WO2020222623A2

WO2020222623A2 - System and method for automatically constructing content for strategic sales

Info

Publication number: WO2020222623A2
Application number: PCT/KR2020/005915
Authority: WO
Inventors: 최윤내; 최치훈; 김민구
Original assignee: 주식회사 옷딜
Priority date: 2019-05-02
Filing date: 2020-05-04
Publication date: 2020-11-05
Also published as: KR20200127520A; WO2020222623A3; WO2020222623A9; KR102309373B1

Abstract

The present invention relates to a system and a method for automatically constructing content for strategic sales and, more specifically, is characterized by automatically generating content for strategic sales on the basis of the shopping mall of another business of the same kind, or data present on an SNS.

Description

System and method for automatically building contents for planning and sales

The present invention relates to a system and method for automatically constructing content for plan sales, and more particularly, to automatically generate content for plan sales based on data existing in shopping malls or SNS of other similar companies. It is characterized.

In general, the contents for sale of online shopping malls in which e-commerce transactions are performed using a communication network such as the Internet are planned and sold for products that consumers want to purchase by corresponding period, product category, and fashion. The content is displayed in the form of a banner on the screen by the user's selection.

The content for planning and sale as described above may be created as a web page displayed on the web through the Internet, or may be implemented as an application program (so-called'app') executed on a mobile communication terminal such as a user's mobile phone.

These planned and sold contents are typically used most effectively in shopping malls for clothing sales. For example,'spring clothing','summer clothing','fall clothing','winter clothing', and'S/S(Spring/Summer) ) Banners are produced in'Product Exhibition','F/W (Fall/Winter) Product Exhibition','Early Seasonal Clothing', etc. by category,'Outdoor Product Exhibition','Men's Discount Exhibition','Women's Clothing Discount Exhibition','Pants Product Exhibition', etc. The banner is produced by fashion, and by fashion, the banner is produced as'Fashionable Skirt Planning Exhibition' and'ㅇㅇ Pattern Clothing Discount Exhibition'. When the consumer selects the banner on the main screen of the shopping mall, the contents for the planned sale are displayed and the consumer Make the purchase possible.

Conventionally, for such planned sales, a person in charge for planning sales of a shopping mall analyzes and researches popular products by period, category or fashion to select popular products, select and process images of selected popular products. , Display and display the image-processed product in the order of popularity in the contents for planning and sale, and if a purchase is made for the displayed product, the purchase result is fed back and reflected in the analysis survey of the product, and whether to continue planning sales You decide.

However, the conventional procedure for planning sales made as described above is limited in the number of reflected products, because the corresponding personnel manually conducts product analysis and investigation, and ideas and images of planning sales and production of products There was a problem that it took a lot of time to select and display the product, and for this reason, it was not possible to frequently carry out planned sales in shopping malls, and there was a problem that the process of reflecting the purchase results of the planned sales as feedback was slow.

On the other hand, as described above, as a banner production method for e-commerce on the Internet, the automatic advertisement banner production method and advertisement banner automatic production server technology of Korean Patent Publication No. 10-2016-95662, and Korean Patent Publication No. 10-2000- The technology of automatic homepage setup and advertisement banner creation method linked to the free web hosting service of No. 7069 and the operation method related thereto is known.

The present invention is invented to solve the problems of the prior art as described above, and automatically creates contents for planning and sales that can automatically build contents for planning and sale by a system that has been manually performed by conventional personnel. Provides a system and its method to build with.

The system and method for automatically building contents for plan sales of the present invention to achieve the above technical problems, automatically generate contents for plan sales based on data existing in shopping malls or SNS of other similar companies Characterized in that.

The effect of the system and method for automatically constructing contents for plan sales of the present invention having the above configuration is, by automatically generating contents for plan sales, so that the person in charge of the conventional shopping mall manually for plan sales Since the process of analyzing, selecting and displaying products is omitted, shopping mall companies can quickly perform planned sales by time, category, or fashion.

In addition, since the planned sales, which were subjectively managed by the shopping mall manager in the prior art, are quantified using data based on other shopping malls or personal SNS, and planned sales are made, objective plan sales that can satisfy a plurality of customers can be made.

