Online Sales System

Abstract

The present invention may include a management server that outputs products for sale as an online sales system.

Description

Technology Field

[0001] The present invention relates to an online sales system. Background Technology

[0002] As global online sales of products become more active, sellers seeking to sell to consumers in various countries are facing significant difficulties due to differences in country-specific regulations, language, customs requirements, and cultural characteristics.

[0003] Particularly for products such as food or cosmetics, which have strict ingredient regulations or labeling requirements specific to each country, it is excessively time-consuming and costly for sellers to personally review the laws of each nation and modify or register product information accordingly.

[0004] Furthermore, handling the entire process manually—including multilingual translation, currency conversion, product localization, customs clearance, and export document preparation—highly increases the risk of errors. Since small and medium-sized sellers find it difficult to secure specialized personnel for these tasks, global sales are effectively restricted.

[0005] Meanwhile, the aforementioned background technology is technical information that the inventor possessed for the derivation of the present invention or acquired during the process of deriving the present invention, and it cannot necessarily be considered publicly known technology disclosed to the general public prior to the filing of the present invention. Prior art literature

[0006] Korean Published Patent No. 1020220101953 The problem to be solved

[0007] The objective of the present invention is to provide an online sales system that, with only a single input of product information by a seller, automatically analyzes regulatory requirements by country, recommends eligible sales countries and product packaging types, and generates export documents using information automatically converted to suit each country's language and price.

[0008] The technical problems of the present invention are not limited to those mentioned above, and other unmentioned technical problems will be clearly understood by those skilled in the art from the description below. means of solving the problem

[0009] An online sales system according to one embodiment of the present invention may include a management server that outputs products for sale.

[0010] According to one embodiment, the management server may include: an information input unit that receives and registers sales product information from a seller terminal; a regulatory analysis unit that automatically collects and analyzes regulatory information of multiple countries related to the sales product based on the sales product information registered through the information input unit; a product recommendation unit that recommends a product available for sale in at least one country and the packaging form of the product based on the regulatory information of the country analyzed by the regulatory analysis unit and the sales product information registered by the information input unit; a translation conversion unit that automatically performs translation and price conversion of the sales product information corresponding to the sales product recommended by the product recommendation unit into multiple languages; a document generation unit that collects information on customs clearance and delivery requirements of multiple countries and automatically generates export documents based on the collected information; and an interface unit that combines and outputs the product recommended by the product recommendation unit and the information translated and price converted by the translation conversion unit, and proceeds with ordering and payment.

[0011] According to one embodiment, the sales product information may include all information corresponding to the sales product, such as the product name, photograph, type, description, main ingredient, capacity, method of use, target user, country of origin, certification information, packaging type, country of intent to sell, consumer price, and export price.

[0012] According to one embodiment, the information input unit may provide the stored sales product information to the regulatory analysis unit.

[0013] According to one embodiment, the regulatory analysis unit may collect country-specific regulatory information to determine whether the relevant laws are satisfied when exporting products for sale to a country, including product ingredients, packaging methods, and labeling elements, based on a database that has been pre-learned of food regulations or cosmetic regulations for each country.

[0014] According to one embodiment, the translation conversion unit can translate translation phrases and product descriptions to suit local preferences and local culture of each product using an artificial intelligence-based machine learning algorithm.

[0015] According to one embodiment, the machine learning algorithm may refer to an artificial intelligence algorithm that generates translation results by including a natural language processing model trained based on user reviews, search keywords, and cultural attribute data, including country-specific buyer preference data.

[0016] According to one embodiment, the document generation unit may be characterized by periodically updating export and import-related laws and necessary documents by country to automatically reflect the latest information.

[0017] According to one embodiment, the document generation unit can generate export documents suitable for country-specific document formats based on language-specific product information and price information output from the translation conversion unit.

[0018] According to one embodiment, the interface unit outputs a product for sale to a consumer terminal that has created an account through the management server, and can visually output the product for sale information with the final translation and price conversion to the consumer terminal.

[0019] According to one embodiment, the interface unit receives purchase information from the consumer terminal, proceeds with and processes payment to generate a purchase information processing result, and can transmit the purchase information processing result to the consumer terminal.

[0020] According to one embodiment, the consumer terminal refers to a terminal possessed by a consumer who has created an account through the management server, and refers to a terminal capable of displaying a plurality of sales products among the sales products output through the management server, and may refer to a terminal that provides order and payment functions for at least one sales product among the plurality of sales products displayed through the display.

[0021] According to one embodiment, the product recommendation unit comprises a product preference index (P pr ), country-specific regulatory scores (C pr ), regulatory sensitivity (R pr Based on ) and weights (w1, w2), the above product preference index (P pr ) and the logarithmic function are used together, and the above country-specific regulatory score (C pr The recommendation score (S) is calculated by computing a formula designed using ) together with the trigonometric function sine. pr ) can be produced.

[0022] According to one embodiment, when the product recommendation unit recommends a product available for sale in at least one country, if the popularity of a specific product in that country is already high, the popularity is further increased or maintained, so the product preference index (P pr The above recommendation score (S) of the corresponding sales product in the corresponding country by ) pr Calculate a formula designed to be used with a logarithmic function to increase the above recommendation score (S) while preventing explosive growth and allowing it to increase gradually. pr It can be characterized by producing ).

[0023] According to one embodiment, the product recommendation unit comprises the recommendation score (S) for the product in question from regulatory information of each country collected and analyzed by the regulatory analysis unit. pr Calculate ) but, the above country-specific regulatory score (C pr ) is located in the molecule, and the regulatory sensitivity (R pr ) is placed in the denominator, so that the more regulations the country has, the higher the above recommendation score (S pr The lower the ) becomes and the fewer regulations the country has, the higher the above recommendation score (S pr Calculate a formula designed to increase the above recommendation score (S pr It can be characterized by producing ).

[0024] According to one embodiment, the product recommendation unit is the product preference index (P) used together with a logarithmic function. pr The above country-specific regulatory score (C) used with the term of ) and trigonometric functions pr ) and the above regulatory sensitivity (R pr By multiplying each term of the ratio of ) by a weight, the above product preference index (P pr ), the above country-specific regulatory scores (C pr ) and the above regulatory sensitivity (R pr Which of the following variables is weighted more importantly to determine the above recommendation score (S pr It can be characterized by producing ). Effects of the invention

[0025] According to one aspect of the present invention described above, the online sales system proposed by the present invention can automatically analyze regulatory requirements by country, recommend countries where sales are possible and product packaging types, and generate export documents with information automatically converted to suit each country's language and price, simply by the seller inputting product information once.

[0026] In addition, by integrating an online interface where consumers can verify and purchase products, the entire process for global product sales can be automated and operated efficiently.

[0027] The effects of the present invention are not limited to those mentioned above, and various effects may be included within the scope obvious to a person skilled in the art from the contents described below. Brief explanation of the drawing

[0028] FIG. 1 is a conceptual diagram of an online sales system according to one embodiment of the present invention. FIG. 2 is a conceptual diagram of a management server according to one embodiment of the present invention. Specific details for implementing the invention

[0029] The following detailed description of the invention refers to the accompanying drawings, which illustrate specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that various embodiments of the invention are different but need not be mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be implemented in other embodiments without departing from the spirit and scope of the invention in relation to one embodiment.

[0030] When it is stated that one component is "connected" or "contracted" to another component, it should be understood that while it may be directly connected or contracted to that other component, there may also be other components in between. Conversely, when it is stated that one component is "directly connected" or "directly contracted" to another component, it should be understood that there are no other components in between.

[0031] Furthermore, it should be understood that the location or arrangement of individual components within each disclosed embodiment may be changed without departing from the spirit and scope of the invention. Accordingly, the following detailed description is not intended to be taken in a limiting sense, and the scope of the invention is limited only by the appended claims, including all equivalents thereof, provided appropriately described. Similar reference numerals in the drawings refer to the same or similar functions across various aspects.

[0032] Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the drawings.

[0034] FIG. 1 is a conceptual diagram of an online sales system according to one embodiment of the present invention.

[0035] Referring to FIG. 1, an online sales system according to one embodiment of the present invention may include a management server (100), a seller terminal (300), a consumer terminal (500), and a consumer terminal (500).

[0036] The management server (100) can output a sales product according to the present invention.

[0037] The seller terminal (300) may refer to a terminal possessed by a seller who intends to generate sales revenue by posting at least one product for sale on the management server (100), and may refer to a terminal capable of providing product information for sale to the management server (100) as images and documents.

[0038] The consumer terminal (500) refers to a terminal possessed by a consumer who has created an account through the management server (100), and refers to a terminal capable of displaying multiple sales products among the sales products output through the management server (100) on a display, and may refer to a terminal that provides order and payment functions for at least one sales product among the multiple sales products displayed on the display.

