KR102615605B1 - Method, device and system for providing styling recommendation and lifestyle item sales platform service based on user interaction - Google Patents

Abstract

A device according to an embodiment acquires the user's personal information from the user's terminal, obtains the user's preference information from the user's terminal, obtains the user's provisioned information from the user's terminal, and obtains the user's personal information, the user's personal information, and the user's personal information. Based on the user's preference information and the user's information, the user's recommended item is selected among the items sold on the platform, and recommendation information for the selected user's recommended item is generated and provided to the user's terminal.

Description

Method, device and system for providing user interaction-based styling recommendation and lifestyle item sales platform service { METHOD, DEVICE AND SYSTEM FOR PROVIDING STYLING RECOMMENDATION AND LIFESTYLE ITEM SALES PLATFORM SERVICE BASED ON USER INTERACTION }

The following embodiments relate to technology for providing a platform service that recommends styling and sells lifestyle items based on user interaction.

As electronic commerce has become more active with the advancement of electronic communication technology, the buying and selling of products online has increased. E-commerce is becoming more and more popular because it is easier to search for desired products compared to offline purchases, and various products can be compared at once.

In addition, as a social atmosphere that pursues and supports one's own unique values rather than popular ones has been formed recently, customized services that maximize customer satisfaction by respecting individual tastes and personalities in the service field have increased. I'm doing it.

In the case of online shopping malls, as the number of people purchasing clothing and lifestyle items online increases, the number of online shopping malls also increases rapidly, causing users to become confused among too much information, and also without users checking the items themselves. There are many cases where people purchase an item without checking the size because it can be purchased, or without confirming whether the item suits them.

Therefore, there is a need for a platform that recommends user styling based on user interaction and sells lifestyle items suitable for the user.

Korean Patent No. 10-2503580 (announced on February 24, 2023) Korean Patent No. 10-2429572 (announced on 2022.08.03) Korean Patent No. 10-2550214 (announced on June 30, 2023) Korean Patent Publication No. 10-2023-0088952 (published on June 20, 2023)

Embodiments seek to recommend items to users based on user interactions.

In embodiments, when more recommended items are selected than the standard number, more suitable selection items are selected among the recommended items.

Embodiments seek to recommend styling by combining selected items.

The object of the present invention is not limited to the object mentioned above, and other objects not mentioned can be clearly understood from the description below.

According to one embodiment, a method of providing a user interaction-based styling recommendation and lifestyle item sales platform service includes the user's name, the user's age, the user's gender, the user's occupation, and the user's body type from the user's terminal. Obtaining the user's personal information including; Obtaining the user's preference information including the user's preferred style and the user's preferred brand from the user's terminal; Obtaining the user's equipment information including information on items the user has from the user's terminal; Selecting a recommended item for the user from among items sold on the platform based on the user's personal information, the user's preference information, and the user's provided information; and generating recommendation information for the user's recommended item and providing it to the user's terminal.

A method of providing a user interaction-based styling recommendation and lifestyle item sales platform service includes: checking the number of items selected as recommended items by the user; and selecting a selection item by selecting the user's recommended items based on the fact that the number of items is greater than a preset standard number. The step of selecting a selection item by selecting the user's recommended items includes: , generating an item-related image data set by inputting each image of an item sold on the platform into a related image extraction artificial neural network, and selecting a selection item using the item-related image data set.

The step of selecting a selection item using the item-related image data set includes obtaining an image of the item possessed by the user, which is an image of the item possessed by the user, based on the information possessed by the user; Inputting a provided item image into the related image extraction artificial neural network to obtain a first image, the item - a second related image whose similarity to the first image among the related images stored in the related image data set is higher than a preset standard ratio It includes acquiring an image, and selecting a recommended item that matches the second image as a selection item.

The step of selecting a selection item using the item-related image data set includes obtaining confirmation location information, which is the user's location information during a preset reference period, from the user's terminal, and preset among the confirmation location information. Extracting a reference location as out-of-town location information, which is location information outside a preset reference range, obtaining an out-out time that is a time corresponding to the out-out location information, Based on the out-out time, the user's main outing time zone A step of selecting, based on the user's main outing time, generating a daytime outing rate and a nighttime outing rate, comparing the daytime outing rate and the nighttime outing rate, wherein the daytime outing rate is greater than the nighttime outing rate. If it is greater than the item-related images stored in the related image data set, selecting a selection item by assigning weight to related images containing a first color more than a target ratio, the night outing ratio is greater than the daytime going out ratio. In this case, selecting a selection item by assigning weight to related images containing a second color greater than the target ratio among the related images stored in the item-related image data set, and the daytime going out ratio and the night going out ratio are If there is a match, the method further includes selecting a selection item by assigning a weight to a related image containing a third color greater than the target ratio among the related images stored in the item-related image data set.

