is smaller than the preset margin margin line, and the product exposure maintenance and profit determination unit 3404 sets the price of product A on the C platform to the lowest price. It is possible to transmit control information to the automatic product registration unit 3404 to register at a discount price (eg, 15,063 won) considering a margin margin line other than . The product exposure maintenance and profit determination unit 3404 considers the advertising cost, the shopping mall exposure maintenance cost, and the expected sales revenue of the target product (product A) to determine the price after discount or adjust the preset margin margin line. it is possible For example, if it is confirmed that the sales volume of product A is increasing by more than a preset standard, it is possible to adjust the margin margin line or adjust the sales price of product A at a discount, rather than simply setting the lowest price or making a decision to stop sales. . For example, the margin margin line is adjusted from 500 won to 400 won, and the selling price of product A is adjusted to 15,000 won. This is because even if margins are reduced, profits can be realized by selling large quantities of products.

When the product exposure maintenance and profit determination unit 3404 updates the price of the target product to the lowest price using the lowest price monitoring information, if it is determined that the margin margin line collapses, the existing selling price of the target product is maintained, Control to register by discounting the existing sales price.

An artificial intelligence-based product transaction automation system according to another embodiment of the present invention has a function of extracting a product having a price difference between platforms, a function of registering the entered product in a shopping platform to sell, and a function of registering the lowest price of the registered product. It includes a function to maintain and a function to register advertisements for product exposure.

An artificial intelligence-based product transaction automation system according to another embodiment of the present invention automatically orders the same product from the lowest price shopping platform and delivers it to consumers when an order for a product registered for sale is generated.

35 illustrates an artificial intelligence-based product transaction automation method according to another embodiment of the present invention.

An artificial intelligence-based product transaction automation method according to another embodiment of the present invention includes the steps of finding and registering a product, determining the lowest price and displaying the product so as to increase the sales probability of the registered product, and performing the product exposure. It includes the step of purchasing and shipping to consumers, and the step of recording orders to use as AI learning data.

Specifically, referring to FIG. 35 , a product list is extracted and a target product is registered on the target platform in consideration of price information and distribution costs for products listed on the platform (S3501).

It is checked whether the price of the corresponding target product is the lowest price in the ECP (S3502), and if the price is the lowest, product exposure management is performed (S3506).

Subsequently, when an order for the target product is received from the customer, the lowest price purchase for the product is performed, and control is performed to deliver the product to the customer (S3507).

If it is confirmed in step S3502 that the lowest price in the ECP is not, in step S3503 the price of the target product is updated to the lowest price (S3503), and it is determined whether or not there is a profit.

It is checked whether the margin margin line collapses according to the lowest price update (S3504), and if the margin margin line does not collapse, product exposure management in step S3506 is performed.

When the margin line collapses, it is determined whether to stop sales (S3505), and product exposure management is performed in step S3506.

As described above, if it is confirmed that the margin margin line collapses, sales of the corresponding target product may be suspended, and it is also possible to adjust the margin margin line and adjust the selling price of the target product.

An artificial intelligence-based product transaction automation system and method according to another embodiment of the present invention applies a purchasing agent as a distribution behavior.

That is, without inventory, a product is registered on the platform, and when an order is generated for the registered product, the product is purchased from the lowest priced platform and delivered to the customer.

The artificial intelligence-based product transaction automation system and method according to another embodiment of the present invention can be applied to the behavior of holding and selling products through purchase of products, but physical automation in that products must be delivered directly when an order occurs. Perform and apply additional calculations related to logistics warehousing costs and processing costs in relation to and remaining goods.

36 illustrates a multi-platform cheapest product search assistance system according to another embodiment of the present invention.

A multi-platform lowest price product search assistance system according to the present invention includes an image extraction and input unit 3601 for extracting an image from a browser and inputting the extracted image, and a multi-shopping platform search unit for searching for a product on a multi-shopping platform for the image. 3602, a search result filtering unit 3603 for filtering results retrieved from the multi-shopping platform, and a purchasing agent unit 3604 for performing a purchasing agent when an order for a product is received.

The image extraction and input unit 3601 utilizes a browser plug-in to extract an image as a button overlaid on an image is provided on a website and the button is clicked.

When a user places a mouse on an image of a specific shopping platform website, a button is displayed in a predetermined area (eg, an upper left area of the image).