1 is a block diagram of a system for automatically building contents for planning and sales of the present invention;

2 and 3 are examples of a web page screen showing a hit rate test for extracting a shopping trend search word in the system of the present invention;

4 is an example of a web page screen showing a hit rate test for collecting images before planning of the system of the present invention;

5 is a flow chart of a method of automatically constructing content for plan sales of the present invention;

6 is a screen of an example showing a special exhibition created for planning sales of the present invention.

Hereinafter, a configuration of a system and method for automatically constructing content for planning and sale of the present invention will be described with reference to the drawings.

However, the disclosed drawings are provided as an example for sufficiently transferring the spirit of the present invention to those skilled in the art. Accordingly, the present invention is not limited to the drawings presented below and may be embodied in other aspects.

In addition, unless otherwise defined in terms of the terms used in the present specification, they have the meanings commonly understood by those of ordinary skill in the art to which the present invention belongs, and the gist of the present invention in the following description and accompanying drawings Detailed descriptions of known functions and configurations that may be unnecessarily obscure will be omitted.

1 is a block diagram of a system for automatically building contents for planning and selling according to the present invention.

The system (1) of the present invention is a content for planning sales that is generated based on data in an online shopping mall selling a specific product through a communication network such as the Internet or other similar company shopping malls or personal SNS (hereinafter, the present invention In the specification, for convenience of explanation, the content for planning and sale is referred to as'planned exhibition'), thereby displaying products that meet the user's wishes at a quick and low cost, and displays the products that meet the user's wishes. As purchasing actions can be made more actively, it can contribute to sales growth and promotion of shopping malls.

The content for planning and sale as described above may be generated as a web page displayed on the web through the Internet, or may be implemented as an application program (so-called'app') executed in a mobile communication terminal such as a user's mobile phone.

In addition, the planning exhibition construction system of the present invention can provide an objectively uniform image and product compared to a conventional exhibition created by a person in charge of a shopping mall in which individual and irregular products are selected.

Hereinafter, the system 1 of the present invention will be described in detail in a modified manner with reference to the drawings.

1) Trend extraction unit (10)

The trend extraction unit 10 accesses a shopping mall website of a similar company or an SNS (Social Network Service) webpage of an individual or group using a communication network (N) such as the Internet, and a keyword of a product to be sold through a special exhibition. Extract.

In order to extract the search word of such a product, preferably, the trend extraction unit 10 utilizes a known crawling technique to extract data by importing the same company shopping mall website or SNS web page of an individual or group as it is. do.

Then, the extracted search words are ranked and sorted from the largest search word extracted the most number of times to the lowest search word extracted the least number of times. The number of rankings can be set.

For example, as shown in FIG. 2, as a result of extracting a search word using a crawling technique for a clothing shopping mall or an SNS web page of an individual or group related to clothing, the word extracted the most number of times is searched from the first place in descending order. The top 10 is long dress, pants, dress, cardigan, blouse, jacket, skirt, knit, sweatshirt, and T-shirt.

In this case, the trend extracting unit 10 may store a keyword list including synonyms for a corresponding product in a database for more precise search word extraction, and recognize a word corresponding to the keyword as indicating a corresponding product.

For example, various keywords can be set for skirt products in the form of Korean, English, Korean and English mixed, English abbreviations, and Korean synonyms as follows.

* (Product Name): (Keyword List)

* Skirt: skirt, skirt, SK

* Frill skirt: frillskirt, frillsk, frill sk, frill skirt, frill skirt

* Knit skirt: knitskirt, knitsk, knit sk, knit skirt, knit skirt

* Long skirt: long skirt, longsk, long sk, long skirt, long skirt

* Midi skirt: midiskirt, midisk, midi sk, midi skirt, midi skirt

* Mini skirt: miniskirt, minisk, mini sk, mini skirt, mini skirt * Leather skirt: leatherskirt, leathersk, leather sk, leather skirt, leather skirt,

Leather sk, leather skirt, leather skirt, leather skirt * Stripe skirt: stripeskirt, stripesk, stripe sk, stripe skirt Dangara sk, Dangara skirt, Dangara skirt, Dangara skirt

Stripe skirt

* Corduroy skirt: corduroyskirt, corduroysk, corduroysk, corduroy skirt, corduroy skirt, golden skirt, golden skirt, golden sk, golden skirt

2) Subject selection section (20)

As described above, when a search word is extracted from a shopping mall website of a similar company or an SNS web page of an individual or group by the trend extraction unit 10, the topic selection unit 20 selects a keyword for a special exhibition from the extracted search word group.