The step of selecting a selection item using the item-related image data set includes obtaining information on a visited brand, which is a brand visited by the user during the reference period, from the user's terminal, based on the information on the visited brand. , acquiring the brand logo of the visited brand, acquiring a third image by inputting the brand logo into the related image extraction artificial neural network, similarity to the third image among the related images stored in the item-related image data set. It further includes obtaining a fourth image, which is a related image higher than a preset standard ratio, and selecting an item matched with the fourth image as a selection item.

The method of providing a user interaction-based styling recommendation and lifestyle item sales platform service further includes generating recommended styling by combining the user's selected items, and providing the generated recommended styling to the user's terminal, The step of generating recommended styling by combining the user's selected items includes classifying the selected items into categories, each category including the color of the item, the texture of the item, the thickness of the item, the material of the item, the pattern of the item, and the thickness of the item. It includes selecting one of the selected items based on size and extracting it as a representative item, and generating recommended styling by combining the representative items extracted for each category.

A method of providing a user interaction-based styling recommendation and lifestyle item sales platform service includes: selecting recommended sales items based on a history of items being sold through the platform; And further comprising transmitting information on the recommended sales item to the seller's terminal, wherein the step of selecting a recommended sales item based on the history of the item being sold through the platform includes: Obtaining sales history, which is the history of items being sold through the platform; determining popular brands, which are the most sold brands, based on the sales history; and popular brands, which are the most sold categories, based on the sales history. A step of checking the category, based on the sales history, selecting an item sold more than a preset target number as a popular item, and checking the popular color that is the most included color among the popular items, the sales history As a basis, identifying a popular brand that is the most sold brand, identifying a popular search keyword that is the most included search keyword during the first period through a search engine of the platform, the popular brand, the popular category , checking whether there is a new item that satisfies all of the popular color and the popular search keyword, and a new item that satisfies all of the popular brand, the popular category, the popular color, and the popular search keyword as the first item. If confirmed, selecting the first item as the recommended sales item; if it is confirmed that there is no new item satisfying all of the popular brand, the popular category, the popular color, and the popular search keyword, the popular brand, the popular search keyword, Confirming a new item that satisfies at least one of a popular category, the popular color, and the popular search keyword as a candidate sales item, and the candidate item among the popular brand, the popular category, the popular color, and the popular search keyword It includes the step of confirming the satisfied number, sorting them in order of increasing the satisfied number, and generating a list including the candidate sales items.

The method of providing a user interaction-based styling recommendation and lifestyle item sales platform service further includes the step of selling the user's recommended item to the user when receiving a purchase request for the user's recommended item from the user's terminal; , the step of selling the user's recommended item to the user includes obtaining the user's purchase history of the same category, which is the history of the user purchasing items of the same category as the user's recommended item, the same category of the user Generating a purchase cycle of the same category based on purchase history; determining a first date, which is the next purchase date of the item of the same category as the user's recommended item, based on the current date and the purchase cycle of the same category; A step of determining whether a user's recommended item whose expiration date is the first date exists among the user's recommended items stored in the warehouse through warehouse data, a user whose item's expiration date is the first date If it is confirmed that the recommended item exists as the first recommended item, selling the first recommended item to the user; if it is confirmed that the user's recommended item whose expiration date is the first date does not exist, the first recommended item A step of generating second dates that fall before a preset range based on the date, a step of determining whether a user's recommended item exists whose expiration date of the item is the second date, and the expiration date of the item is the second date If it is confirmed that the user's recommended item exists as a second recommended item, selling the second recommended item to the user, and if it is confirmed that the user's recommended item whose expiration date is the second date does not exist, Confirming a third recommended item among the user's recommended items stored in the warehouse whose expiration date is the user's recommended item with the earliest date after the first date, and selling the third recommended item to the user. do.

Embodiments may recommend items to users based on user interactions.

In embodiments, when more recommended items are selected than the standard number, a more suitable item to the user can be recommended by selecting the recommended items and selecting the selected items.