[0039] The management server (100), seller terminal (300), and consumer terminal (500) may be their own servers or cloud servers for providing the service according to the present invention, or they may be a peer-to-peer (P2P) set of distributed nodes.

[0040] The management server (100) can perform one or more of the operations, storage, reference, input / output, and control functions of a general computer, and may include an artificial neural network described later based on input data.

[0041] The management server (100) may include a processor and memory. The processor may output a sales product according to the present invention and may include devices capable of performing this. The processor may execute a program or control the management server (100). Program code executed by the processor may be stored in memory. The memory may store relevant information for performing a service according to the present invention or a program for implementing a method. The memory may be volatile memory or non-volatile memory.

[0042] The management server (100) can send data to an external device or receive data from an external device using a network.

[0043] The management server (100) can train an artificial neural network and can also use an artificial neural network that has been trained. The processor can train or execute an artificial neural network stored in memory, and the memory can store an artificial neural network that has been trained. The electronic device that trains the artificial neural network and the electronic device that uses it may be the same, but they may also be separate.

[0044] Artificial intelligence is a computer system that partially implements the functions of the human brain and is capable of learning, speculating, and making judgments on its own. As learning progresses, the probability of extracting the correct answer can increase. Artificial intelligence can be composed of learning and component technologies that utilize it. The learning aspect of AI is an algorithmic technology that classifies and learns features based on input data, while the component technologies may be techniques that utilize these learning algorithms to partially implement the functions of the human brain.

[0045] Artificial intelligence is a technology that facilitates the approach to problems where multiple probabilistic answers are possible, enabling it to logically and probabilistically infer optimal cycles, methods, and plans based on input data. AI inference techniques can include evaluating input data, optimization prediction, knowledge and probability-based reasoning, and preference-based planning.

[0046] Artificial neural networks are learning algorithms in the field of machine learning that programmatically implement the connections between neurons and synapses in the brain. By creating a neural network structure through programming and then training it, artificial neural networks can acquire desired functions. Although errors may exist, they can learn from massive datasets to produce appropriate output data from input data. They have the advantage of being able to obtain output data that has yielded statistically good results and are similar to human reasoning.

[0047] The management server (100) can infer individual characteristics and interests by analyzing consumers' online behavior data, social media activities, search history, etc., using an artificial intelligence algorithm built based on big data, and may include a number of pre-trained artificial neural networks for this purpose.

[0048] The network is a high-speed backbone network of a large-scale communication network capable of high-capacity, long-distance voice and data services, and may be a next-generation wired and wireless network for providing the Internet or high-speed multimedia services.

[0049] If the network is a mobile communication network, it may be a synchronous mobile communication network or an asynchronous mobile communication network. As an example of an asynchronous mobile communication network, a WCDMA (Wideband Code Division Multiple Access) network may be cited. In this case, although not shown in the drawing, the network may include an RNC (Radio Network Controller). Meanwhile, although a WCDMA network was given as an example, it may be a 3G LTE network, a 4G network, a next-generation communication network such as 5G, or other IP-based IP networks.

[0050] The management server (100), seller terminal (300), and consumer terminal (500) may include any terminal capable of exchanging data over a network, such as a desktop computer, laptop, tablet, or smartphone.

[0051] The management server (100), seller terminal (300), and consumer terminal (500) may include one or more of the computational function, storage function, reference function, input / output function, and control function of a computer to perform the service according to the present invention.

[0052] The management server (100), the seller terminal (300), and the consumer terminal (500) may access a website or install an application to receive the service according to the present invention. The management server (100), the seller terminal (300), and the consumer terminal (500) may exchange data through the website or the application.

[0053] The network is a high-speed backbone network of a large-scale communication network capable of high-capacity, long-distance voice and data services, and may be a next-generation wired and wireless network for providing the Internet or high-speed multimedia services.

[0054] If the network is a mobile communication network, it may be a synchronous mobile communication network or an asynchronous mobile communication network. As an example of an asynchronous mobile communication network, a WCDMA (Wideband Code Division Multiple Access) network may be cited. In this case, although not shown in the drawing, the network (300) may include an RNC (Radio Network Controller). Meanwhile, although a WCDMA network was given as an example, it may be a 3G LTE network, a 4G network, a 5G network, or other next-generation communication networks, or other IP-based IP networks.

[0055] A system (1) according to one embodiment of the present invention can automate the entire process for global sales of a product by automatically analyzing regulatory requirements by country, recommending countries where sales are possible and product packaging types, generating export documents with information automatically converted to suit each country's language and price, and finally providing an integrated online interface where consumers can verify and purchase the product, all while the seller inputs product information only once.

[0057] FIG. 2 is a conceptual diagram of a management server according to one embodiment of the present invention.

[0058] Referring to FIG. 2, a management server (100) according to one embodiment of the present invention may include an information input unit (110), a regulation analysis unit (120), a product recommendation unit (130), a translation conversion unit (140), a document generation unit (150), and an interface unit (160).

[0059] The information input unit (110) can receive and register product information for sale from a seller terminal and can provide the stored product information for sale to the regulatory analysis unit (120).

[0060] The aforementioned product information may refer to all information applicable to the product, including the product name, photograph, type, description, main ingredients, volume, method of use, target users, country of origin, certification information, packaging type, countries of intent to sell, consumer price, and export price.

[0061] In addition, the information input unit (110) can not only simply store the input information but also automatically check for any missing or unusual information, for example, if an ingredient is missing or the description is too short, it can send a warning message to the seller's terminal.

[0062] In addition, the information input unit (110) may also include a function that uses a consistency verification algorithm or a sentence analyzer (NLP parser), etc., to extract and organize only the necessary information so that product information can be analyzed and understood more easily from sentences freely written through a seller terminal.

[0063] The regulatory analysis unit (120) can automatically collect and analyze multiple country-specific regulatory information related to the product for sale based on the product for sale information registered through the information input unit (110).

[0064] In addition, the regulatory analysis department (120) can collect country-specific regulatory information to determine whether the product satisfies relevant laws when exporting for sale to a country, including product ingredients, packaging methods, and labeling elements, based on a database of food or cosmetic regulations for each country that has been learned in advance.

[0065] The regulatory analysis department (120) can automatically look up and compare regulations of various countries based on the input product information, for example, it can make a judgment such as "this ingredient is okay in the United States but prohibited in Europe."

[0066] Through this process, the regulatory analysis department (120) can analyze regulatory information by country by comparing product information and pieces of each country's regulations, rather than the seller having to find the laws one by one directly.

[0067] In addition, if the laws (regulations) related to products sold overseas in a specific country are ambiguous, past cases or AI training data may be referenced to determine that it "could be problematic," and a message regarding the issue may be sent to the administrator or operator of the management server.

[0068] The product recommendation unit (130) can recommend products that can be sold in at least one country and packaging types of said products based on country-specific regulatory information analyzed by the regulatory analysis unit (120) and sales product information registered by the information input unit (110).

[0069] For example, the product recommendation unit (130) can determine whether a product popular in Korea will also be popular in Japan, and can recommend it by also considering whether there will be no legal issues in that country.

[0070] As described in the example above, by combining information regarding popularity and legal issues, a message stating "We recommend selling the product in Japan, the United States, and Thailand" can be provided to the administrator or operator of the management server or to the seller's terminal.

[0071] That is, the recommendation of the product recommendation unit (130) is not a simple guess, but can be calculated mathematically by calculating the popularity score, regulation score, etc. by country, and in this calculation, a method can be used in which products that are searched for a lot by people or have good reviews receive a higher score.

[0072] In addition, the product recommendation section (130) can reduce the number of unreasonable export attempts by lowering the score for countries with overly strict regulations.

[0073] That is, the product recommendation unit (130) is configured to determine 'which country is most suitable for selling products?' and can consider the regulatory situation of each country and the consumer preferences of that country together based on the information entered by the seller.

[0074] Meanwhile, the product recommendation unit (130) has a product preference index (P pr ), country-specific regulatory scores (C pr ), regulatory sensitivity (R pr Product preference index (P) based on ) and weights (w1, w2) pr Use ) and the logarithmic function together, and country-specific regulatory scores (C pr The recommendation score (S) is calculated by computing a formula designed using ) together with the trigonometric function sine. pr ) can be produced.

