KR102567458B1 - Method, device and system for providing commerce platform service based on influencer and community connection using artificial intelligence model - Google Patents

Abstract

An embodiment of the present invention uses an artificial intelligence model to advertise/promote products for sale through influencers and community bulletin boards, but provide customized advertisements to consumers based on a predetermined artificial intelligence model/algorithm. It relates to a method, device, and system for providing influencer and community linkage-based commerce platform services.

Description

Method, device and system for providing commerce platform service based on influencer and community linkage using artificial intelligence model

In the following examples, influencers using artificial intelligence models advertise/promote products for sale through influencers and community bulletin boards, but provide customized advertisements to consumers based on predetermined artificial intelligence models/algorithms. and a method, apparatus, and system for providing community linkage-based commerce platform services.

As the social networking service market grows, product advertising/promotion using social networking services (SNS) such as Instagram is in the limelight.

Influencers with a large number of followers are using social networking services for personal branding, sales, and business rather than simply as a means of communication.

However, when a sales page for a product advertised/promoted is located in an open market, a predetermined fee is charged to the open market, so both the advertisement orderer and the orderer have a limitation in that the rate of return is reduced.

On the other hand, when displaying advertisements related to products to consumers, a platform that proposes advertisements based on data collected through consumers' smart devices is generally used, but some consumers use such platforms to protect personal information. There is a tendency to want to block advertising information.

In addition, when data is collected through PCs used for both personal and business purposes, the effect of advertising is reduced because products that reflect the contents searched for at work are displayed instead of products that individual consumers are interested in. there is.

In addition, in promoting energy-saving, carbon-reducing, and eco-friendly products, conventional advertising methods have limitations in that they are inappropriate due to their low advertising effectiveness/efficiency.

KR 10-2497809 B KR 10-1790312 B KR 10-2110495 B

The problem to be solved by one embodiment of the present invention is to overcome the limitations of the conventional influencer-based advertising promotion platform / technology as described above, customized only with data collected from social networking services (SNS) and community bulletin boards. By proposing advertisements, it is to provide a method, device, and system for providing influencer and community-linked commerce platform services using an artificial intelligence model that can increase the accuracy of customized advertisements and prevent unnecessary personal information from being provided to platforms, etc. .

According to one embodiment, in the method of providing a commerce platform service based on influencer and community linkage using an artificial intelligence model performed by a device, the step of recording stock information of a product to be sold; registering the product to be sold on a predetermined product page to be sold; Transmitting a first feed posting request message including a link of the product page to be sold to a first influencer terminal; posting a first posting including a link of the product page to be sold on a predetermined community bulletin board to correspond to the first feed; receiving a purchase request message from a consumer terminal based on the product page to be sold; recording sales information of the product to be sold; Calculating a fee for the product to be sold in the first influencer terminal; And based on a predetermined artificial intelligence model, marking a customized advertisement on the consumer terminal; provides a method of providing influencer and community linkage-based commerce platform services using an artificial intelligence model, including.

In addition, marking the customized advertisement on the consumer terminal may include: recording a history of access to a product page through the first feed for each unique number of the consumer terminal; extracting second influencers that are interested influencers of the consumer terminal based on the access history; Extracting first data, which is the frequency of each hashtag described in the second feeds, which are all feeds of a plurality of first influencers, and the frequency by 'alt' attribute value described in the image tags of the second feeds of the first influencers step; By merging the first data extracted for each of the first influencers, the first influencer is a 'row', the hashtag and the 'alt' attribute value of the first data are a 'column', and the first influencer is a 'row'. generating a first matrix having frequency values of 1 data as 'values'; clustering the first influencers into first clusters by K-Means clustering on the first influencers based on the first matrix; extracting a second cluster, which is a cluster to which the second influencer belongs, from among the first clusters; designating a single third influencer from among the first influencers belonging to the second cluster, excluding the second influencer; designating the first feed posted by the third influencer as a third feed; designating a first posting including a link of the product page for sale corresponding to the third feed as a second posting; and providing the third feed or the second posting as a customized advertisement to the consumer terminal.

The displaying of the customized advertisement on the consumer terminal may include: extracting words from first postings posted on a plurality of community bulletin boards, and extracting second data, which is a frequency of each word by community bulletin board; A second method in which the second data extracted for each community bulletin board is merged, the community bulletin board name is a 'row', the word of the second data is a 'column', and the frequency value of the second data is a 'value' generating a matrix; clustering the community bulletin boards into third clusters by K-Means clustering of the community bulletin boards based on the second matrix; extracting, as third data, words having a frequency greater than or equal to a predetermined ratio in each of the third clusters; generating fourth clusters by merging the third and second clusters; extracting, as fourth data, hashtags and 'alt' attribute values having a frequency equal to or greater than a predetermined ratio for each of the second clusters; generating fifth data obtained by merging the third data and the fourth data; Generating a third matrix having the fourth clusters as 'rows', words, hashtags, and 'alt' attribute values of the fifth data as 'columns', and frequency values of the fifth data as 'values' step; clustering the fourth clusters into fifth clusters by K-Means clustering of the community message boards based on the third matrix; designating a sixth cluster that is one of the fifth clusters; and posting a fourth feed posted by a fourth influencer included in the sixth cluster as a third post on a community bulletin board included in the sixth cluster.