Embodiments may recommend styling by combining selected items.

Meanwhile, the effects according to the embodiments are not limited to those mentioned above, and other effects not mentioned can be clearly understood by those skilled in the art from the description below.

1 is a diagram schematically showing the configuration of a system according to an embodiment.
Figure 2 is a flowchart for explaining the process of selecting a user's recommended item according to an embodiment.
Figure 3 is a flow chart to explain the process of selecting a user's recommended item through an artificial neural network according to an embodiment.
Figure 4 is a flowchart for explaining the process of selecting a selection item according to an embodiment.
Figure 5 is a flow chart to explain the process of selecting a selection item through an artificial neural network for extracting related images according to an embodiment.
FIG. 6 is a flowchart illustrating a process of selecting a selection item through items and item-related image data sets owned by a user according to an embodiment.
Figure 7 is a flowchart for explaining the process of selecting a selection item through a user's main outing time and item-related image data set according to an embodiment.
Figure 8 is a flowchart illustrating a process of selecting a selected item through a visited brand and item-related image data set according to an embodiment.
Figure 9 is a flowchart illustrating a process for recommending styling by combining selected items according to an embodiment.
Figure 10 is a flow chart to explain the process of selecting recommended sales items and providing them to the seller according to an embodiment.
Figure 11 is a flowchart for explaining the process of selecting a recommended sales item according to an embodiment.
Figure 12 is a flow chart to explain the process of selling a user's recommended item to a user according to an embodiment.
Figure 13 is an exemplary diagram of the configuration of a device according to an embodiment.

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

Terms such as first or second may be used to describe various components, but these terms should be interpreted only for the purpose of distinguishing one component from another component. For example, a first component may be named a second component, and similarly, the second component may also be named a first component.

When a component is referred to as being “connected” to another component, it should be understood that it may be directly connected or connected to the other component, but that other components may exist in between.

The terms used in the examples are for descriptive purposes only and should not be construed as limiting. Singular expressions include plural expressions unless the context clearly dictates otherwise. In this specification, terms such as “comprise” or “have” are intended to designate the presence of features, numbers, steps, operations, components, parts, or combinations thereof described in the specification, but are not intended to indicate the presence of one or more other features. It should be understood that this does not exclude in advance the possibility of the existence or addition of elements, numbers, steps, operations, components, parts, or combinations thereof.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by a person of ordinary skill in the technical field to which the embodiments belong. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related technology, and unless explicitly defined in the present application, should not be interpreted in an ideal or excessively formal sense. No.

In addition, when describing with reference to the accompanying drawings, identical components will be assigned the same reference numerals regardless of the reference numerals, and overlapping descriptions thereof will be omitted. In describing the embodiments, if it is determined that detailed descriptions of related known technologies may unnecessarily obscure the gist of the embodiments, the detailed descriptions are omitted.

Embodiments may be implemented in various types of products such as personal computers, laptop computers, tablet computers, smart phones, televisions, smart home appliances, intelligent vehicles, kiosks, and wearable devices.

1 is a diagram schematically showing the configuration of a system according to an embodiment.

Referring to FIG. 1, a system according to an embodiment may include a user terminal 100 and a device 200 that can communicate with each other through a communication network.

First, a communication network can be configured regardless of the communication mode, such as wired or wireless, and can be implemented in various forms to enable communication between servers and between servers and terminals.

The user's terminal 100 is a terminal used by a user who has a high probability of purchasing, that is, a lifestyle item suitable for the user is recommended, and the user wishes to purchase the recommended item. It is a mobile phone, a desktop PC, a laptop PC, and a tablet PC. , It may be implemented as a smartphone, etc., but is not limited to this, and may be implemented as various types of communication devices that can be connected to an external server. For example, as shown in FIG. 1, the user's terminal 100 may be a smartphone, and may be employed differently depending on the embodiment.

The user's terminal 100 may be configured to perform all or part of the calculation function, storage/reference function, input/output function, and control function of a typical computer. The user's terminal 100 may be configured to communicate with the device 200 wired or wirelessly.

The user's terminal 100 is connected to a website operated by a person or organization that provides services using the device 200, or an application developed and distributed by a person or organization that provides services using the device 200. This can be installed. The user's terminal 100 may be linked to the device 200 through a website or application.