[0075] The product recommendation department (130) analyzes the marketability and regulatory environment of each country comprehensively and calculates a formula to quantify the recommendation score (S) to solve the practical problems faced by sellers who intend to sell products in the global market, as it is difficult to clearly determine which country to prioritize exporting products to because regulations and consumer reactions vary widely from country to country. pr It can calculate ). The formula for comprehensively analyzing and quantifying the marketability and regulatory environment of each country is the recommendation score (S for the relevant product) derived from the regulatory information of each country collected and analyzed by the Regulatory Analysis Department. pr Calculate ) but country-specific regulatory scores (C pr ) is located in the molecule, and the regulatory sensitivity (R pr ) is placed in the denominator, so that the more regulations a country has, the higher the recommendation score (S pr The lower the ) becomes and the fewer regulations the country has, the higher the recommendation score (S pr It can be characterized by being designed to increase the ) as described above. pr ) and regulatory sensitivity (R prThe ratio of ) can be characterized by its use with the trigonometric function sine; since it is a curve that rises and falls uniformly from 0 to 180 degrees, the calculated recommendation score (S pr ) can be smoothly adjusted. In other words, the trigonometric function sine is the country-specific regulation score (C pr ) and regulatory sensitivity (R pr It can be applied to a combination of ) to serve as an adjustment mechanism to ensure smooth score changes according to regulatory conditions, and this is characterized by being used to gradually reflect differences without excessively differentiating between countries with slightly high regulations and those with very high regulations. In addition, the formula for comprehensively analyzing and quantifying the marketability and regulatory environment of each country is the product preference index (P) used together with a logarithmic function. pr Country-specific regulatory scores (C) used with the term of ) and trigonometric functions pr ) and regulatory sensitivity (R pr By multiplying each term of the ratio of ) by a weight, the product preference index (P pr ), country-specific regulatory scores (C pr ) and regulatory sensitivity (R pr Which of the following variables is weighted more importantly to determine the recommendation score (S pr It may be characterized by calculating ). That is, the product recommendation unit (130) first collects information about the product and calculates the product preference index (P pr ) and country-specific regulatory scores (C pr ) can be derived, and regulation sensitivity (R) can be derived from information previously entered by an administrator or operator of the management server (100). pr ) and weights (w1, w2) can be derived. Afterwards, the product recommendation unit (130) can derive the product preference index (P) according to each collected variable. pr It can be calculated using a logarithmic function to ensure smooth reflection, and country-specific regulatory scores (C pr ) is regulation sensitivity (R prIt can be used with ), but can be calculated using the trigonometric function sine. Multiply each calculated term by a weight and add them to obtain the recommendation score (S pr ) can be calculated, and the product recommendation unit (130) calculates the recommendation score (S pr The higher the value of ), the higher the recommendation value of the product can be judged.

[0076] As described above, the product recommendation unit (130) comprehensively analyzes the marketability and regulatory environment of each country for each product for sale, calculates a formula for quantification, and calculates a recommendation score (S pr It can calculate ) and the calculated recommendation score (S pr You can recommend products for sale by country by listing products in order of highest value of ).

[0077] delete

[0078] Here, the product preference index (P pr ) can mean a numerical representation of how much people in each country like and are interested in the product in question.

[0079] More specifically, the product recommendation unit (130) can collect reviews of the product if the product is sold in the country, detect positive reactions from the collected reviews, and calculate the average value of positive reaction reviews relative to the total reviews to obtain a product preference index (P pr ) can be derived.

[0080] For example, if a specific product is found to have a large number of positive reviews in the country's reviews, the product preference index (P pr The value of ) can increase.

[0081] Country-specific regulatory scores (C pr ) can be used as a score value between 0 and 100, and can represent a numerical value indicating how many regulations there are when selling the product in each country.

[0082] More specifically, the product recommendation unit (130) calculates the ratio of the number of regulatory restrictions of a country to the average number of regulatory restrictions of a country collected by the regulatory analysis unit (120) to obtain a country-specific regulatory score (C pr ) can be derived.

[0083] Country-specific regulatory scores (C pr The larger the value of ), the more regulations in the country may be, making it difficult to sell, and it may represent a value that quantifies the number of legal restrictions by analyzing the export standards of food / cosmetics in each country.

[0084] Regulation sensitivity (R pr ) can be used as a dimensionless value, a real number between 0 and 1, and can represent a value indicating how much the type of product is restricted by the regulations of the country and how sensitively it is received.

[0085] More specifically, the product recommendation unit (130) is regulated by an administrator or operator managing the management server (100) with a sensitivity (R pr After receiving and storing the value of ), it can be used to calculate formulas for comprehensively analyzing and quantifying the marketability and regulatory environment of each country.

[0086] Managers or operators shall consider the regulatory sensitivity (R) of the product. pr You can set ) but enter a corresponding value depending on the category or attribute of the product, and for products where regulations can be strictly applied, enter a value close to 1, and for products where regulations can be lightly applied, enter a value close to 0.

[0087] For example, if a manager or operator determines that strict regulations must be applied to products sold as food or pharmaceuticals, the regulation sensitivity (R) is set to a value of 1. pr ) can be directly entered into the management server (100) or product recommendation unit (130).

[0088] The product preference index (P), a key variable used in the formula for comprehensively analyzing and quantifying the marketability and regulatory environment of each country mentioned above. pr ) is a numerical representation of how popular a product is in a specific country, and can be derived by calculating based on user response data such as search volume and positive review ratio.

[0089] In addition, the country-specific regulatory score (C), which is a key variable used in the formula to comprehensively analyze and quantify the marketability and regulatory environment of each country mentioned above. pr ) indicates how many legal restrictions there are when exporting products from the country and can be derived based on the number or complexity of constraints.

[0090] In addition, regulatory sensitivity (R), a key variable used in the formula to comprehensively analyze and quantify the marketability and regulatory environment of each of the aforementioned countries, pr ) is a value indicating how sensitive a product is to regulations, and for product groups with strict legal standards, such as food or pharmaceuticals, the sensitivity can be set high.

[0091] The weights (w1, w2) can be used as dimensionless values ​​and real numbers between 0 and 1, and can represent the popularity weight of the product for sale and the regulation rate weight of the product for sale, respectively, and can be derived based on values ​​directly input by an administrator or operator managing the management server (100).

[0092] More specifically, the product recommendation unit (130) can store values ​​previously entered by an administrator or operator, and can use the stored values ​​to calculate a formula for comprehensively analyzing and quantifying the marketability and regulatory environment of each country.

[0093] For example, a manager or operator comprehensively analyzes the marketability and regulatory environment of each country, calculates a formula for quantification, and obtains the recommendation score (S) of the product. pr When calculating the product preference index (P pr If ) is determined to be more important, w1 can be entered as a value greater than w2.

[0094] The product recommendation unit (130) may be characterized by using weights (w1, w2) to comprehensively analyze the marketability and regulatory environment of each country and to calculate a formula for quantification in order to prevent all products for sale and all countries from being evaluated and quantified by the same standard.

[0095] That is, the product recommendation unit (130) can set which factors to consider more important through weights (w1, w2). For example, some sellers may say they want to prioritize selling to countries with better consumer response, while others may say they want to export mainly to countries with less strict regulations. Therefore, the manager or operator can adjust the direction of recommendation score calculation by reflecting the seller's strategic choice.

[0096] Recommendation Score (S pr ) is a value calculated by a formula to comprehensively analyze and quantify the marketability and regulatory environment of each country, and can be a score assigned to each product for sale, and the higher the score, the higher the recommendation ranking of the corresponding product for sale.

[0097] Also, recommendation score (S pr ) is the product preference index (P pr In a proportional relationship with ), the product preference index (P pr If ) increases, the recommendation score (S) for the corresponding product for sale pr Although it can also increase, using a logarithmic function can prevent a rapid rise.

[0098] The formula for comprehensively analyzing and quantifying the marketability and regulatory environment of each country is the Product Preference Index (P), because the popularity of each product and whether it can be sold in that country are also important for determining which product is appropriate to recommend in which country. pr ), country-specific regulatory scores (C pr ) and regulatory sensitivity (R pr It can be characterized by the fact that the formula is designed using all of ).

[0099] delete

[0100] delete

[0101] delete

[0102] delete

[0103] delete

[0104] delete

[0105] delete

[0106] delete

[0107] delete

[0108] delete

[0109] The product preference index (P), which is one of the variables in the formula for comprehensively analyzing and quantifying the marketability and regulatory environment of each country mentioned above. pr ) can vary depending on the number of searches and reviews for products sold; as time passes and the number of searches and reviews increases, the product preference index (P pr The value of ) can also increase.

[0110] In addition, if each country's regulatory policy changes, the country-specific regulatory score (C pr ) may change, and if it is a newly registered product for sale, regulatory sensitivity (R pr The value of ) can also be changed.

[0111] As described above, the product recommendation unit (130) is a product preference index (P) that can be changed every hour. pr ), country-specific regulatory scores (C pr ), regulatory sensitivity (Rpr Based on ), the marketability and regulatory environment of each country are comprehensively analyzed at predetermined intervals (e.g., every day), and a formula for quantification is recalculated to obtain a recommendation score (S pr ) can be recalculated.