For example, as described above, in descending order of the words that appear the most as a result of extracting the search terms, long dresses, pants, dresses, cardigans, blouses, jackets, skirts, knits, sweatshirts, and T-shirts are listed in descending order. In this case, the subject selection unit 20 may select'Long One Piece' as a keyword to perform a special exhibition among them, and after that, a special exhibition selling products based on the selected keyword is established.

3) Exhibition image analysis department (30)

The pre-plan image analysis unit 30 searches the image database 90 for an image to be displayed on a banner for a special exhibition on the main screen of the shopping mall based on the keyword selected by the subject selection unit 20 and analyzes it. This is the part.

Preferably, the pre-planning image analysis unit 30 is preferably performed using a machine learning technique based on artificial intelligence, and a classifier that extracts and classifies characteristic elements from raw data. It is possible to search and analyze the image to be displayed on the banner based on the result of training by repeatedly operating.

At this time, the pre-planning image analysis unit 30 collects images from shopping mall websites and related SNS web pages of the same industry through a communication network by a crawling technique when a suitable image is not found in the image database 90. Search and analyze.

For example, when the topic selection unit 10 selects'Long One Piece' as the keyword, the pre-planning image analysis unit 30 converts the image corresponding to the long dress to the image database 90 and a shopping mall website of the same industry. And related SNS web pages are collected, searched, and analyzed.

4) Planning exhibition image generation unit (40)

The pre-plan image generation unit 40 is a part that attaches appropriate phrases to the image to be displayed on the banner searched and analyzed by the pre-plan image analysis unit 30, and generates a banner image with the phrase.

For example, a phrase corresponding to the long dress is added to the image corresponding to the long dress analyzed by the pre-planning image analysis unit 30, and an example of the phrase is'It is now to buy! Good style, want to wear every day, everybody has at least one insert, a popular store, this much, what do you wear every day?, The completion of the fashion is daily, comfortable but atmospheric, every morning Codygomin?, believe in living, now What you need, good to wear tomorrow, good to wear on a sunny weekend, it's perfect for this season, I'll take responsibility for the life cut, my own style, only pretty clothes, one more time because it's pretty comfortable, mandatory check before buying clothes, hot items, actually Some examples include'I want to wear it, it is easy to use coordination, and it is easy and comfortable to styling.'

5) Planning exhibition generation unit (50)

The planning exhibition generation unit 50 is a part that defines the characteristics of the exhibition exhibition, such as the style, season, region, and coordination of the banner image generated by the exhibition image generation unit 40.

The data of the personality of the exhibition is stored in the exhibition personality database 91, and the exhibition creation unit 50 may generate a plurality of exhibitions of various personalities by using the personality data stored in the exhibition personality database 91. The generated exhibition is stored in the exhibition database 92.

For example, styles (round neck, V neck, swear neck, high neck, boat neck, off-shoulder, turtleneck, basic color, high neck) that match the banner image of the'long one piece' generated by the pre-planning image generation unit 40 Color, etc.), season (spring, summer, autumn, winter, interseason, etc.), region (mountain, sea, tourist destination, amusement park, etc.), coordination (campus, wedding, date, festival, homeware, party, company, hot place, etc.) Vacation, etc.)

6) Customized exhibition creation unit (60)

This is a part for extracting a list of exhibitions suitable for the user from the exhibitions stored in the exhibition database 92, and registering recommended products corresponding to the characteristics of the extracted exhibitions in the exhibition.

At this time, it is possible to provide a special exhibition to a user in various forms by utilizing various templates stored in the planning exhibition template 93.

In addition, the recommended product is recommended from the product recommendation unit 70 to be described later.

7) Product recommendation unit (70)

The product recommendation unit 70 is a part that searches the product database 94 for a product to be registered before the customized planning and provides it to the customized exhibition generation unit 60.