In FIG. 1 and the following description, for convenience of explanation, only the user terminal 100 is shown and described, but the number of terminals may vary depending on the embodiment. As long as the processing capacity of the device 200 allows, the number of terminals is not particularly limited.

The device 200 may be its own server owned by a person or organization that provides services using the device 200, a cloud server, or a p2p (peer-to-peer) set of distributed nodes. It may be possible. The device 200 may be configured to perform all or part of the calculation function, storage/reference function, input/output function, and control function of a typical computer. The device 200 may be configured to communicate with the user's terminal 100 wired or wirelessly.

Additionally, the device 200 can communicate wired or wirelessly with websites including web pages including articles and SNS including blogs, cafes, Instagram, Facebook, Twitter, and YouTube, and the device 200 You can obtain information by accessing the website.

The device 200 obtains the user's personal information from the user's terminal 100, obtains the user's preference information from the user's terminal 100, and obtains the user's provisioning information from the user's terminal 100, Based on the user's personal information, user's preference information, and user's provided information, the user's recommended item is selected among the items sold on the platform, and recommendation information for the selected user's recommended item is generated to provide the user's terminal (100) It can be provided as .

In the present invention, artificial intelligence (AI) refers to technology that imitates human learning ability, reasoning ability, perception ability, etc. and implements this with a computer, and includes concepts such as machine learning and symbolic logic. may include. Machine Learning (ML) is an algorithmic technology that classifies or learns the characteristics of input data on its own. Artificial intelligence technology is a machine learning algorithm that analyzes input data, learns the results of the analysis, and makes judgments or predictions based on the results of the learning. Additionally, technologies that mimic the functions of the human brain, such as cognition and judgment, using machine learning algorithms can also be understood as the category of artificial intelligence. For example, technical fields such as verbal understanding, visual understanding, reasoning/prediction, knowledge representation, and motion control may be included.

Machine learning can refer to the process of training a neural network model using experience processing data. Machine learning can mean that computer software improves its own data processing capabilities. A neural network model is built by modeling the correlation between data, and the correlation can be expressed by a plurality of parameters. A neural network model extracts and analyzes features from given data to derive correlations between data. Repeating this process to optimize the parameters of the neural network model can be called machine learning. For example, a neural network model can learn the mapping (correlation) between input and output for data given as input-output pairs. Alternatively, a neural network model may learn the relationships by deriving regularities between given data even when only input data is given.

An artificial intelligence learning model or neural network model may be designed to implement the human brain structure on a computer, and may include a plurality of network nodes with weights that simulate neurons of a human neural network. A plurality of network nodes may have a connection relationship with each other by simulating the synaptic activity of neurons in which neurons exchange signals through synapses. In an artificial intelligence learning model, multiple network nodes are located in layers of different depths and can exchange data according to convolutional connection relationships. The artificial intelligence learning model may be, for example, an artificial neural network model (Artificial Neural Network), a convolution neural network (CNN) model, etc. As an example, an artificial intelligence learning model may be machine-learned according to methods such as supervised learning, unsupervised learning, and reinforcement learning. Machine learning algorithms for performing machine learning include Decision Tree, Bayesian Network, Support Vector Machine, Artificial Neural Network, and Ada-boost. , Perceptron, Genetic Programming, Clustering, etc. can be used.

Among them, CNN is a type of multilayer perceptrons designed to use minimal preprocessing. CNN consists of one or several convolution layers and general artificial neural network layers on top of them, and additionally utilizes weights and pooling layers. Thanks to this structure, CNN can fully utilize input data with a two-dimensional structure. Compared to other deep learning structures, CNN shows good performance in both video and audio fields. CNNs can also be trained via standard back propagation. CNNs have the advantage of being easier to train and using fewer parameters than other feedforward artificial neural network techniques.

Convolutional networks are neural networks that contain sets of nodes with bound parameters. The increasing size of available training data and the availability of computational power, combined with algorithmic advances such as piecewise linear unit and dropout training, have led to significant improvements in many computer vision tasks. For extremely large data sets, such as those available for many tasks today, overfitting is not critical, and increasing the size of the network improves test accuracy. Optimal use of computing resources becomes a limiting factor. For this purpose, distributed, scalable implementations of deep neural networks can be used.

Figure 2 is a flowchart for explaining the process of selecting a user's recommended item according to an embodiment.

Referring to FIG. 2, first, in step S201, the device 200 may obtain the user's personal information from the user's terminal 100.