[0112] The product recommendation unit (130) can quickly and accurately determine which products to recommend in various countries by comprehensively analyzing the marketability and regulatory environment of each country and calculating a formula to quantify them. For example, if a product has fewer regulations and high popularity in the United States, the recommendation score (S pr ) can be calculated at a high level.

[0113] Conversely, in some countries, if popularity is high but there are too many regulations, the recommendation score (S pr ) can be lowered.

[0114] Accordingly, when the product recommendation department (130) makes a judgment that "it is okay to recommend this product" for each country, it comprehensively analyzes the marketability and regulatory environment of each country and calculates a recommendation score (S) calculated by a formula for quantification. pr It can be judged based on ).

[0115] The greatest benefit that the product recommendation unit (130) can obtain through a formula to comprehensively analyze and quantify the marketability and regulatory environment of each country is that it automatically recommends the most suitable export target country without the seller having to investigate country-specific regulations or make a judgment based on intuition.

[0116] This not only reduces the seller's time and costs but can also prevent risks such as violations of export regulations in advance.

[0117] Furthermore, from the perspective of the management server administrator or operator, strategic flexibility can be secured by adjusting variables and weights to respond quickly to market conditions.

[0118] The formula for comprehensively analyzing and quantifying the marketability and regulatory environment of each country can automatically collect all variable values ​​to calculate the equation, and since it includes only logarithmic functions, sine functions, and simple multiplication / addition, it can be easily implemented by an ordinary technician.

[0119] The translation conversion unit (140) can automatically translate and price conversion of the saleable products recommended by the product recommendation unit (130) and the saleable product information corresponding to the saleable products into multiple national languages.

[0120] In addition, the translation conversion unit (140) can translate the translation phrases and product descriptions to suit the local preferences and local culture of each product using an artificial intelligence-based machine learning algorithm.

[0121] Here, machine learning algorithms may refer to artificial intelligence algorithms that generate translation results, including natural language processing models trained based on user reviews, search keywords, and cultural attribute data, which include country-specific buyer preference data.

[0122] In addition, the translation conversion unit (140) can evaluate the quality of the translation so that people of that country can understand it naturally, rather than just a simple machine translation. For example, if a US consumer provides feedback such as "this sentence is awkward," the translation can be modified and re-uploaded.

[0123] Additionally, the translation conversion unit (140) may detect if culturally sensitive expressions (e.g., whitening, dieting, etc.) are problematic and translate them into other expressions.

[0124] Meanwhile, the translation conversion unit (140) has a translation accuracy (R acc ), cultural difference score (C diff ), Click count for suspected mistranslations (E click ), average clicks (E avg ) and weights(w a , w bCalculated based on ) translation accuracy score (T score ) can be produced.

[0125] The translation conversion unit (140) calculates a formula to prevent consumers from rejecting a product when the translated phrase is unnatural or the meaning is distorted, in the case of products where trust is important, such as cosmetics, health foods, and pharmaceuticals, so that consumers may reject the product. score The quality of translation can be quantified by calculating ). A formula to prevent consumers from rejecting a product if the translated phrase is unnatural or the meaning is distorted is used to ensure that the accuracy and suitability of the translation are reflected more highly as the cultural difference is smaller. This is achieved through the cultural difference score (C diff It can be characterized as being designed using ) together with the trigonometric function cosine. That is, the culture difference score (C diff When ) is 0, the value applied to the trigonometric function cosine can be derived as 1, and the culture difference score (C diff When ) is 1, the value applied to the trigonometric function cosine can be -1, so the greater the cultural difference, the higher the translation accuracy score (T score ) can approach a deduction. As mentioned above, according to the formula designed to prevent consumers from rejecting a product if the translated phrase is unnatural or the meaning is distorted, the cultural difference score (C diff ) is applied to the trigonometric function cosine, allowing for a natural reflection where larger cultural differences are reflected as a deduction factor, and smaller differences are given bonus points. In addition, to prevent consumers from turning away from a product if translated phrases are unnatural or their meaning is distorted, the formula used to evaluate consumer reactions to mistranslations is the number of clicks on translations suspected of mistranslation (E click ) and average clicks (E avgIt may be characterized by being designed to reflect the ratio of ). Accordingly, based on the number of times a product is clicked, if the number of clicks for a translation suspected of being a mistranslation of the product is higher than the average, the translation accuracy score (T score It may be reflected as a deduction in the translation accuracy score (T score It can be reflected in the form of bonus points. In other words, logarithmic functions and the trigonometric function cosine were used in the formula to prevent consumers from turning away from the product if the translated phrase is unnatural or the meaning is distorted, and the logarithmic function is used for translation accuracy (R acc It can be applied to ) to make the increase in score gradually slow down as accuracy increases, and the trigonometric function cosine is the culture difference score (C diff It can be applied to ) so that the smaller the cultural difference, the higher the score is reflected.

[0126] The logarithmic function makes it difficult for translations that have already received high scores to receive even higher scores, thereby suppressing excessive score increases and maintaining stability. Additionally, thanks to its curvilinear characteristics, the cosine function reflects changes smoothly rather than abruptly in the score, and can be adjusted to respond sensitively, particularly when cultural differences are moderate.

[0127] delete

[0128] Here, translation accuracy (R acc ) can be used as a normalized real number between 0 and 1, and can mean an accuracy indicator that evaluates how well a machine-translated sentence matches the original sentence.

[0129] For example, translation accuracy (R acc ) can be derived by utilizing tools for machine translation quality evaluation indicators such as BLEU (Bilingual Evaluation Understudy), ROUGE, and BERTScore.

[0130] In addition, translation accuracy (R acc) is a normalized real value ranging from 0 to 1, and a value closer to 1 may indicate higher translation quality.

[0131] The translation conversion unit (140) compares a predefined translated sentence (Reference Translation) with a machine translation sentence generated by a model to obtain a translation accuracy (R acc It can derive ) and can automatically derive it in real-time or at regular intervals, and may also generate a warning signal that administrator review is required if it falls below a certain level (e.g., 0.4).

[0132] Cultural Difference Score (C diff ) can refer to an indicator that quantifies how much a product description or advertising copy differs from local culture, language habits, consumer sentiment, etc.

[0133] In addition, the cultural difference score (C diff ) is a normalized real number between 0 and 1, where a value closer to 0 indicates that the cultural expression fits the language / culture of local consumers well, and a value closer to 1 indicates that the translation is far removed from local sentiment.

[0134] Additionally, the translation conversion unit (140) compares and analyzes the similarity of the machine translation result with country-specific user reviews, search keywords, local advertising text datasets, etc., using a natural language processing model (NLP) to obtain a culture difference score (C diff ) can be derived.

[0135] For example, while the word "whitening" may be sensitive in the U.S. market, it is often perceived positively in Korea; cultural distance can be calculated using NLP models (such as BERT and RoBERTa) that reflect this contextual sensitivity.

[0136] Click count for translations suspected of mistranslation (E click) may refer to the number of feedback clicks made by consumers on end consumer terminals regarding translated product descriptions or advertising text, judging that 'this phrase is awkward' or 'it is suspected to be a mistranslation.'

[0137] Also, the number of clicks on translations suspected of being mistranslated (E click ) is an integer value that can have a value greater than or equal to 0, and can be reflected in the evaluation only when a minimum number of exposures (e.g., 500 or more) is met to ensure reliability.

[0138] The translation conversion unit (140) can provide buttons such as "This phrase is strange," "Report mistranslation," and "Request more natural translation" on the web / mobile UI through the management server (100), and accumulate the number of times these are clicked to obtain the number of clicks on translations suspected of mistranslation (E click ) can be derived.

[0139] Cumulative number of clicks on translations suspected of being mistranslations (E click ) is recorded on a log collection server in real-time or on a daily basis, and can be aggregated on a daily basis and stored in the DB.

[0140] Average number of clicks (E avg ) may refer to the average number of mistranslation feedback clicks aggregated for products sold in the same period, same language region, and same category, and the number of clicks for translations suspected of being mistranslations (E click It can be used as a numerical value that serves as a standard for comparison with ).

[0141] Also, average click count (E avg ) is an integer value and can be calculated as an average value based on the entire product family.

[0142] The translation conversion unit (140) can classify similar product families based on the language, product type, country, and marketing channel to which the product being evaluated belongs, and the number of clicks on translations suspected of mistranslation (E) collected from the similar group click It can be derived by averaging ).

[0143] For example, if the average number of mistranslation clicks for 100 products in the Japanese cosmetics category over the past week is 2.4, then the average number of clicks (E avg ) can be derived as 2.4.

[0144] As mentioned above, the formula designed to prevent consumers from rejecting a product when translated phrases are unnatural or their meaning is distorted is based on a total of four main variables, including translation accuracy (R acc ) is a value indicating how well a machine-translated sentence matches the original text, and can be derived through natural language processing-based evaluation metrics such as BLEU, ROUGE, and BERTScore.