In addition, based on the sales information, predicting the amount of demand for each item of the product to be sold; and calculating a selling price for each item of the product to be sold based on the warehousing information and the demand amount for each item, wherein the step of estimating the demand quantity for each item of the product to be sold includes: a first prediction step of generating a fourth matrix having item names of the products to be sold as 'rows' and first monthly sales volumes of the products to be sold by month in the previous year as 'values'; a second prediction step of clustering the items of the product to be sold by K-Means clustering for the first monthly sales volume of the product to be sold based on the fourth matrix; a third prediction step of calculating a first monthly average sales volume for each seventh group of products to be sold; a fourth prediction step of generating a predetermined first function by interpolating the first monthly average sales volume and calculating a first monthly rate of change by differentiating the first function; a fifth prediction step of extracting first periods and eighth clusters in which the first monthly change rate is negative, based on the first monthly rate of change; a sixth prediction step of extracting ninth clusters having a positive first monthly rate of change corresponding to the first periods from among the seventh clusters; a seventh prediction step of matching the eighth and ninth clusters one-to-one; an eighth prediction step of calculating 'the first monthly change rate of the ninth cluster compared to the first monthly change rate of the eighth cluster' at a predetermined first ratio for the matched eighth and ninth clusters; a ninth prediction step of reading a second monthly sales volume, which is the monthly sales volume of the current year of the product to be sold; a tenth prediction step of reading a first comparison target sales volume of an eighth cluster for a 'third period corresponding to the latest one month corresponding to the first period' among the second monthly sales volume; an eleventh prediction step of calculating a first expected sales volume for each product to be sold based on the first comparison target sales volume and a first ratio for a ninth cluster having a first ratio value for the eighth cluster; a twelfth prediction step of reading the second comparison target sales volume of the 'tenth cluster in which the first comparison target sales volume does not exist among the seventh clusters' for the third period; a thirteenth prediction step of calculating a second expected sales volume for each product to be sold based on the second sales volume to be compared and a first monthly rate of change; and a fourteenth prediction step of merging the first predicted sales volume and the second predicted sales volume to calculate the demand quantity for each item.

The calculating of the selling price may include: calculating a first warehousing price based on the warehousing information recorded at the time of warehousing for each product to be sold; Calculating a second stock price at the current point in time for each product to be sold; Calculating a price change rate that is a second warehousing price compared to the first warehousing price; Calculating a first expected rate of change, which is 'the first expected sales volume compared to the sales volume during the last month' for each product to be sold or 'the second expected sales volume compared to the sales volume during the last month' for each product to be sold; When the first expected rate of change is greater than or equal to the 'positive first reference rate of change': calculating a first selling price by reflecting 'a first margin rate exceeding the absolute value of the price change rate' in the first storage price; When the first expected rate of change is less than the first standard rate of change and greater than or equal to the second standard rate of change, which is a positive number: The second selling price is calculated by reflecting the “second margin rate, which is a positive number designated for each seventh group,” to the first warehousing price doing; When the first expected rate of change is greater than or equal to a 'third reference rate of change that is less than or equal to 0': calculating a third selling price by reflecting a 'third margin rate that is a negative number specified for each seventh group' to the first storage price; When the first expected change rate is greater than or equal to 'a fourth reference change rate that is less than the third reference change rate and is a negative number': stopping sales of the product to be sold; and when the first expected rate of change is less than the fourth reference rate of change: calculating a fourth selling price by reflecting the 'fourth margin rate, which is a negative number designated for each seventh group', to the first stock price. there is.

An apparatus according to an embodiment may be combined with hardware and controlled by a computer program stored in a medium to execute any one of the methods described above.

According to one embodiment, an advertisement for a product to be sold may be consigned to an influencer or posted on a predetermined community bulletin board.

In addition, by suggesting customized advertisements only with data collected from social networking services (SNS) of consumer terminals and community bulletin boards, it is possible to increase the accuracy of customized advertisements and prevent unnecessary personal information from being provided to platforms.

In addition, based on the advertisement provided to a certain consumer terminal, other related advertisements are displayed in association with each other, so that advertisements of a necessary classification (group) can be continuously provided to consumers.

In addition, it is possible to improve the effective number of impressions of an advertisement by linking an advertisement between an influencer's social networking service feed and a community bulletin board.

In addition, based on the sales history of the product, future demand can be predicted.

In addition, sales/revenues can be maximized by adjusting the sales price based on the predicted product demand.

In addition, the use of green technology can be promoted by advertising/promoting products related to energy saving, carbon reduction, and eco-friendliness through influencers and community bulletin boards.

In addition, it is possible to contribute to energy saving and carbon reduction by improving the efficiency of advertisement display itself by not providing consumers with overlapping advertisements that consumers have already checked.

1 is a diagram showing a commerce platform service providing system based on linkage with an influencer and a community using an artificial intelligence model according to an embodiment of the present invention.
2 is a flowchart illustrating a method of providing a commerce platform service based on linkage with an influencer and a community using an artificial intelligence model according to an embodiment of the present invention.
3 is a flowchart illustrating steps of displaying a customized advertisement on a consumer terminal of a method for providing a commerce platform service based on influencer and community linkage using an artificial intelligence model according to an embodiment of the present invention.
4 is a flowchart illustrating the steps of predicting the demand for each item of a product to be sold in the method of providing a commerce platform service based on influencer and community connection using an artificial intelligence model according to an embodiment of the present invention.
5 is a flowchart illustrating steps for calculating a sales price of a method for providing a commerce platform service based on influencer and community connection using an artificial intelligence model according to an embodiment of the present invention.
6 is a diagram showing a fourth matrix used in the step of predicting the amount of demand for each item of a product to be sold in the method of providing a commerce platform service based on influencer and community connection using an artificial intelligence model according to an embodiment of the present invention. .
7 is a first month for each seventh cluster calculated in the step of predicting the amount of demand for each item of a product to be sold in the method of providing a commerce platform service based on influencer and community connection using an artificial intelligence model according to an embodiment of the present invention. It shows the average sales volume.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. However, since various changes can be made to the embodiments, the scope of the patent application is not limited or limited by these embodiments. It should be understood that all changes, equivalents or substitutes to the embodiments are included within the scope of rights.

Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only, and may be modified and implemented in various forms. Therefore, the embodiments are not limited to the specific disclosed form, and the scope of the present specification includes changes, equivalents, or substitutes included in the technical spirit.

Although terms such as first or second may be used to describe various components, such terms should only be construed for the purpose of distinguishing one component from another. For example, a first element may be termed a second element, and similarly, a second element may be termed a first element.

It should be understood that when an element is referred to as being “connected” to another element, it may be directly connected or connected to the other element, but other elements may exist in the middle.