Preferably, the product recommendation unit 70 is preferably performed using a machine learning technique based on artificial intelligence, and includes a classifier for extracting and classifying characteristic elements from raw data. Iteratively operates to provide products based on training results.

At this time, the product recommendation unit 70 collects products by crawling images from shopping mall websites and related SNS web pages of the same industry through a communication network when a suitable product is not found in the product database 94. You can provide the product.

8) User experience collection unit (80)

The user experience collection unit 80 receives data from a user terminal 100 such as a mobile phone, collects user experience information (login, likes, shopping cart, purchase, product recommendation history information), and is linked to a database (not shown) This is the part to store in.

In addition, the customized exhibition creation unit 60 may use the user experience information stored in the user experience collection unit 80, and through this, it is possible to provide a more accurate customized product to the user through the exhibition.

The system for automatically building contents for planning and sales of the present invention constructed as described above was requested to the Korea Information and Communication Technology Association Software Test Certification Research Institute to test the hit rate for extracting shopping trend search words, the hit rate for collecting images before planning, and the speed for generating before planning, The results are as follows.

1. Shopping trend search word extraction hit rate

We checked the hit rate of search word extraction corresponding to the shopping trend. At this time, the hit rate for keyword extraction was calculated by the following equation 1.

(Equation 1)

Search word extraction hit rate (%)={Σ number of successes, number of measurements (10 times)×100}/number of measurements (10 times)

* Test Methods

1) Access the test target product URL (website to which the system of the present invention is applied) with a Chrome web browser.

2) Log in with a test account.

3) From the screen menu, select the'Automatic Exhibition' or'Automatic Exhibition Registration' menu and go to the'Automatic Exhibition Registration' page.

4) Click the'Search Word Extraction' button on the page of the'Automatic Exhibition Registration' content page.

5) From the list of companies' products collected to extract search terms, sort in descending order by the frequency of'extracted keywords' to check the set of the top 10'extracted keywords'.

6) If the set of the top 10'extracted keywords' identified in test procedure 5) and the list displayed in the search word list (select box) match, it is judged that the search word extraction was successful.

7) Repeat test procedure 4)-6) 10 times.

8) Calculate'Search word extraction hit rate (%)'.

* Test result

-Shopping trend keyword extraction hit rate: 100%

* Details of test results

1) Extraction keyword detailed search result: shown in FIG. 3

: Check the set of the top 10'extracted keywords' by sorting in descending order by the frequency of'extracted keywords' in the list of company products collected for search term extraction.

In Fig. 3, only the 1st to the 54th place are shown, and the top 10 extracted keywords in order of appearance frequency are'Long One Piece','Pants','One Piece','Cardigan','Blouse','Jacket','Skirt', Displayed as'knit','man-to-man', and't-shirt'.

2) Experiment result screen: shown in Fig. 4

: The set of the top 10'extracted keywords' identified in test procedure 5) and the list displayed in the search word list (select box) as shown in FIG. 4 ('long one piece','pants','one piece','cardigan', 'Blouse','Jacket','Skirt','Knit','Man-to-Man','T-Shirt') match each other, and this is judged as a successful search term extraction.

Hereinafter, although not shown, as a result of performing the above test procedure 4)-8) 9 times, the set of the top 10'extracted keywords' identified in the test procedure 5) and the list displayed in the search word list (select box) coincide with each other box.

2. Hit rate for collecting images before planning

When performing the image generation function before planning, the hit rate of image collection matching the search word was checked. At this time, the image collection hit ratio was calculated by Equation 2 below.

(Equation 2)

Image collection hit rate (%) = {Σ number of successes, number of measurements (10 times) × 100} / number of measurements (10 times)

* Test Methods

2) Log in with a test account.

5) Select the search term ('Long One Piece') item from the search term list and click the'Create Banner' button.

6) Among the images collected from the interlocking database server, it was confirmed that an image corresponding to the search word was created.

7) Click the'Related Products Link' button at the bottom of the created image.

8) If the search word ('Long One Piece') is included in the top product name on the page to be moved after clicking the'Related Product Link' button at the bottom of the created image, it is judged as a successful image collection hit.