Specifically, the device 200 may obtain the user's personal information including the user's name, user's age, user's gender, user's occupation, and user's body type from the user's terminal 100. Also, at this time, the user's body shape may include the user's body size, including the user's height and user's weight.

In step S202, the device 200 may obtain the user's preference information from the user's terminal 100.

Specifically, the device 200 may obtain the user's preference information including the user's preferred style, which is the user's preferred style, and the user's preferred brand, which is the user's preferred brand. Here, the user's preferred style may include cuteness, simplicity, casualness, elegance, etc., and may also include styles preferred by other users.

In step S203, the device 200 may obtain the user's information from the user's terminal 100.

Specifically, the device 200 may obtain the user's equipment information, including information on items the user owns, from the user's terminal 100. At this time, the information on the item that the user has may include the image of the item that the user has, the name of the item, the category of the item, the brand of the item, the color of the item, etc. It is not limited. In addition, the device 200 can obtain the user's equipment information, including information on one equipment item that the user has, from the user's terminal 100, and obtain information on a plurality of equipment items that the user has. It is possible to obtain the user's included information. In other words, the user may have one item or multiple items.

In step S204, the device 200 may select a recommended item for the user from among the items sold on the platform based on the user's personal information, user's preference information, and user's equipment information. Here, one user's recommended item may be selected, or multiple items may be selected.

Specifically, the device 200 applies the user's personal information, user's preference information, user's equipment information, and information on sales items being sold on the platform to the first artificial neural network to Among the items on sale, you can check the user's personal information, user's preference information, and the item that best matches the user's equipment information and select it as the user's recommended item. To this end, the device 200 may operate a platform for selling items. Specifically, the device 200 receives from the seller's terminal the name of the sale item sold by the seller, the image of the sale item, the category of the sale item, the characteristics of the sale item, the size of the sale item, the color of the sale item, and the brand of the sale item. It is possible to obtain information on sales items including, and the device 200 can upload sales items to be sold on the platform through the information on the acquired sales items. Additionally, the device 200 can obtain information about the sales items including By obtaining the user's personal information, user's preference information, and user's equipment information from the terminal 100, the user's recommended items among sales items sold on the platform can be recommended. That is, the device 200 can communicate wired or wirelessly with the seller's terminal, where the seller is a seller who sells items through a platform and may be the same person as the device manager who manages the device 200. Additionally, the device 200 may store information on sales items to be sold on the platform obtained from the seller's terminal in a database provided in the device 200.

Meanwhile, the process of selecting a user's recommended item from among the items for sale on the platform through the first artificial neural network will be described with reference to FIG. 3.

In step S205, the device 200 may generate recommendation information for the user's recommended item and provide it to the user's terminal 100.

Specifically, the device 200 can select a user's recommended item through a first artificial neural network, and when the user's recommended item is selected, obtain a link where the user's recommended item is sold through the platform, and link and Recommendation information for the user's recommended item including information on the user's recommended item can be generated. Additionally, the device 200 may provide recommendation information about the generated user's recommended item to the user's terminal 100.

For this reason, the device 200 uses the platform through the user's personal information, which is information related to the user, the user's preference information, which is information about the user's preferred style and brand, and the user's equipment information, which is information about the items the user has. It has the effect of recommending items that users are likely to purchase among the items on sale.

Figure 3 is a flow chart to explain the process of selecting a user's recommended item through an artificial neural network according to an embodiment.

Referring to FIG. 3, first, in step S301, the device 200 generates a first input signal based on the user's personal information, user's preference information, user's equipment information, and information on items sold on the platform. You can.

Specifically, the device 200 may perform a process of preprocessing the user's personal information, user's preference information, user's equipment information, and information on sales items being sold on the platform. The pre-processed user's personal information, user preference information, user's equipment information, and information on sales items sold on the platform are used as input to the first artificial neural network, or are processed through normal processing to remove unnecessary information. 1 An input signal can be generated.

In step S302, the device 200 may generate a first output signal by applying the first input signal to the first artificial neural network.

Specifically, the device 200 may include a first artificial neural network or communicate wired or wirelessly with a separate first artificial neural network, where the first artificial neural network includes the user's personal information, the user's preference information, and the user's equipment. It may be an algorithm that receives information and information on sales items sold on a platform as input and predicts the matching score between the user and the sales item.