Terms used in the examples are used only for descriptive purposes and should not be construed as limiting. Singular expressions include plural expressions unless the context clearly dictates otherwise. In this specification, terms such as "include" or "have" are intended to designate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, but one or more other features It should be understood that the presence or addition of numbers, steps, operations, components, parts, or combinations thereof is not precluded.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by a person of ordinary skill in the art to which the embodiment belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and unless explicitly defined in the present application, they should not be interpreted in an ideal or excessively formal meaning. don't

In addition, in the description with reference to the accompanying drawings, the same reference numerals are given to the same components regardless of reference numerals, and overlapping descriptions thereof will be omitted. In describing the embodiment, if it is determined that a detailed description of a related known technology may unnecessarily obscure the gist of the embodiment, the detailed description will be omitted.

An artificial intelligence (AI) system is a computer system that implements human-level intelligence, and unlike existing rule-based smart systems, machines learn and judge on their own. The more AI systems are used, the higher the recognition rate and the more accurate understanding of user preferences. Conventional rule-based smart systems are gradually being replaced by deep learning-based AI systems.

Artificial intelligence technology consists of machine learning and element technologies using machine learning. Machine learning is an algorithm technology that classifies/learns the characteristics of input data by itself, and element technology is a technology that uses machine learning algorithms such as deep learning to mimic functions such as recognition and judgment of the human brain. It consists of technical fields such as understanding, inference/prediction, knowledge expression, and motion control.

The various fields where artificial intelligence technology is applied are as follows. Linguistic understanding is a technology for recognizing and applying/processing human language/characters, including natural language processing, machine translation, dialogue systems, question and answering, voice recognition/synthesis, and the like. Visual understanding is a technology for recognizing and processing objects like human vision, and includes object recognition, object tracking, image search, person recognition, scene understanding, space understanding, image improvement, and the like. Inference prediction is a technique of reasoning and predicting logically by judging information, and includes knowledge/probability-based reasoning, optimization prediction, preference-based planning, and recommendation. Knowledge expression is a technology that automatically processes human experience information into knowledge data, and includes knowledge construction (data creation/classification) and knowledge management (data utilization). Motion control is a technology for controlling the autonomous driving of a vehicle and the movement of a robot, and includes motion control (navigation, collision, driving), manipulation control (action control), and the like.

In general, in order to apply machine learning algorithms to real life, learning is performed in a trial and error method due to the nature of the basic methodology of machine learning. In particular, deep learning requires hundreds of thousands of iterations. It is impossible to execute this in an actual physical external environment, so instead, the actual physical external environment is virtually implemented on a computer and learning is performed through simulation.

According to one embodiment, in the method of providing a commerce platform service based on influencer and community linkage using an artificial intelligence model performed by the device 1, the step of recording stock information of a product to be sold (S100); Registering the product to be sold on a predetermined product page to be sold (S200); Transmitting a first feed posting request message including a link of the product page for sale to a first influencer terminal (S300); Posting a first posting including a link of the product page to be sold on a predetermined community bulletin board to correspond to the first feed (S400); Receiving a purchase request message from a consumer terminal (3) based on the product page to be sold (S500); Recording sales information of the product to be sold (S600); Calculating a fee for the product to be sold at the first influencer terminal (S700); And based on a predetermined artificial intelligence model, displaying a customized advertisement on the consumer terminal 3 (S800); a method for providing a commerce platform service based on influencer and community linkage using an artificial intelligence model, including to provide.

The apparatus 1 may be a predetermined server computer or/and a predetermined administrator's computer. The device 1 may be connected to a predetermined database server 4 to manage/process data.

The device 1 may be connected to a plurality of influencer terminals, post an advertisement feed including a link to a product page for sale, and provide a fee for the advertisement feed to the influencer.

The term 'nth influencer terminal' used in the description of the present invention may be interpreted as meaning a terminal used by the 'nth influencer'. 'Consumer terminal 3' can be interpreted as meaning a terminal used by 'consumer' in the same way.

Meanwhile, consumers may access the product page for sale through a link to the product page for sale using the consumer terminal 3 (smart device, PC, etc.).

In the step of recording the warehousing information of the product to be sold (S100); the purchase information (name, part number, quantity, order date, payment amount, delivery departure date, delivery arrival date, Wholesale) seller information, manufacturer information, product type, brand, catalog, etc.) can be recorded/stored in the database server 4 or device 1.

In the step of registering the product to be sold in a predetermined product page to be sold (S200), information such as description data, product name, and price of the product to be sold is described on a pre-built shopping mall web page or an open market product page. can be registered.

In the step (S300) of sending a first feed posting request message including a link to the product page for sale to the first influencer terminal, the first influencer terminal uses SMS, SNS message, application push notification, etc. A message requesting advertisement feed posting may be transmitted to the unser terminal.

In the step (S400) of posting a first posting including a link to the product page for sale on a predetermined community bulletin board to correspond to the first feed, the contents of the first feed uploaded by the first influencer can be copied as it is or edited by an administrator or a predetermined artificial intelligence algorithm and registered on a predetermined community bulletin board.

Receiving a purchase request message from the consumer terminal 3 based on the product page to be sold (S500); when the product to be sold, which was an advertisement target, is sold (when a consumer purchases), a purchase request is sent. and performs payment processing on the consumer terminal 3.

In the step of recording the sales information of the product to be sold (S600), by accumulating and recording information such as when, at what price, and to whom the product to be sold was sold, which product to sell is how much by utilizing this later Predetermined statistical information such as whether or not it has been sold may be generated.

In the step (S700) of calculating the fee for the product to be sold in the first influencer terminal, after a predetermined period has elapsed, based on the amount of sales/profit of the product to be sold for a predetermined period, The fee to be received by the first influencer can be calculated and settled (deposited). Alternatively, a certain amount determined at the time of the contract can be settled.

In step (S800) of displaying a customized advertisement on the consumer terminal 3 based on a predetermined artificial intelligence model, the behavior pattern of the consumer terminal 3 (mainly accessed influencer, mainly accessed community bulletin board) , a product mainly purchased, etc.), a customized advertisement for a product to be sold with a high probability of being purchased by the consumer may be displayed.

The customized advertisement may be provided in the form of a predetermined Google Admob or a banner displayed on one side of a community bulletin board.