9) Repeat test procedure 4)-8) 10 times.

10) Calculate the'image collection hit rate (%)'.

* Test result

-Hit rate for collecting images before planning: 100%

* Details of test results

1) Experiment result screen: shown in FIG. 5

: As shown in Fig. 5, an image containing the phrase of the keyword'Long One Piece' (the phrase of'Caramel Suspenders Long One Piece') is collected in the top product name from the contents displayed by clicking the'Related Product Link' button at the bottom of the image And judged this as a success of image collection.

Hereinafter, although not shown, the above test procedure 4)-8) was repeated 9 times, and as a result of performing the above test procedure 4)-8), the'Lace Layered Bending Spring Long One Piece','Chiffon Beauty Flower Long One Piece','Flower Chiffon Long' One Piece','Flora Bustier Long One Piece','Flower Frill Wave Neck Long One Piece','Pleated Ella Long One Piece','V-Neck Chiffon Frill Dot Long One Piece','(omitted)... Images containing the phrase of a long dress, such as'Strep Nitron One Piece' are collected.

3. Planning exhibition creation speed

* Test Methods

Measure the time required to import product information and create a screen before planning. At this time, the image collection hit ratio was calculated by the following equation (3).

(Equation 3)

Average time required = Σ time required / number of measurements (10 times)

-Time required (recall product information) = Time to complete product list display (-) Start time to load product

-Time required (creating the exhibition screen) = Completion time to display the exhibition screen (-) Start time of creating the exhibition

-Test environment

1) Product to be tested: Website to which the system of the present invention is applied

2) Number of repeated tests: 1-0 times

3) Registration information for automatic planning

-Event start: 2019-02-01

-Event end: 2019-02-28

-Title: Now to live! Long dress

-Content: Instem Long One Piece

-Keyword: Long One Piece

* Test procedure

2) Log in with a test account.

4) Enter the registration information (event start, event end, title, content, search word) before automatic planning.

5) Click the'Search Word Extraction' button of the'Automatic Exhibition Registration' content.

6) Select the search term ('Long One Piece') from the search term list and click the'Load Product' button to measure the time required (retrieve product information).

7) After selecting the search term ('Long One Piece'), click the'Create Banner' button.

8) Click the'Generate Exhibition' button to measure the time required (generate the exhibition screen).

9) Repeat test procedure 6)-8) 10 times,

*10) Calculate the average time required (retrieving product information) and the average time required (creating the pre-planning screen).

* Test result

-Average time required (recall product information): 74.72 seconds

-Average time required (creating the exhibition screen): 3.92 seconds

: As can be seen from the above test results, it can be seen that the system of the present invention has a very fast execution speed of product information retrieval as 74.72 seconds, and the speed of creating a plan exhibition screen based on this is very rapid as 3.92 seconds.

* Remarks: The hardware specifications of the server and client computer to which the system of the present invention used for the above project creation speed is applied are as follows.

1) server

-OS: CentOS 6.9 (64bit)

-CPU: Intel Core i7-4770K (3.50 GHz)

-Memory: 8GB

-Storage: 1TB

-Pre-Requisite: Apache Tomcat 7.0.65, MariaDB 10.1.26

2) client computer

-OS: Apple macOS Sierra 10.12.5 (64bit)

-CPU: Intel Core i7-8750H (2.20 GHz)

-Memory: 16GB

-Storage: 250GB

-Pre-Requisite: Google Chrome 72.0.3626.96

Therefore, according to the test results as described above, the system for automatically building contents for planning sales of the present invention can very quickly and accurately generate a special exhibition according to trends extracted from a shopping mall website or SNS web page in the same industry. Is showing.

Next, according to the test results as described above, a method of automatically constructing content for plan sales according to the present invention will be described. 5 is a flowchart of a method of automatically constructing content for planning sales according to the present invention, and hereinafter, each step will be described in a retrofit manner, but descriptions of the same components as those described in the system of the present invention will be omitted.

1) Trend extraction step (S1)

The trend extraction step (S1) is to access a shopping mall website of a similar company or an SNS (Social Network Service) webpage of an individual or group using a communication network (N) such as the Internet, ) Is extracted.