[0145] In addition, the cultural difference score (C diff ) is a numerical value indicating how unnatural or alien a translated sentence feels to the local culture, and is also a real number from 0 to 1, where a value closer to 0 may indicate an expression that fits the local culture well.

[0146] Also, the number of clicks on translations suspected of being mistranslated (E click ) can be derived by measuring the number of times consumers clicked on feedback such as "this sentence is strange" or "it is unnatural."

[0147] Also, average click count (E avg ) refers to the average mistranslation feedback figures appearing in the same language group and product family, and can be used as a reference point for comparison.

[0148] weight(w a , w b ) can refer to a value that determines which part of the formula's elements is evaluated as more important, and respectively translation accuracy (R acc Weights assigned to ), culture difference score (C diff It can mean the weight assigned to ).

[0149] Also, weights (w a , w b) is a real value between 0 and 1, and can be entered such that the sum is 1.

[0150] Also, weights (w a , w b ) can be set by receiving direct input from an administrator or operator managing the management server (100), and the administrator or operator can set country-specific or product-specific weights based on country-specific sensitivity and cultural characteristics.

[0151] Accordingly, the translation conversion unit (140) can calculate the formula and retrieve the corresponding weight values ​​based on the country or product of the product being sold and substitute them into the formula.

[0152] For example, in the case of pharmaceutical products sensitive to semantic distortion from translation, a manager or operator w a You can enter a larger value for w when marketing copy is more important than meaning distortion. b You can enter a larger value for the value.

[0153] Accordingly, when the translation conversion unit (140) calculates the translation accuracy score of a pharmaceutical product sensitive to semantic distortion, the weight (w) set for the pharmaceutical product a , w b You can retrieve the value.

[0154] That is, weights (w a , w b ) is translation accuracy (R acc ) and cultural difference score (C diff It can be used to adjust the influence of ) and to flexibly adjust which element to give more weight in translation quality evaluation.

[0155] The translation conversion unit (140) can secure flexibility to customize translation quality evaluation criteria according to the nature of the product being sold and the characteristics of the market through weight setting.

[0156] Translation accuracy score (T score) is translation accuracy (R acc Translation accuracy (R) of each product sold is in a proportional relationship with ) acc As ) increases, the translation accuracy score (T score ) can also increase.

[0157] In addition, translation accuracy score (T score ) is the culture difference score (C diff In an inverse relationship with ), if cultural differences are large, the translation accuracy score (T score ) can decrease.

[0158] In addition, translation accuracy score (T score ) is the number of clicks on translations suspected of mistranslation (E click ) and average clicks (E avg It can decrease based on the ) ratio, the number of clicks on translations suspected of mistranslation (E click The higher the ), the higher the translation accuracy score (T score ) can become smaller.

[0159] To prevent consumers from turning away from a product due to unnatural or distorted translated phrases, translation accuracy (R) is used. While higher accuracy generally indicates a better translation, translation accuracy is designed so that the perceived improvement is minimal when the accuracy is close to 100%. acc It can be characterized by being designed to suppress unnecessary and excessive weighting by using a logarithmic function to reduce the increase.

[0160] delete

[0161] delete

[0162] delete

[0163] delete

[0164] delete

[0165] delete

[0166] delete

[0167] delete

[0168] The following explains each of the aforementioned variables and formulas step by step.

[0169] First, the translation conversion unit (140) can input data that can be collected from the system and external users.

[0170] That is, the translation conversion unit (140) evaluates the machine-translated sentences of the sales product information received from the seller terminal (300) for each sales product using artificial intelligence and a tool such as BLEU, and the translation accuracy (R acc It can derive ), and through NLP analysis, identify how different the words and writing style used in the product descriptions differ from the local culture to obtain a culture difference score (C diff ) can be derived.

[0171] Additionally, the translation unit (140) aggregates the number of clicks indicating dissatisfaction with the translation from consumers of each country who have accessed the server, and the number of clicks for translations suspected of mistranslation (E click It can derive ) in real-time, and calculate the average number of feedback clicks for similar product families (same language / type) to calculate the average number of clicks (E avg ) can be derived.

[0172] Second, the translation conversion unit (140) can determine the cause and nature of the problem based on the input value, specifically by calculating a formula to prevent consumers from rejecting the product if the translated phrase is unnatural or the meaning is distorted, and a translation accuracy score (T score It can calculate ) and classify the translation quality status of each product based on this score.

[0173] That is, the translation conversion unit (140) has a translation accuracy (R acc ) can be adjusted through a logarithmic function so as not to be sensitive to high accuracy, and can be designed so that low accuracy causes a large score difference, and the culture difference score (C diff) can reflect the stability of the translation by inducing more instability when cultural differences are moderate through a sine function, and reducing the impact when localization is good or very distant, and the number of clicks on translations suspected of mistranslation (E click ) and average clicks (E avg It is possible to give a warning signal to the system by reflecting the normalized relative level of dissatisfaction based on consumer response through the ratio of ).

[0174] The translation conversion unit (140) calculates a translation accuracy score (T) by comprehensively calculating each of the aforementioned variables and functions. score The lower the value, the lower the translation quality can be judged.

[0175] Finally, the translation conversion unit (140) can perform subsequent operations automatically or manually based on the result of the identification step described above, which is a translation accuracy score (T score It may operate differently depending on the result value of ) and the characteristics of each variable.

[0176] For example, translation accuracy score (T score If ) is greater than or equal to 1, the translation quality of the product being sold is judged to be excellent, and the translation can be displayed as is.

[0177] As another example, the translation accuracy score (T score If ) is calculated as a value between 0.5 and 1, the cultural difference score (C) among the translated sentences diff The retranslation algorithm can be partially executed only on sections with high ) and can be readjusted based on the sentences clicked by consumer feedback.

[0178] In addition, translation accuracy score (T score If ) is calculated to be less than 0.5, the translation sentence of the product can be classified for administrator review, and a consumer feedback-based translation template conversion or similar product phrases can be utilized.

[0179] To prevent consumers from rejecting a product if the translated phrase is unnatural or the meaning is distorted, the formula allows each variable to be accumulated and updated over time, so the translation conversion unit (140) can evaluate the translation quality of each product for sale at a predetermined period (e.g., one week).

[0180] The translation conversion unit (140) uses formulas to prevent consumers from rejecting the product if the translated phrase is unnatural or the meaning is distorted, thereby comprehensively reflecting not only simple machine translation quality but also local cultural suitability and actual user reactions, and can adjust the importance of each element, so that a flexible policy can be established depending on the product characteristics or the country of sale.

[0181] In addition, the translation conversion unit (140) can apply an automated system for judging translation quality by using a formula to prevent consumers from rejecting the product if the translated phrase is unnatural or the meaning is distorted, thereby enabling real-time quality evaluation and automatic approval of translation information.

[0182] The effect that can be realized through a formula to prevent consumers from rejecting a product when the translated phrase is unnatural or the meaning is distorted by the translation conversion unit (140) is that the quality of the automatically translated product information can be determined numerically, so that low-quality translations can be automatically filtered out or a request for review by an administrator can be made.

[0183] In addition, if the number of clicks (Eclick) suspected of mistranslation exceeds a certain threshold, the sentence may be re-translated or replaced with a similar phrase, which can reduce consumer complaints.

[0184] The formula, designed to prevent consumers from rejecting a product if the translated text is unnatural or the meaning is distorted, can automatically collect all variable values ​​to calculate the formula; since it includes only logarithmic functions, sine functions, and simple multiplication / addition, it can be easily implemented by an ordinary technician.

[0185] The document generation unit (150) can collect information on customs clearance and delivery requirements for multiple countries and automatically generate export documents based on the collected information.

[0186] In addition, the document generation unit (150) may be characterized by periodically updating export and import related laws and necessary documents for each country to automatically reflect the latest information.

[0187] In addition, the document generation unit (150) can generate export documents that match the document format of each country based on the language-specific product information and price information output from the translation conversion unit (140), thereby performing the function of automatically generating various documents required for export.

[0188] In addition, the document generation unit (150) can automatically retrieve a list of export documents required for each country and fill them in according to product information, and can separately mark countries where the document format changes frequently and send relevant messages to the manager to pay attention.

[0189] In addition, the document generation unit (150) may notify the manager that "automatic generation is difficult" if there are too many documents or if the writing is difficult, and conversely, if it is easy, the document may be automatically completed and used immediately.

[0190] Meanwhile, the document generation unit (150) has a number of documents (N) required for export. forms ), update cycle of relevant country's laws (U gap ), translation accuracy (T acc ), accuracy of product description (R clarity ) and weights(w x , w yBased on ), calculate the document creation difficulty score (D) by performing a calculation similar to a formula to determine the document creation difficulty in advance. complex ) can be produced.