In addition, the step of displaying a customized advertisement on the consumer terminal 3 (S800) is: Recording a history of accessing the product page through the first feed for each unique number of the consumer terminal 3 (S811) ; Extracting second influencers that are interested influencers of the consumer terminal 3 based on the access history (S812); Extracting first data, which is the frequency of each hashtag described in the second feeds, which are all feeds of a plurality of first influencers, and the frequency by 'alt' attribute value described in the image tags of the second feeds of the first influencers Step (S813); By merging the first data extracted for each of the first influencers, the first influencer is a 'row', the hashtag and the 'alt' attribute value of the first data are a 'column', and the first influencer is a 'row'. generating a first matrix having frequency values of 1 data as 'values' (S814); Clustering the first influencers into first clusters by K-Means clustering on the first influencers based on the first matrix (S815); extracting a second cluster to which the second influencer belongs from among the first clusters (S816); designating a single third influencer excluding the second influencer from among the first influencers belonging to the second cluster (S817); designating the first feed posted by the third influencer as a third feed (S818); designating a first posting including a link of the product page for sale corresponding to the third feed as a second posting (S819); and providing the third feed or the second posting to the consumer terminal 3 as a customized advertisement (S820).

In the step (S811) of recording the history of accessing the product page through the first feed for each unique number of the consumer terminal 3, based on the account or / and unique number (ID) of each consumer It records which product page each consumer accesses through which primary feed and when.

In the step (S812) of extracting second influencers who are interested influencers of the consumer terminal 3 based on the access history, based on the URL information redirected to the product page (the access history) , The first feed that causes the consumer terminal 3 to flow into the product page and the second influencer that posted the first feed are extracted.

Basically, general SNS services such as Instagram are not dependent on the community bulletin board or product page configured based on the device 1 implementing an embodiment of the present invention, so in a general way, such interest influencer data It is difficult to obtain, but based on the access history as described above, it is possible to record which influencer the product page was visited by consumers, and based on this, it is possible to determine which influencer each consumer is interested in.

Extracting first data, which is the frequency of each hashtag described in the second feeds, which are all feeds of a plurality of first influencers, and the frequency by 'alt' attribute value described in the image tags of the second feeds of the first influencers In step S813, all hashtags are extracted from the second feeds posted on the SNS service, and a frequency for each hashtag is calculated. In addition, while uploading an image to the feed, the 'alt' attribute value of the <img> tag automatically generated and described by the SNS service is extracted, and the frequency for each alt attribute value is calculated.

By merging the first data extracted for each of the first influencers, the first influencer is a 'row', the hashtag and the 'alt' attribute value of the first data are a 'column', and the first influencer is a 'row'. In the step of generating a first matrix having a frequency value of 1 data as a 'value' (S814), the first matrix is generated as described above to determine which keywords are frequently posted for each first influencer. generate

For example, if there are 200 types of hashtags and 300 types of 'alt' attribute values extracted from 100 first influencers, a first matrix having a size of 100 X 500 may be generated.

At this time, if there are overlapping items in the hash tag and the 'alt' attribute value, they can be merged and processed.

Clustering the first influencers into first clusters by K-Means clustering for the first influencers based on the first matrix (S815); Based on the frequency appearing in , the type/similarity of each first influencer is determined and classified into a plurality of first clusters.

In step (S816) of extracting a second cluster, which is a cluster to which the second influencer belongs, among the first clusters, the second influencer who has already introduced the consumer terminal 3 to the product page belongs. The second cluster, which is a cluster, is extracted. In other words, influencers of a type (group) that are determined to be of interest for each consumer are extracted.

In the step (S817) of designating a single third influencer, excluding the second influencer from among the first influencers belonging to the second cluster, the consumer terminal 3 has already been configured once. If the advertisement of the second influencer, which is likely to have been purchased through the advertisement, is displayed on the consumer terminal 3 again, the efficiency / effect of the advertisement may be reduced, so the similar type (second cluster) Designate one third influencer as an influencer.

According to one embodiment of the present invention, the use of green technology can be promoted by advertising/promoting products related to energy saving, carbon reduction, and eco-friendliness through influencers and community bulletin boards.

In addition, it is possible to contribute to energy saving and carbon reduction by not providing consumers with redundant advertisements that consumers have already checked.

In the step of designating the first feed posted by the third influencer as the third feed (S818), the first feed, which is the advertisement feed posted by the designated third influencer, is designated as the third feed. .

In step S819, designating a first posting including a link of the product page for sale corresponding to the third feed as a second posting (S819); the first posting posted on the community bulletin board in response to the third feed. as the second posting.

Finally, in the step of providing the third feed or the second post as a customized advertisement to the consumer terminal 3 (S820), the advertisement that the consumer has flowed into or / and the first advertisement that has flowed the consumer to the product page An advertisement feed (third feed) of a third influencer similar to the second influencer or a second post corresponding to it is provided as a customized advertisement for each consumer.

In conclusion, consumers will not see an advertisement feed or post about a product page that they have already been introduced to through an influencer, but instead will be able to see a similar advertisement feed or post.

And, the step of marking the customized advertisement on the consumer terminal 3 is: extracting words from the first posts posted on a plurality of community bulletin boards, and extracting second data, which is the frequency of each word by community bulletin board ( S831); A second method in which the second data extracted for each community bulletin board is merged, the community bulletin board name is a 'row', the word of the second data is a 'column', and the frequency value of the second data is a 'value' generating a matrix (S832); clustering the community bulletin boards into third clusters by K-Means clustering of the community bulletin boards based on the second matrix (S833); extracting, as third data, words having a frequency greater than or equal to a predetermined ratio in each of the third clusters (S834); generating fourth clusters obtained by merging the third and second clusters (S835); extracting a hashtag having a frequency equal to or greater than a predetermined ratio and an 'alt' attribute value for each of the second clusters as fourth data (S836); generating fifth data obtained by merging the third data and the fourth data (S837); Generating a third matrix having the fourth clusters as 'rows', words, hashtags, and 'alt' attribute values of the fifth data as 'columns', and frequency values of the fifth data as 'values' Step (S838); clustering the fourth clusters into fifth clusters by K-Means clustering for the community message boards based on the third matrix (S839); designating a sixth cluster that is one of the fifth clusters (S840); and posting a fourth feed posted by a fourth influencer included in the sixth cluster as a third post on a community bulletin board included in the sixth cluster (S841).