This button may be a mark of a service provided by a service provider, or may be a button including text such as "search for the lowest price overseas product", for example.

The multi-shopping platform search unit 3602 provides overseas product information for an image extracted when a button is clicked through a website of a service provider.

That is, a new window or tab is opened in the web browser, and information about a product clicked on a button overlaid on an image is controlled to be displayed through a website of a service provider.

At this time, if the user selects any one of the overseas product information listed through the service provider's website, the contents of the web page where the product can be ordered is displayed through a link in the service provider's website. is expressed

The filtering unit 3603 filters the search results according to classification criteria including lowest price and sales volume.

The filtering unit 3603 performs a list-up of products corresponding to the same image as the image searched by the user through image search within the multi-platform, and provides filtered search results in order of price and sales volume.

Also, the filtering unit 3603 may display a product related to a product of interest to the user as a recommended product through a separate area.

For example, if the product that the user clicks on the overlaid button is “duvet organizer”, “vacuum vacuum pack, compression pack” may be displayed as a recommended product through a separate area.

At this time, it is possible to utilize a database including related product information classified for each target product according to a preset classification criterion, and such a database is information on search results in chronological order of other users or group purchases of other users. It can be built using information, etc.

Through this, the input information is only the user's click on the overlaid button, but the user's search intention (to arrange the futons) is identified, and multi-shopping platform search results for recommended products related to it are provided to the user. to increase satisfaction with search results.

According to another embodiment of the present invention, a search result of a multi-shopping platform for a corresponding image is provided by utilizing an image URL, and a multi-shopping platform is searched by using shopping information of a URL of a webpage with the corresponding image together. It is possible to improve the accuracy.

In this case, the shopping information of the URL of the web page with the corresponding image may include category information and product name information.

for example,

<https://www.OOOO.com/vp/products/1893870707?vendorItemId=71204845840&sourceType=HOME_RELATED_ADS&searchId=feed-1629680350077-related_ads&clickEventId=40e94b5c-e8ed-4124-86cc-955ec4208065&isAddedCart=>

On the page, by using the category information "OO shopping platform website > Home appliances digital > Mobile phone accessories > Selfie stick/cradle > Cradle" and the product name information "US ZJ059 desktop iPad holder tap holder portable foldable height adjustment" information In multi-shopping platform search results, it is possible to first filter products in the category of “cradle” or “cell phone accessories” and provide them as search results.

The provider of the image search device according to the prior art is a shopping platform operator, and there is a problem of providing only products within the shopping platform, and when using image search, the user must perform a search by uploading a captured or downloaded image, If the overseas product is the lowest price among the searched products, there is a problem in that the actual purchase is reluctant because the payment method and delivery are complicated.

According to the multi-platform cheapest product search assistance system according to another embodiment of the present invention, by clicking a button overlaid on an image on all websites using a browser plug-in, similar product results on multi-platforms are obtained. , simplifies (simplifies) the product results of multi-platform through image search in terms of user experience, and combines purchasing agents to support easy purchase of products without complicated processes, even if they are overseas products.

Claims (5)

an input unit for receiving review data on an analysis target product on an online sales page;
a memory storing a program for generating merit keyword information and demerit keyword information for the product to be analyzed using the review data; and
A processor for executing the program;
The processor analyzes the review data to classify it into positive review information and negative review information, analyzes the positive review information to generate the merit keyword information, and analyzes the negative review information to generate the demerit keyword information.
Keyword Extraction Device for Advantages and Disadvantages Based on In-Product Reviews.
According to claim 1,
The processor analyzes the review data and classifies it into the positive review information and the negative review information based on a predetermined review score criterion.
Keyword Extraction Device for Advantages and Disadvantages Based on In-Product Reviews.
According to claim 1,
The processor extracts keywords by analyzing the frequency of occurrence and the relationship between words in the review data composed of sentences, and generating the strength keyword information and the weakness keyword information
Keyword extraction device based on in-person product reviews.
According to claim 3,
The processor decomposes the extracted keyword into morpheme units to generate a basic keyword, and based on this, generates the advantage keyword information and the disadvantage keyword information.
Keyword Extraction Device for Advantages and Disadvantages Based on In-Product Reviews.
According to claim 3,
The processor generates and provides the merit keyword information and the demerit keyword information in words, and provides them in a predetermined number of words.
Keyword Extraction Device for Advantages and Disadvantages Based on In-Product Reviews.

Images