[0191] The document generation unit (150) must prepare various country-specific import and export documents to export products to foreign countries in a global e-commerce environment. Since the volume and complexity of the documents vary by country and also differ significantly depending on the type of product, the difficulty of generating documents can be determined in advance and quantified by a formula to determine criteria for determining whether automation is possible or manual intervention is required. The formula for determining the difficulty of generating documents in advance indicates that complexity increases as the number of documents increases, but the range of change should decrease once a certain number is exceeded, so the number of documents required for export (N forms ) can be characterized as being designed for use with a logarithmic function. In addition, regarding the formula for determining the difficulty of document generation in advance, complexity increases as the information becomes uncertain the longer the statutes have not been updated, but since the rate of increase should decrease after a certain period, the statute update cycle of the relevant country (U gap ) can be characterized as being designed for use with the inverse tangent function. In addition, regarding the formula for determining the difficulty of document generation in advance, since automated document generation becomes difficult as translation accuracy is low and explanations are unclear, translation accuracy (T acc ) and accuracy of product descriptions (R clarity It can be characterized by being designed with an inverse proportional structure. In addition, the formula for determining the difficulty of generating documents in advance is weight (w x , w y Flexibility can be ensured by adjusting importance according to operational objectives through ). In particular, the formula for determining the difficulty of document creation in advance is the logarithmic function for the number of documents required for export (N formsBy applying it to ), it is possible to prevent the difficulty score from increasing exponentially even if the number of documents increases beyond a certain level, and the inverse tangent function is applied to the statute update cycle of the relevant country (U gap By applying it to ), if a statute is old or older than a certain amount, uncertainty becomes saturated, allowing the increase in the score to become more gradual. Formulas used to determine the difficulty of document creation in advance utilize logarithmic and inverse tangent functions to ensure numerical realism and prevent score distortion due to extreme values. As time passes, the statute update cycle of the relevant country (U gap The value of ) may increase, and accordingly, the document creation difficulty score (D) calculated by a formula to automatically determine the document creation difficulty in advance is complex ) can also rise.

[0192] Accordingly, the update cycle must be renewed, and translation accuracy (T acc ) and accuracy of product descriptions (R clarity ) can also change in real time, and the document generation unit (150) recalculates a formula to determine the difficulty of generating the document in advance in real time to obtain a document generation difficulty score (D complex ) can be recalculated.

[0193] delete

[0194] Here, the number of documents required for export (N forms ) can use the unit 'piece' and may mean the total number of documents required to export to the country.

[0195] More specifically, the document generation unit (150) automatically counts the number of documents required for export sales of products through the export and import regulations DB of the country concerned, and the number of documents required for export (N forms ) can be derived.

[0196] Update cycle of relevant country's laws (U gap) may use the unit of 'months' and may refer to the period elapsed since the last update of a statute or regulation.

[0197] More specifically, the document generation unit (150) automatically calculates the difference between the latest statute update date and the current date and the statute update cycle (U) of the country. gap ) can be derived.

[0198] Translation accuracy (T acc ) can be used as a real number between 0 and 1, and can refer to a machine translation quality indicator measured using tools for machine translation quality evaluation indicators such as BLEU (Bilingual Evaluation Understudy), ROUGE, BERTScore, etc.

[0199] More specifically, the document generation unit (150) compares the similarity between the result of translation using artificial intelligence and the reference sentence, which is sales product information received from the seller terminal (300), to obtain a translation accuracy (T acc ) can be produced.

[0200] Accuracy of product descriptions (R clarity ) can be used as a real number between 0 and 10, and can mean clarity in sentence structure and meaning transmission in product descriptions or labels.

[0201] More specifically, the document generation unit (150) evaluates the accuracy (R) of the product description for sale through the number of grammatical errors, sentence complexity, and reading difficulty using an AI sentence analyzer. clarity ) can be derived.

[0202] As mentioned above, the formula for determining the difficulty of generating documents in advance can be calculated based on four key variables, namely the number of documents required for export (N forms ) is the total number of documents required for export, and since the difficulty of the work naturally increases as the number increases, it can be a major criterion.

[0203] In addition, the update cycle of the relevant country's laws (U gap ) is a value indicating how much time has passed since the country's statutes were last updated. The older the statutes, the less up-to-date they are, and the greater the likelihood that interpretations or documentation requirements will be unclear, which can increase the difficulty of drafting documents.

[0204] In addition, translation accuracy (T acc ) indicates how accurately the automatically translated product description matches the original text, and the accuracy of the product description for sale (R clarity ) can mean a value that measures how structured and clear the product description entered by the seller is.

[0205] weight(w x , w y ) can be set to a real value such that the sum of the two values ​​is 1, and each represents the number of documents required for export (N forms ) and the update cycle of relevant country's laws (U gap It can mean a value for setting the importance of ).

[0206] weight(w x , w y ) can be set by receiving direct input from an administrator or operator managing the management server (100), and the administrator or operator can set weights based on the number of documents to be prepared per country and the update cycle.

[0207] For example, a manager or operator has a document creation difficulty score (D complex In cases where it is determined that the influence of the number of documents to be prepared is greater when calculating ), w x If you can input a larger value for w and it is determined that the influence of frequently changing laws or regulations is greater y You can enter a larger value for the value.

[0208] That is, weights (w x , w yThe importance of each element can be adjusted according to various products and national situations, and the administrator or operator of the management server can flexibly adjust the importance to suit each situation.

[0209] Document creation difficulty score (D complex ) is the number of documents required for export (N forms In a proportional relationship with ), the more documents there are, the higher the document generation difficulty score (D) for generating documents. complex ) can be increased, but by using the logarithmic function ln, it is possible to prevent the difficulty from increasing exponentially even as the number grows.

[0210] In addition, document creation difficulty score (D complex ) is the statute update cycle of the relevant country (U gap In a proportional relationship with ), as the longer the update is, the more uncertain the information becomes, so the document generation difficulty score (D complex ) can be used, but it can be characterized by reflecting a saturation characteristic in which the increase in difficulty decreases even after a certain period of time when used with an inverse tangent function.

[0211] In addition, document creation difficulty score (D complex ) is translation accuracy (T acc In an inverse relationship with ), the more accurate the translation, the higher the document generation difficulty score (D complex ) can be lowered.

[0212] In addition, document creation difficulty score (D complex ) is the accuracy of the product description for sale (R clarity In an inverse relationship with ), the clearer the sales product information received from the seller terminal (300), the lower the difficulty of generating the document can be.

[0213] The formula for determining the difficulty of generating a document in advance is a formula designed to quantitatively calculate the difficulty in the process of generating a document by the document generation unit (150), and the resulting value is a document generation difficulty score (D complex) can indicate how difficult it is to generate documents.

[0214] That is, the document creation difficulty score (D complex The larger the value of ), the more difficult it may be to generate export documents in order to export the product to the corresponding country from the document generation unit (150).

[0215] delete

[0216] delete

[0217] delete

[0218] delete

[0219] delete

[0220] delete

[0221] delete

[0222] The document generation unit (150) can be divided into three stages: input, identification, and operation, in order to calculate a formula to determine the difficulty of generating the document in advance.

[0223] First, the document generation unit (150) can derive all data necessary for the calculation of a formula to determine the difficulty of generating the document in advance by utilizing information collected from a system or a specific server (such as a national law DB).

[0224] Second, the document generation unit (150) can interpret and quantify each element by extracting and calculating each variable based on the input data and performing preprocessing to put it into a mathematical formula.

[0225] Third, the document generation unit (150) executes a mathematical formula based on the identified value to obtain a document generation difficulty score (D complex It can produce ) and perform various automation operations based on the result.

[0226] For example, the document generation unit (150) has a document generation difficulty score (D complexIf ) is calculated to be greater than 4.0, automatic document generation is suspended, and a request can be made to the administrator to review the document for that country.

[0227] In addition, the document generation unit (150) has a document generation difficulty score (D complex If ) is calculated as a value between 2.5 and 4.0, some documents are automatically generated as templates, but verification and review by an administrator may be requested during the final review stage.

[0228] In addition, the document generation unit (150) has a document generation difficulty score (D complex If ) is calculated to be a value less than 2.5, it is determined that automatic generation and automatic submission of all documents are possible, and automatic generation and submission of all documents can be performed.

[0229] The document generation unit (150) can derive variables by automatically analyzing input product descriptions, automatic translation results, DB-based regulatory information of export target countries, etc., and calculates a formula in real time to determine the difficulty of generating documents in advance through an automation engine to obtain a document generation difficulty score (D complex ) can be produced.