In step S831 of extracting words from the first posts posted on a plurality of community bulletin boards and extracting second data, which is the frequency of each word by community bulletin board, for each of the plurality of community bulletin boards included in the community site. , Calculate the word appearance frequency for each community bulletin board.

For example, the frequency of words appearing on the 'game bulletin board' and the frequency of words appearing on the 'exercise bulletin board' may be extracted and included in the second data.

A second method in which the second data extracted for each community bulletin board is merged, the community bulletin board name is a 'row', the word of the second data is a 'column', and the frequency value of the second data is a 'value' In the step of generating a matrix (S832), a second matrix is generated based on the frequency of occurrence of words for each community bulletin board.

For example, if there are 20 total community bulletin boards and 1000 words appearing in all community bulletin boards, a second matrix having a size of 20*1000 may be generated.

In the step (S833) of clustering the community bulletin boards into third clusters by K-Means clustering of the community bulletins based on the second matrix, community bulletin boards having a high word similarity between each other are selected from among the community bulletin boards. grouped into the same cluster.

In the step (S834) of extracting, as third data, words having a frequency equal to or greater than a predetermined ratio for each of the third clusters, the frequency of occurrence of words in community bulletin boards included in each of the third clusters is averaged, and each third cluster Extract words with a high frequency of star appearance.

In the step of generating fourth clusters by merging the third clusters and the second clusters (S835), the second clusters (the cluster including the second influencer) and the third cluster (the group of community bulletin boards) ) to generate fourth clusters.

For example, if the third clusters include the clusters of A0, A1, A2, ..., A30, and the second clusters include the clusters of B0, B1, B2, ..., B40, they are respectively By summing, a total of 70 fourth clusters can be generated.

extracting a hashtag having a frequency equal to or greater than a predetermined ratio and an 'alt' attribute value for each of the second clusters as fourth data (S836); generating fifth data obtained by merging the third data and the fourth data (S837); Generating a third matrix having the fourth clusters as 'rows', words, hashtags, and 'alt' attribute values of the fifth data as 'columns', and frequency values of the fifth data as 'values' In step S838, a third matrix is generated to include third data and fourth data that are occurrence words/keywords for each of the fourth clusters.

In step S839 of clustering the fourth clusters into fifth clusters by K-Means clustering of the community bulletin boards based on the third matrix, the third cluster including the community bulletin boards and the fifth clusters Classify the second cluster, which includes 2 influencers, as a cluster at the same time.

For example, a third cluster that includes A bulletin board and B bulletin board and a second cluster that includes C influencer, D influencer, and E influencer may be classified into the same fifth cluster.

In step S840 of designating a sixth cluster that is one of the fifth clusters, one of the fifth clusters for which classification has been completed is selected.

In the step (S841) of posting the fourth feed posted by the fourth influencer included in the sixth cluster as a third post on the community bulletin board included in the sixth cluster, An advertisement feed (fourth feed) posted by a fourth influencer is posted in conjunction with the community bulletin board included in the sixth cluster.

In other words, an advertisement feed posted by an influencer can be automatically uploaded to a community bulletin board with similar tendencies to that influencer.

Accordingly, consumers can access feeds and posts about products they are interested in as advertisements through SNS and community bulletin boards.

In addition, based on the sales information, predicting the amount of demand for each item of the product to be sold (S900); and calculating a sales price for each item of the product to be sold based on the warehousing information and demand quantity for each item (S1000).

In step (S900) of predicting demand quantity for each item of the product to be sold based on the sales information, the product to be sold is based on past sales data accumulated for a predetermined period of time (eg, one year or more). It is possible to predict how much demand will be sold for each item in the future.

In step (S1000) of calculating the sales price for each item of the product to be sold based on the warehousing information and the amount of demand for each item, a manager implementing an embodiment of the present invention selects the product to be sold (from a wholesaler, manufacturer, etc.) ), the selling price may be calculated based on the purchased price and the demand quantity predicted how much the product to be sold will be sold.

For example, when the storage price of a product to be sold is 1,000,000 won and the demand quantity of the product to be sold is predicted to be 1,000 units, 1,300,000 won obtained by adding a certain margin to the storage price may be calculated as the selling price. In contrast, when the demand quantity of the same product to be sold is predicted to be 3,000, 1,400,000 won, which is a higher margin attached to the warehousing price, may be calculated as the selling price.

In the step of estimating the quantity demanded by item of the product to be sold (S900), the first month, which is the monthly sales volume of the product to be sold, in the previous year, with the month as 'column' and the item name of the product to be sold as 'row'. A first prediction step (S901) of generating a fourth matrix having sales volume as a 'value'; a second prediction step (S902) of clustering the items of the product to be sold by K-Means clustering for the first monthly sales volume of the product to be sold based on the fourth matrix; a third prediction step (S903) of calculating a first monthly average sales volume for each seventh group of products to be sold; a fourth prediction step (S904) of generating a predetermined first function by interpolating the first monthly average sales volume and calculating a first monthly change rate by differentiating the first function; a fifth prediction step (S905) of extracting first periods and eighth clusters in which the first monthly change rate is negative, based on the first monthly rate of change; a sixth prediction step (S906) of extracting ninth clusters having a positive first monthly rate of change corresponding to the first periods from among the seventh clusters; a seventh prediction step (S907) of matching the eighth and ninth clusters one-to-one; An eighth prediction step of calculating 'the first monthly change rate of the ninth cluster compared to the first monthly change rate of the eighth cluster' with respect to the matched eighth and ninth clusters at a predetermined first ratio (S908); a ninth prediction step (S909) of reading a second monthly sales volume, which is the monthly sales volume of the current year of the product to be sold; A 10th prediction step (S910) of reading the first comparison target sales volume of the eighth cluster for 'a third period corresponding to the latest one month corresponding to the first period' among the second monthly sales volume; An eleventh prediction step (S911) of calculating a first expected sales volume for each product to be sold based on the first comparison target sales volume and a first ratio with respect to the ninth cluster having a first ratio value for the eighth cluster (S911) ; a twelfth prediction step (S912) of reading the second comparison target sales volume of the 'tenth cluster in which the first comparison target sales volume does not exist among the seventh clusters' for the third period; a thirteenth prediction step (S913) of calculating a second expected sales volume for each product to be sold based on the second sales volume to be compared and a first monthly rate of change; and a fourteenth prediction step ( S914 ) of merging the first predicted sales volume and the second predicted sales volume to calculate the demand quantity for each item.