At this time, when extracting the search word, it is desirable to use a known crawling technique that extracts data by bringing the same company shopping mall website or SNS web page of an individual or group as it is, and extracting the extracted search word the most number of times. Since the ranking is determined and sorted from the largest number of search words to the lowest search word extracted with the least number of times, and the number of rankings can be set as described in the above-described trend extraction unit 10, detailed examples are omitted.

2) Subject selection step (S2)

In the trend extraction step, when a search word is extracted from a shopping mall website of a similar company or an SNS content of an individual or group, a keyword for a special exhibition is selected from the extracted search word group.

An example of the subject selection step (S2) is the same as described in the topic selection unit 20 described above, so detailed examples are omitted.

3) Image analysis stage before planning (S3)

Based on the keyword selected in the subject selection step (S2), an image to be displayed on a banner for a special exhibition on the main screen of the shopping mall is searched and analyzed in an image database.

At this time, the search and analysis of the image is preferably performed using a machine learning technique based on artificial intelligence, and a classifier that extracts and classifies characteristic elements from raw data is repeatedly operated. An image to be displayed on a banner is searched and analyzed based on the training result.

At this time, when a suitable image is not searched in the image database, images may be collected, searched and analyzed from shopping mall websites and related SNS web pages of the same industry through a communication network through a crawling technique.

An example of the pre-plan image analysis step is as described in the above-described pre-plan image analysis unit 30, so a detailed example will be described.

4) Pre-plan image creation step (S4)

This is a step of adding appropriate phrases to the image to be displayed on the banner searched and analyzed from the pre-planning image analysis step (S3), and generating a banner image containing the phrases.

The example of the pre-plan image generation step is the same as described in the above-described pre-plan image generation unit 40, so detailed examples are omitted.

5) Planning exhibition creation step (S5)

This is a step of defining the characteristics of the exhibition exhibition, such as the style, season, region, and coordination of the banner image generated in the exhibition image generation step S4.

The data of the personality of the exhibition is stored in a database of personality of the exhibition, and a number of exhibitions of various characteristics can be generated by using the stored personality data. The created exhibition is stored in the exhibition database.

The example of the planning exhibition generation step (S5) is the same as described in the aforementioned planning exhibition generation unit 50, so a detailed example is omitted.

6) Creation of customized exhibitions (S6)

In the pre-planning image generation step (S5), a list of exhibitions suitable for a user is extracted from the exhibitions stored in the exhibition database, and recommended products that match the characteristics of the extracted exhibitions are registered before the planning.

At this time, it is possible to provide a special exhibition to a user in various forms by utilizing various templates stored in the special exhibition template.

In addition, the recommended product is provided by searching in a product database.

Preferably, the recommended product is performed using a machine learning technique based on artificial intelligence, and learning by repeatedly operating a classifier that extracts and classifies characteristic elements from raw data. Offer products based on training results.

In this case, when a suitable product is not searched in the product database, the product may be provided by collecting the image from the shopping mall website and related SNS web page of the same industry through a communication network using a crawling technique.

In addition, data is received from a user terminal such as a mobile phone, and user experience information (login, likes, shopping cart, purchase, product recommendation history information) is collected and stored in a linked database (not shown), and the customized exhibition creation step ( In S6), the previously stored user experience information can be used. Through this, a more accurate customized product can be provided to the user through a special exhibition.

6 is a screen showing an example of a special exhibition created for planning sales according to the present invention, and shows that a special exhibition related to'Office Look' is automatically generated by the system and method of the present invention.

On the other hand, the method of the present invention made as described above allows the seller of the product to pay the advertising fee determined by the product seller whenever the consumer clicks the product for the product displayed in the created exhibition. The advertising revenue can be provided to the system operator of the present invention by setting the minimum or maximum amount of the advertising amount, and exhausting the amount of advertising as the amount set, and preventing products that have exhausted the advertising amount or have no advertising amount from being displayed before planning. have.