[0230] The document generation unit (150) calculates the document generation difficulty score (D complex You can determine whether automation is required, and perform document generation tasks by predicting the likelihood of problems and assigning priorities.

[0231] More specifically, the document generation unit (150) calculates the document generation difficulty score (D complex If ) is above a certain threshold, a guidance message can be provided to allow an administrator or operator to manually create documents.

[0232] In addition, the document generation unit (150) may be made to process (generate documents) first if the difficulty level is high but is below a certain threshold.

[0233] That is, the document generation unit (150) can predict the difficulty of automatic document generation in advance as a score through a formula for determining the difficulty of generating documents in advance, and can use it to establish a strategy to prioritize selecting a simpler export country by comparing the difficulty levels by country.

[0234] The formula for determining the difficulty of document generation in advance can automatically collect all variable values ​​to calculate the formula, and since it includes only logarithmic functions and simple multiplication / addition, it can be easily implemented by an average technician.

[0235] The interface section (160) can combine and output information regarding products recommended by the product recommendation section (130) and translation and price conversion by the translation conversion section (140), and can proceed with ordering and payment.

[0236] Additionally, the interface unit (160) outputs sales products to a consumer terminal (500) that has created an account through the management server (100), and can visually output sales product information with the final translation and price conversion to the consumer terminal (500).

[0237] Additionally, the interface unit (160) receives purchase information from the consumer terminal (500), proceeds with and processes payment to generate a purchase information processing result, and can transmit the purchase information processing result to the consumer terminal (500).

[0238] That is, the interface section (160) can function as a screen that a consumer views through a consumer terminal, that is, an online shopping mall where a consumer can check and purchase products for sale.

[0239] In addition, the interface unit (160) can automatically reflect different languages, currencies, and cultural expressions for each country to output a screen suitable for each consumer, and for example, for Japanese consumers, explanations in Japanese and prices in yen can be displayed.

[0240] In addition, the interface unit (160) can reduce consumer confusion by displaying the phrase "Translation quality under review" if the product's translation quality is evaluated as low.

[0242] The embodiments described above are for illustrative purposes only, and those skilled in the art will understand that the embodiments described above can be easily modified into other specific forms without altering the technical concept or essential features of the embodiments described above. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single unit may be implemented in a distributed manner, and components described as distributed may likewise be implemented in a combined form.

[0244] The scope of protection sought through this specification is defined by the claims set forth below rather than by the detailed description, and should be interpreted to include all modifications or variations derived from the meaning and scope of the claims and the concept of equivalents. Explanation of the symbols