A first prediction generating a fourth matrix having a month as a 'column', an item name of the product to be sold as a 'row', and a first monthly sales volume of the product to be sold by month in the previous year as a 'value'. An example of the fourth matrix generated according to step S901 is shown in FIG. 6 .

The fourth matrix may be stored in the form of a predetermined database or spreadsheet.

In a second prediction step (S902) of clustering the items of the product to be sold by K-Means clustering for the first monthly sales volume of the product to be sold based on the fourth matrix, the fourth matrix is clustered, , The products to be sold may be clustered based on the sales volume data of each product to be sold included in the fourth matrix.

For example, products A and B, which show similar trends in sales volume, may belong to the same cluster.

6 shows clusters calculated for each product to be sold (by item).

In a third prediction step (S903) of calculating the first monthly average sales volume for each seventh cluster of products to be sold, for each of the plurality of seventh clusters, monthly sales volume of products to be sold belonging to each seventh cluster ( The average sales volume for the first month may be calculated by averaging data for one year last year or average data accumulated for n years.

For example, if the January sales of products A, B, and C belonging to cluster 0 were 30, 50, and 70, respectively, the average sales in January of cluster 0 is calculated as 50, and this calculation process is performed for each cluster and It repeats each month.

In a fourth prediction step (S904) of generating a predetermined first function by interpolating the first monthly average sales volume and calculating a first monthly change rate by differentiating the first function, as shown in FIG. 7, each First functions for each seventh cluster are generated by interpolating the first monthly average sales calculated for each seventh cluster for each seventh cluster.

A graph according to the first function generated for each of the seventh clusters is shown in FIG. 7 (a).

Meanwhile, a first monthly change rate, which is a change rate of monthly sales volume, is calculated by differentiating the calculated first functions for each cluster.

In a fifth prediction step (S905) of extracting first periods and eighth clusters in which the first monthly change rate is negative, based on the first monthly change rate, the first monthly change rate is negative in a certain period (several months). , That is, the first periods in which the monthly sales volume is decreasing are extracted for each cluster, and the seventh clusters in which the first period exists are designated as the eighth clusters.

In other words, the seventh cluster in which the monthly sales volume does not decrease is not designated as the eighth cluster.

In the sixth prediction step (S906) of extracting ninth clusters having a positive first monthly rate of change corresponding to the first periods from among the seventh clusters, the first monthly rate of change for the same period as the first periods is extracted. The assignees, that is, the seventh clusters with increasing monthly sales are designated as the ninth clusters.

In other words, while the monthly sales volume decreases due to the end of the fashion of the products to be sold in the eighth cluster, the clusters in which the monthly sales volume increases due to the onset of the trend are extracted as the ninth clusters.

In the seventh prediction step of matching the eighth and ninth clusters one-to-one (S907), clusters having a correlation among the extracted eighth and ninth clusters are matched with each other.

For example, in FIG. 7 (b) in which only some of the 7th clusters shown in FIG. 7 (a) are marked, the 8th cluster marked in orange shows sales volume during the period between May and June. Conversely, since the sales volume of the ninth cluster marked in yellow green increases during the period from May to June, the orange second cluster and the yellow green ninth cluster can be matched with each other.

An eighth prediction step (S908) of calculating 'the first monthly change rate of the ninth cluster compared to the first monthly change rate of the eighth cluster' with respect to the matched eighth and ninth clusters at a predetermined first ratio; , calculates the ratio of the first monthly change rates of the matched clusters.

For example, if the first monthly rate of change (slope) of the eighth cluster is -15 and the first monthly rate of change of the ninth cluster is 3, the first rate is calculated as 3/-15 = -1/5 = -0.2 It can be.

The first ratio is a value explaining the ratio of the increase in the sales volume of the 9th cluster to the decrease in the sales volume of the 8th cluster of the matched pair. When the sales volume of the ninth cluster matched thereto is calculated to be in a moderately increasing trend, the amount of demand for the ninth cluster can be predicted based on this (first ratio).

In the ninth prediction step (S909) of reading the second monthly sales volume, which is the current year's monthly sales volume of the product to be sold, the actual sales volume of the current year (data from the point of calculation until one year ago) is calculated from the sales information for each item of the product to be sold. read it

In the 10th prediction step (S910) of reading the first comparison target sales volume of the eighth cluster for the 'third period corresponding to the last one month corresponding to the first period' of the second monthly sales volume, Among the monthly sales volume, data of this month (data from the time of calculation until one month ago, data from the previous month, or data from this month if it is the end of the month) is read and designated as the first comparison target sales volume.

An eleventh prediction step (S911) of calculating a first expected sales volume for each product to be sold based on the first comparison target sales volume and a first ratio with respect to the ninth cluster having a first ratio value for the eighth cluster (S911) In ;, the absolute value of the first ratio (of the first comparison target sales volume calculated for the eighth cluster matched with the ninth cluster for the ninth cluster, which is a cluster in which a new epidemic begins when the epidemic of a certain cluster ends) ) may be reflected to calculate the first expected sales volume.

For example, when the sales volume of the 8th cluster for the first comparison target is calculated as 500 units and the sales volume of the 9th cluster for this month is calculated as 300 units, the first ratio value of the 9th cluster to the 8th cluster is -0.7 .

In the twelfth prediction step (S912) of reading the second comparison target sales volume of the 10th cluster in which the first comparison target sales volume does not exist among the seventh clusters for the third period (S912); In order to predict the demand for the 10th cluster, which is a cluster where it is difficult to grasp the trend of the trend because there is no prevalence, among the second monthly sales volume of the 10th cluster, If it is the end of the month, this month's data) is read and designated as the second comparison target sales volume.

In a thirteenth prediction step (S913) of calculating a second expected sales volume for each product to be sold based on the second sales volume to be compared and the first monthly change rate, the change rate of sales volume in the last year is not determined because the relationship with other clusters is not determined. Based on this, the sales volume change rate of this year is estimated, and based on this, the second expected sales volume, which is the expected sales volume for the next month, is calculated.