[Explanation of code]

* Description of the symbols for the main parts of the drawing *

10; Trend extraction unit

20; Subject Selection Department

30; Image analysis unit

40; Image generator

50; Planning exhibition generation department

60; Customized exhibition generation department

70; Product Recommendation Department

80; User experience collection unit

90; Image database

91; Rules database

92; Exhibition database

93; Template database

100; User terminal

Claims

A system for automatically building contents for planned sales, characterized in that automatically generating contents for planned sales based on data existing in a shopping mall or SNS of other similar companies.
The method of claim 1, wherein the system,

A trend extraction unit 10 for extracting a search word of a product to be sold by accessing a shopping mall website or an SNS web page of the same company using a communication network;

A topic selection unit 20 for selecting a keyword for a special exhibition from the extracted search word group when a search word is extracted by the trend extraction unit 10;

An exhibition image analysis unit 30 for searching and analyzing an image to be displayed on a banner for a special exhibition in the image database 90 based on the keyword selected by the subject selection unit 20;

An exhibition image generation unit 40 that attaches appropriate phrases to the images to be displayed on the banner searched and analyzed by the exhibition image analysis unit 30, and generates a banner image containing the words;

A planning exhibition generation unit 50 defining the characteristics of the exhibition exhibition based on the banner image generated by the exhibition planning image generation unit 40 and storing it in the exhibition exhibition database 92;

A customized exhibition creation unit 60 for extracting a list of exhibitions from the exhibitions stored in the exhibition database 92, and registering recommended products matching the characteristics of the extracted exhibitions before the exhibitions;

System characterized in that it is configured to include.
The method of claim 2, wherein the trend extraction unit 10,

A system characterized in that the search word is extracted using a crawling technique.
The method of claim 3, wherein the trend extraction unit 10,

A system characterized in that a keyword list including synonyms for a corresponding product is stored in a database, and a word corresponding to the keyword is also recognized as indicating a corresponding product.
The method of claim 2, wherein the pre-plan image analysis unit 30,

It is performed using a machine learning technique based on artificial intelligence,

When a suitable image is not found in the image database 90, the image is collected from a shopping mall website of the same industry and a related SNS web page through a communication network.
The method of claim 2,

A product recommendation unit 70 that searches for a product in the product database 94 and provides it to a plan to be created by the customized exhibition generation unit 60;

The system characterized in that it is configured to further include.
The method of claim 6, wherein the product recommendation unit (70),

It is performed using a machine learning technique based on artificial intelligence,

When a suitable product is not found in the product database 94, an image is collected from a shopping mall website of the same industry and a related SNS web page through a communication network.
The method of claim 6,

A user experience collection unit 80 that receives data from the user terminal 100, collects user experience information, stores it in an interlocked database, and provides the stored user experience information to the customized exhibition generation unit 60; The system characterized in that it is configured to further include.
In order to automatically build content for planned sales,

A trend extraction step (S1) of extracting a search word of a product to be sold by accessing a shopping mall website or an SNS web page of a similar company using a communication network;

When a search word is extracted in the trend extraction step (S1), a topic selection step (S2) of selecting a keyword for a special exhibition from the extracted search word group;

Pre-plan image analysis step (S3) of searching an image database for an image to be displayed on a banner for a special exhibition based on the keyword selected in the subject selection step (S2);

Pre-plan image generation step (S4) of adding appropriate phrases to the image to be displayed on the banner searched and analyzed in the pre-planning image analysis step (S3), and generating a banner image with the phrase;

A planning exhibition generation step (S5) of defining the characteristics of the exhibition exhibition based on the banner image generated in the exhibition planning image generation step (S4) and storing it in the exhibition exhibition database;

A customized exhibition creation step (S6) of extracting a list of exhibitions from the exhibitions stored in the exhibition database, and registering recommended products matching the characteristics of the extracted exhibitions before the exhibitions (S6); Method for automatically building content for planning sales, characterized in that configured to include.
The method of claim 9

The trend extraction step (S1),

A method characterized in that the search word is extracted using a crawling technique.
The method of claim 9, wherein the pre-plan image analysis step (S3),

The method, characterized in that the search and analysis of the image is performed using a machine learning technique based on artificial intelligence.
The method of claim 9, wherein the customized exhibition creation step (S6),

A method comprising selecting a recommended product using a machine learning technique based on artificial intelligence.
The method of claim 9,

A method, characterized in that the seller of the product pays an advertisement fee for the product displayed in the exhibition.