[0245] 100: Management Server

Claims

Claim 1 An online sales system comprising a management server for outputting sales products; wherein the management server comprises: an information input unit that receives and registers sales product information from a seller terminal; a regulatory analysis unit that automatically collects and analyzes multiple country-specific regulatory information related to sales products based on the sales product information registered through the information input unit; a product recommendation unit that recommends products available for sale in at least one country and packaging forms of said products based on the country-specific regulatory information analyzed by the regulatory analysis unit and the sales product information registered by the information input unit; a translation conversion unit that automatically performs translation and price conversion of the sales-available products recommended by the product recommendation unit and the sales product information corresponding to the sales-available products into multiple country-specific languages; and a document generation unit that collects information on customs clearance and delivery requirements for multiple countries and automatically generates export documents based on the collected information. and includes an interface unit that integrates and outputs products recommended by the product recommendation unit and information translated and price-converted by the translation conversion unit, and proceeds with ordering and payment; wherein the sales product information refers to all information corresponding to the sales product, including the product name, photo, type, description, main ingredients, capacity, method of use, target users, country of origin, certification information, packaging type, desired sales country, consumer price, and export price; wherein the information input unit provides the stored sales product information to the regulatory analysis unit; wherein the regulatory analysis unit collects country-specific regulatory information to determine whether the product satisfies relevant laws for export to the relevant country for sales, including product ingredients, packaging method, and labeling elements, based on a database that has been pre-learned of food or cosmetic regulations for each country; wherein the translation conversion unit translates the translation phrase and product description to match the local preferences and local culture of each product using an artificial intelligence-based machine learning algorithm; and wherein the machine learning algorithm includes user reviews including country-specific buyer preference data,It refers to an artificial intelligence algorithm that generates translation results, including a natural language processing model trained based on search keywords and cultural attribute data; the document generation unit is characterized by periodically updating export and import-related laws and necessary documents by country to automatically reflect the latest information; the document generation unit generates export documents suitable for country-specific document formats based on language-specific product information and price information output from the translation conversion unit; the interface unit outputs sales products to a consumer terminal that has created an account through the management server, and visually outputs sales product information that has been finally translated and price-converted to the consumer terminal; the interface unit receives purchase information from the consumer terminal, proceeds with and processes payment to generate a purchase information processing result, and transmits the purchase information processing result to the consumer terminal; the consumer terminal refers to a terminal possessed by a consumer that has created an account through the management server; it refers to a terminal capable of displaying multiple sales products among the sales products output through the management server; it refers to a terminal that provides order and payment functions for at least one sales product among the multiple sales products displayed through the display; and the product recommendation unit refers to a product preference index (P, pr ), country-specific regulatory scores (C pr ), regulatory sensitivity (R pr Based on ) and weights (w1, w2), the above product preference index (P pr ) and the logarithmic function are used together, and the above country-specific regulatory score (C pr The recommendation score (S) is calculated by computing a formula designed using ) together with the trigonometric function sine. pr ) calculates, and when the product recommendation unit recommends a product available for sale in at least one country, if the popularity of a specific product in that country is already high, that popularity increases further or is maintained, so the product preference index (P pr The above recommendation score (S) of the corresponding sales product in the corresponding country by ) pr Calculate a formula designed to be used with a logarithmic function to increase the above recommendation score (S) while preventing explosive growth and allowing it to increase gradually. pr Characterized by calculating ) and the product recommendation unit, wherein the recommendation score (S) for the corresponding product is derived from regulatory information of each country collected and analyzed by the regulatory analysis unit. pr Calculate ) but, the above country-specific regulatory score (C pr ) is located in the molecule, and the regulatory sensitivity (R pr ) is placed in the denominator, so that the more regulations the country has, the higher the above recommendation score (S pr The lower the ) becomes and the fewer regulations the country has, the higher the above recommendation score (S pr Calculate a formula designed to increase the above recommendation score (S pr Characterized by calculating ), and the product recommendation unit is characterized by the product preference index (P) used together with a logarithmic function. pr The above country-specific regulatory score (C) used with the term of ) and trigonometric functions pr ) and the above regulatory sensitivity (R pr By multiplying each term of the ratio of ) by a weight, the above product preference index (P pr ), the above country-specific regulatory scores (C pr ) and the above regulatory sensitivity (R pr Which of the following variables is weighted more importantly to determine the above recommendation score (S pr Characterized by calculating ), and the translation conversion unit, wherein translation accuracy (R acc ), cultural difference score (C diff ), Click count for suspected mistranslations (E click ), average clicks (E avg ) and weights(w a , w b A formula is calculated to prevent consumers from turning away from the product if the translated phrase based on ) is unnatural or the meaning is distorted, and the translation accuracy score (T score ) calculates, and the above translation conversion unit calculates a 'formula for preventing consumers from turning away from a product when translated phrases are unnatural or their meaning is distorted' to prevent situations where consumers turn away from a product when translated phrases are unnatural or their meaning is distorted, in the case of products where trust is important such as cosmetics, health foods, and pharmaceuticals, thereby calculating the translation accuracy score (T) of the product for sale. score A formula for quantifying translation quality by calculating ) and preventing consumers from rejecting the product if the translated phrase is unnatural or its meaning is distorted, wherein the accuracy and suitability of the translation are reflected more highly as the cultural difference is smaller, the cultural difference score (C diff It is characterized by being designed using ) and the trigonometric function cosine together, and the above culture difference score (C diff When ) is 0, the value applied to the trigonometric function cosine is derived as 1, and when 1, the value applied to the trigonometric function cosine is -1, so the greater the cultural difference, the higher the translation accuracy score (T score ) is characterized by being reflected to approach a deduction, and the formula to prevent consumers from turning away from the product if the above-mentioned translated phrase is unnatural or its meaning is distorted is the number of clicks on translations suspected of mistranslation (E) to evaluate consumer reactions due to mistranslation. click ) and average clicks (E avg It is characterized by being designed to reflect the ratio of ), and based on the number of clicks on a sales product, if the number of clicks on a translation suspected of being a mistranslation of the corresponding sales product is higher than the average, the translation accuracy score (T score It is reflected as a deduction in the translation accuracy score (T score It is reflected in the form of bonus points in ), and the formula to prevent consumers from turning away from the product if the above translated phrase is unnatural or its meaning is distorted uses a logarithmic function and a trigonometric function cosine, wherein the logarithmic function is the above translation accuracy (R acc It is applied to ) so that the increase in score gradually slows down as accuracy increases, and the trigonometric function cosine is the above culture difference score (C diff It is applied to ) so that the smaller the cultural difference, the higher the score is reflected; the above logarithmic function includes the feature of making it difficult for translations that have already received a high score to receive a higher score, thereby suppressing excessive score increases and maintaining stability; the cosine function includes the feature that, thanks to its curvilinear characteristics, changes are reflected smoothly in the score without abruptness, and is adjusted to respond sensitively, especially when the cultural difference is moderate; and the above translation accuracy (R acc ) is used as a normalized real number between 0 and 1 and refers to an accuracy metric that evaluates how well a machine-translated sentence matches the source sentence; it is derived using tools for machine translation quality evaluation metrics including BLEU (Bilingual Evaluation Understudy), ROUGE, and BERTScore, and the above translation accuracy (R acc ) is a normalized real value ranging from 0 to 1, and a value closer to 1 indicates higher translation quality, and the translation conversion unit compares a predefined translated sentence (Reference Translation) with a machine translation sentence generated by the model to determine the translation accuracy (R acc It is characterized by deriving ) and automatically deriving it in real-time or at regular intervals, and generating a warning signal that administrator review is required if it falls below a certain level, and the above culture difference score (C diff ) refers to an indicator that quantifies how much a product description or advertising copy differs from local culture, language habits, consumer sentiment, etc., and is a normalized real number between 0 and 1; a value closer to 0 means that the cultural expression fits the language / culture of local consumers well, while a value closer to 1 means that the translation is far removed from local sentiment. The above translation conversion unit compares and analyzes the similarity of the machine translation result with country-specific user reviews, search keywords, local advertising copy datasets, etc., using a natural language processing model (NLP) to obtain the above culture difference score (C diff Deriving ) and the number of clicks on the above suspected mistranslations (E click ) refers to the number of feedback clicks made by consumers on end-consumer terminals regarding translated product descriptions or advertising text, judging that "this phrase is awkward" or "a mistranslation is suspected," and is an integer value greater than or equal to 0, but is reflected in the evaluation only when the minimum number of exposures is met to ensure reliability, and the translation conversion unit provides buttons including "This phrase is strange," "Report mistranslation," and "Request more natural translation" on the web or mobile UI through the management server, and accumulates the number of clicks to [represent] the number of clicks on the suspected mistranslation translation (E click Deriving ), and the cumulative number of clicks on translations suspected of mistranslation (E click ) is recorded on the log collection server in real-time or on a daily basis, aggregated on a daily basis and stored in the DB, and the above average click count (E avg ) refers to the average number of mistranslation feedback clicks aggregated for products sold in the same language region and same category during the same period, and the above number of clicks for translations suspected of being mistranslations (E click It is used as a numerical value serving as a comparison standard with ), is an integer value, is calculated as an average value based on the entire product family, and the above translation conversion unit classifies similar product families based on the language, product type, country, and marketing channel to which the product under evaluation belongs, and the number of clicks on the above suspected mistranslations (E) collected from the corresponding similar group click The formula derived by averaging ) and designed to prevent consumers from rejecting the product if the translated phrase is unnatural or its meaning is distorted is based on a total of four major variables, and is a value indicating how closely the machine-translated sentence matches the original text, derived through natural language processing-based evaluation indicators such as BLEU, ROUGE, and BERTScore, which is the translation accuracy (R acc It utilizes the cultural difference score (C) as a numerical value quantifying how unnatural or alien a translated sentence feels to the local culture; it is also a real number ranging from 0 to 1, where a value closer to 0 indicates an expression that fits the local culture better. diff Utilizing ), the number of clicks on translations suspected of mistranslation (E) derived by measuring the number of times consumers clicked on feedback such as "this sentence is strange" or "it is unnatural" click Utilizing ), the average number of clicks (E) used as a comparison standard, referring to the average mistranslation feedback figures appearing within the same language group and product family. avg It utilizes ), refers to a value that determines which part of the formula's elements is evaluated as more important, and respectively represents translation accuracy (R acc Weights assigned to ), culture difference score (C diff The weight (w) assigned to ) a , w b Characterized by utilizing ), and the above weight (w a , w b ) is entered as a real value between 0 and 1 inclusive, such that the sum is 1, and is configured to be directly entered by an administrator or operator managing the management server; the corresponding weight values ​​are retrieved according to the country or product of the determined sales product and input into a formula to prevent consumers from rejecting the product if the translated phrase is unnatural or its meaning is distorted, and the weight (w a , w b ) is the above translation accuracy (R acc ) and the above culture difference score (C diff It is used as a constant that controls the influence of ) to flexibly adjust which element to give more weight in the above translation quality evaluation, and the above translation conversion unit secures flexibility to customize translation quality evaluation criteria according to the nature of the product for sale and market characteristics through weight setting, and the above translation accuracy score (T score ) is the above translation accuracy (R acc Translation accuracy (R) of each product sold is in a proportional relationship with ) acc As ) increases, it increases, and the above culture difference score (C diff In an inverse relationship with ), it decreases if cultural differences are large, and the number of clicks on the above-mentioned suspected mistranslations (E click ) and average clicks (E avg Decrease based on the ) ratio, but the number of clicks on the above suspected mistranslation translation (E click It is characterized by becoming smaller as ) increases, and the above translation accuracy score (T score ) indicates a better translation as it increases, but translation accuracy (R acc It is characterized by being extracted by a formula designed to prevent consumers from turning away from the product if the translated phrase is unnatural or its meaning is distorted, by using a logarithmic function to reduce the rate of increase and suppress unnecessary and excessive weighting; and the translation conversion unit evaluates the machine-translated sentences obtained through artificial intelligence from the sales product information input from the seller terminal for each sales product using tools such as BLEU to obtain translation accuracy (R acc ) derive ), and use NLP analysis to identify how different the words and writing style used in the product descriptions differ from the local culture, thereby determining the culture difference score (C diff ) derive, and aggregate the number of clicks indicating dissatisfaction with the translation from consumers in each country who accessed the server, and the number of clicks on translations suspected of mistranslation (E click ) derives in real-time, calculates the average feedback click count of similar product families (same language / type), and calculates the average click count (E avg ) derives, and the translation conversion unit determines the cause and nature of the problem based on the derived and input value, and calculates a formula to prevent consumers from turning away from the product if the translated phrase is unnatural or the meaning is distorted, thereby obtaining the translation accuracy score (T score Calculate ) and the above translation accuracy score (T score It is characterized by classifying the translation quality status of each product for sale based on ), and the translation conversion unit is characterized by the translation accuracy (R acc ) is adjusted through a logarithmic function so as not to be sensitive to high accuracy, and is designed so that low accuracy causes a large score difference, and the above culture difference score (C diff Reflecting the stability of the translation, ) induces more instability when cultural differences are moderate through a sine function, and reduces the impact when localization is good or very distant, and the above number of clicks on suspected mistranslations (E click ) and average clicks (E avg It is characterized by generating a warning signal to the system by reflecting the normalized relative degree of dissatisfaction based on consumer response through the ratio of ), and the translation conversion unit is characterized by a translation accuracy score (T) calculated through a formula to prevent consumers from turning away from the product if the translated phrase is unnatural or its meaning is distorted. score It is determined that the lower the value, the lower the translation quality; and the translation conversion unit performs a subsequent operation based on the result calculated through a formula to prevent consumers from rejecting the product if the translated phrase is unnatural or its meaning is distorted, wherein the translation accuracy score (T score It is characterized by operating differently depending on the result value of ) and the characteristics of each variable, and the formula for preventing consumers from rejecting a product if the translated phrase is unnatural or its meaning is distorted is calculated such that each variable accumulates and is updated over time, and the translation conversion unit evaluates the translation quality of each product for sale at predetermined intervals. The translation conversion unit uses the formula for preventing consumers from rejecting a product if the translated phrase is unnatural or its meaning is distorted to comprehensively reflect simple machine translation quality, local cultural suitability, and actual user reactions, and establishes flexible policies according to product characteristics or sales countries by adjusting the importance of each element. The translation conversion unit uses the formula for preventing consumers from rejecting a product if the translated phrase is unnatural or its meaning is distorted to enable the application of an automated system for judging translation quality, thereby applying real-time quality evaluation and automatic approval status to translation information. The translation conversion unit uses the formula for preventing consumers from rejecting a product if the translated phrase is unnatural or its meaning is distorted to numerically determine the actual quality of the automatically translated product information, so that low-quality translations are automatically filtered or An online sales system that enables administrator review requests. Claim 2 delete Claim 3 delete Claim 4 delete Claim 5 delete Claim 6 delete

Images