In a 14th prediction step (S914) of merging the first expected sales volume and the second expected sales volume to calculate the demand for each item, the first expected sales volume calculated for the matched ninth clusters and the unmatched Demand amounts for all seventh clusters are calculated by merging the data of the second predicted sales volumes calculated for the tenth clusters.

In addition, the calculating of the selling price may include: calculating a first warehousing price based on the warehousing information recorded at the time of warehousing for each product to be sold (S1001); Calculating a second stock price at the current point in time for each product to be sold (S1002); Calculating a price change rate that is a second warehousing price compared to the first warehousing price (S1003); Calculating a first expected rate of change, which is 'the first expected sales volume compared to the sales volume during the last month' for each product to be sold or 'the second expected sales volume compared to the sales volume during the last month' for each product to be sold (S1004); When the first expected rate of change is greater than or equal to the 'positive first reference rate of change': Calculating a first selling price by reflecting 'a first margin rate exceeding the absolute value of the price change rate' in the first storage price (S1005); When the first expected rate of change is less than the first standard rate of change and greater than or equal to the second standard rate of change, which is a positive number: The second selling price is calculated by reflecting the “second margin rate, which is a positive number designated for each seventh group,” to the first warehousing price Step (S1006); When the first expected rate of change is greater than or equal to the 'third reference rate of change that is less than or equal to 0': Calculating a third selling price by reflecting 'the third margin rate, which is a negative number designated for each seventh group,' to the first storage price (S1007) ; When the first expected rate of change is greater than or equal to 'the fourth standard rate of change which is less than the third standard rate of change and is a negative number': stopping sales of the product to be sold (S1008); and when the first expected rate of change is less than the fourth reference rate of change: calculating a fourth selling price by reflecting the 'fourth margin rate, which is a negative number designated for each seventh group', to the first stock price (S1009). can include

In step S1001 of calculating a first warehousing price for each product to be sold based on the warehousing information recorded at warehousing, the warehousing price at the time of warehousing of the product to be sold is calculated (recorded).

In step S1002 of calculating a second storage price at the current point in time for each product to be sold, a second storage price, which is a price at which the product to be sold is currently available for storage, is calculated (recorded).

In step (S1003) of calculating a price change rate that is the second warehousing price compared to the first warehousing price, how much the current second warehousing price fluctuates compared to the first warehousing price when the product to be sold is actually warehousing. yield

For example, when the first warehousing price is 1,000,000 won and the second warehousing price is 1,200,000 won, it can be seen that a profit of 200,000 won has occurred according to the change in the warehousing price.

In step S1004 of calculating a first expected change rate, which is 'the first expected sales volume compared to the sales volume in the last month' for each product to be sold or 'the second expected sales volume compared to the sales volume for the latest one month' for each product to be sold (S1004). A first expected change rate for each product to be sold is calculated by comparing the first expected sales volume or the second expected sales volume, which is the amount of demand predicted according to the steps, with the sales volume for this month (for the last month).

The first expected rate of change calculated for each product to be sold may be regarded as an index indicating how much more or less each product to be sold will be sold in the next month.

When the first expected rate of change is equal to or greater than the 'positive first reference rate of change': Calculating a first selling price by reflecting 'a first margin rate exceeding the absolute value of the price change rate' in the first storage price (S1005); In , when the first expected rate of change is very large (when it is predicted that the product to be sold will be sold much more in the next month), the profit (margin per unit) can be maximized by raising the price.

When the first expected rate of change is greater than or equal to the 'second standard rate of change that is less than the first standard rate of change and is a positive number': The second selling price is calculated by reflecting 'the second margin rate that is a positive number designated for each first group' in the first warehousing price In the step of doing, if the first expected rate of change is not very large but is a positive number (when the product to be sold is expected to be sold a little more next month), an appropriate margin rate can be applied to maximize profits (sales volume).

When the first expected rate of change is greater than or equal to the 'third reference rate of change that is less than or equal to 0': calculating a third selling price by reflecting 'the third margin rate, which is a negative number designated for each first group,' to the first storage price; When the first expected rate of change is a negative number (when the product to be sold is expected to be sold less in the next month), an appropriate discount rate may be applied to minimize loss (bad inventory).

When the first expected rate of change is equal to or greater than the 'fourth standard rate of change that is less than the third standard rate of change and is a negative number': in the step of stopping sales of the product to be sold (S1008); In this case (when it is predicted that the product to be sold will be sold less in the next month), the sales of the product to be sold can be temporarily suspended so that the product can be sold at an appropriate price in the future according to scarcity by adjusting the amount of the product to be sold.

When the first expected rate of change is less than the fourth reference rate of change: calculating a fourth selling price by reflecting the 'fourth margin rate, which is a negative number designated for each first group', in the first stock price; Even if a product is completely out of fashion and sales are stopped, it is determined that it is difficult to compensate for losses in the future, and a high discount rate is applied so that the product to be sold can be quickly sold.

The embodiments described above may be implemented as hardware components, software components, and/or a combination of hardware components and software components. For example, the devices, methods and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate (FPGA). array), programmable logic units (PLUs), microprocessors, or any other device capable of executing and responding to instructions. A processing device may run an operating system (OS) and one or more software applications running on the operating system. A processing device may also access, store, manipulate, process, and generate data in response to execution of software. For convenience of understanding, there are cases in which one processing device is used, but those skilled in the art will understand that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that it can include. For example, a processing device may include a plurality of processors or a processor and a controller. Other processing configurations are also possible, such as parallel processors.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. Program commands recorded on the medium may be specially designed and configured for the embodiment or may be known and usable to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. - includes hardware devices specially configured to store and execute program instructions, such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of program instructions include high-level language codes that can be executed by a computer using an interpreter, as well as machine language codes such as those produced by a compiler. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

Software may include a computer program, code, instructions, or a combination of one or more of the foregoing, which configures a processing device to operate as desired or processes independently or collectively. The device can be commanded. Software and/or data may be any tangible machine, component, physical device, virtual equipment, computer storage medium or device, intended to be interpreted by or provide instructions or data to a processing device. , or may be permanently or temporarily embodied in a transmitted signal wave. Software may be distributed on networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer readable media.

As described above, although the embodiments have been described with limited drawings, those skilled in the art can apply various technical modifications and variations based on the above. For example, the described techniques may be performed in an order different from the method described, and/or components of the described system, structure, device, circuit, etc. may be combined or combined in a different form than the method described, or other components may be used. Or even if it is replaced or substituted by equivalents, appropriate results can be achieved.

Therefore, other implementations, other embodiments, and equivalents of the claims are within the scope of the following claims.

1 : device (server)
2: Influencer Terminal
3: consumer terminal
4: Database Server
S100: Step of recording the warehousing information of the product to be sold
S200: Step of registering a product to be sold on a predetermined product page to be sold
S300: Transmitting a first feed posting request message to a first influencer terminal
S400: Posting the first post on a predetermined community bulletin board
S500: Receiving a purchase request message from a consumer terminal
S600: Recording sales information
S700: A step of settling the commission for the product to be sold
S800: Displaying customized advertisements on consumer terminals based on a predetermined artificial intelligence model
S811: Recording a history of access to a product page through the first feed for each unique number of the consumer terminal
S812: Step of extracting second influencers
S813: Step of extracting first data
S814: generating a first matrix
S815: Step of clustering the first influencers into first clusters
S816: Extracting a second cluster
S817: Designating a third influencer
S818: Designating the first feed posted by the third influencer as the third feed
S819: Designating the first posting corresponding to the third feed as the second posting
S820: Step of providing a third feed or a second posting as a customized advertisement to the consumer terminal
S831: Step of extracting second data
S832: generating a second matrix
S833: Step of clustering community message boards into third clusters
S834: extracting as third data words whose frequency is equal to or greater than a predetermined ratio in each of the third clusters
S835: generating fourth clusters
S836: Extracting, as fourth data, hashtags and 'alt' attribute values whose frequencies are equal to or greater than a predetermined ratio for each second cluster
S837: Step of generating fifth data
S838: generating a third matrix
S839: Clustering the fourth clusters into fifth clusters
S840: Designating a 6th cluster
S841: Posting the fourth feed as a third post on the community bulletin board included in the sixth cluster
S900: Step of predicting demand by item
S901: first prediction step
S902: Second Prediction Step
S903: Third Prediction Step
S904: 4th Prediction Step
S905: Fifth Prediction Step
S906: 6th Prediction Step
S907: Seventh Prediction Step
S908: Eighth Prediction Step
S909: ninth prediction step
S910: 10th Prediction Step
S911: 11th Prediction Step
S912: twelfth prediction step
S913: 13th prediction step
S914: 14th prediction step
S1000: step of calculating selling price for each item
S1001: Step of calculating the first warehousing price
S1002: Step of calculating the second warehousing price
S1003: Calculating the rate of change in price
S1004: Calculating a first expected rate of change
S1005: Calculating a first selling price
S1006: Calculating a second selling price
S1007: Calculating a third selling price
S1008: step of discontinuing the sale of the product to be sold
S1009: Calculating a fourth selling price

Claims (3)

In the method of providing a commerce platform service based on influencer and community linkage using an artificial intelligence model performed by a device,
Recording the warehousing information of the product to be sold;
registering the product to be sold on a predetermined product page to be sold;
Transmitting a first feed posting request message including a link of the product page to be sold to a first influencer terminal;
posting a first posting including a link of the product page to be sold on a predetermined community bulletin board to correspond to the first feed;
receiving a purchase request message from a consumer terminal based on the product page to be sold;
recording sales information of the product to be sold;
Calculating a fee for the product to be sold in the first influencer terminal; and
Displaying a customized advertisement on the consumer terminal based on a predetermined artificial intelligence model;
The step of displaying a customized advertisement on the consumer terminal is:
recording a history of access to a product page through the first feed for each unique number of the consumer terminal;
extracting second influencers that are interested influencers of the consumer terminal based on the access history;
Extracting first data, which is the frequency of each hashtag described in the second feeds, which are all feeds of a plurality of first influencers, and the frequency by 'alt' attribute value described in the image tags of the second feeds of the first influencers step;
By merging the first data extracted for each of the first influencers, the first influencer is a 'row', the hashtag and the 'alt' attribute value of the first data are a 'column', and the first influencer is a 'row'. generating a first matrix having frequency values of 1 data as 'values';
clustering the first influencers into first clusters by K-Means clustering on the first influencers based on the first matrix;
extracting a second cluster, which is a cluster to which the second influencer belongs, from among the first clusters;
designating a single third influencer from among the first influencers belonging to the second cluster, excluding the second influencer;
designating the first feed posted by the third influencer as a third feed;
designating a first posting including a link of the product page for sale corresponding to the third feed as a second posting; and
Providing the third feed or second posting as a customized advertisement to the consumer terminal; including,
Method of providing commerce platform service based on influencer and community connection using artificial intelligence model delete The method of claim 1,
The step of displaying a customized advertisement on the consumer terminal is:
extracting words from first postings posted on a plurality of community bulletin boards, and extracting second data, which is a frequency of each word by community bulletin board;
A second method in which the second data extracted for each community bulletin board is merged, the community bulletin board name is a 'row', the word of the second data is a 'column', and the frequency value of the second data is a 'value' generating a matrix;
clustering the community bulletin boards into third clusters by K-Means clustering of the community bulletin boards based on the second matrix;
extracting, as third data, words having a frequency greater than or equal to a predetermined ratio in each of the third clusters;
generating fourth clusters by merging the third and second clusters;
extracting, as fourth data, hashtags and 'alt' attribute values having a frequency equal to or greater than a predetermined ratio for each of the second clusters;
generating fifth data obtained by merging the third data and the fourth data;
Generating a third matrix having the fourth clusters as 'rows', words, hashtags, and 'alt' attribute values of the fifth data as 'columns', and frequency values of the fifth data as 'values'step;
clustering the fourth clusters into fifth clusters by K-Means clustering of the community message boards based on the third matrix;
designating a sixth cluster that is one of the fifth clusters; and
Posting a fourth feed posted by a fourth influencer included in the sixth cluster as a third post on a community bulletin board included in the sixth cluster;
Method of providing commerce platform service based on influencer and community connection using artificial intelligence model