In other words, the first artificial neural network predicts and outputs items that the user is likely to purchase among the sales items on sale on the platform through the user's personal information, user preference information, user's equipment information, and information on sales items being sold on the platform. can do.

The learning device in which learning of the first artificial neural network is performed may be the same device as the device 200 that generates a matching score using the learned first artificial neural network, or may be a separate device.

The learning device may apply input to the first artificial neural network. The first artificial neural network may be an artificial neural network learned according to reinforcement learning. The first artificial neural network may be a Q-Network, Depp Q-Network (DQN), or relational network (RL) structure suitable for outputting abstract inference through reinforcement learning.

The first artificial neural network learned according to reinforcement learning can be updated and optimized by reflecting the evaluation of various rewards. For example, the first reward may have a higher reward value as the matching score of a suitable sales item is selected according to the user's personal information, and the second reward may increase as the matching score of a suitable sales item is selected according to the user's preference information. The value may increase, and the third reward may increase the reward value as the matching score of a suitable sales item is selected according to the user's information. The fourth reward may be a matching score of an unsuitable sales item according to the user's personal information. The more you do not select, the higher the reward value can be. The fifth reward can increase the reward value the more you do not select the matching score of an unsuitable sales item according to the user's preference information. The sixth reward is the user's possession information. Accordingly, the reward value may increase as the matching score for unsuitable sales items is not selected.

The learning device may obtain an output from the first artificial neural network. The output of the first artificial neural network may be a matching score of sales items being sold on the platform according to the user's personal information, user's preference information, and user's information. At this time, the first artificial neural network can analyze and predict a matching score suitable for the sales item by considering the user's personal information, user's preference information, and user's equipment information, and output the predicted matching score.

The learning device may evaluate the output of the first artificial neural network and pay a reward. At this time, the evaluation of the output can be divided into first compensation, second compensation, third compensation, fourth compensation, fifth compensation, and sixth compensation.

Specifically, the learning device awards a large number of first rewards when it outputs a matching score of a suitable sales item according to the user's personal information, and grants a large number of second rewards when it outputs a matching score of a suitable sales item according to the user's preference information. If the matching score of a suitable sales item is output according to the user's information, a large number of third rewards are awarded, and if the matching score of an unsuitable sales item is not output according to the user's personal information, a large number of fourth rewards are awarded. , if the matching score of an unsuitable sales item is not output according to the user's preference information, the fifth reward is awarded more, and if the matching score of an unsuitable sales item is not output according to the user's information, the sixth reward is awarded more. It can be awarded.

The learning device may update the first artificial neural network based on the evaluation.

Specifically, the learning device is a first artificial neural network that analyzes information about the matching score of a sale item by considering the user's personal information, the user's preference information, the user's equipment information, and information on the sale item being sold on the platform. In the environment, an artificial neural network is used through the process of optimizing the policy that determines the actions to be taken in specific states so that the expected value of the sum of rewards is maximized. It can be renewed.

Meanwhile, the process of optimizing the policy can be accomplished through estimating the maximum expected value of the sum of rewards or the maximum value of the Q-function, or estimating the minimum value of the loss function of the Q-function. Estimation of the minimum value of the loss function can be done through stochastic gradient descent (SGD), etc. The process of optimizing the policy is not limited to this, and various optimization algorithms used in reinforcement learning can be used.

The learning device can gradually update the first artificial neural network by repeating the learning process of the first artificial neural network described above. Through this, the learning device can learn an artificial neural network that predicts and outputs a matching score for a sale item by considering the user's personal information, user preference information, user's equipment information, and information on the sale item.

In other words, the learning device considers the user's personal information, user's preference information, user's equipment information, and information on sale items, and outputs the matching score of the item, including the first reward, the second reward, the third reward, and the The first artificial neural network can be trained by adjusting the analysis standard to reflect reinforcement learning through the 4th reward, the 5th reward, and the 6th reward.

In step S303, the device 200 may generate a matching score corresponding to an item being sold on the platform, based on the first output signal. At this time, the matching score corresponding to the item may be a real number between 0 and 1.

For example, the device 200 generates a matching score of the first sale item as 0.8, a matching score of the second sale item as 0.1, and a matching score of the third sale item as 0.75 through the first artificial neural network. , the matching score of the fourth sales item can be generated as 0.3, and the matching score of the fifth sales item can be generated as 0.9.

In step S304, the device 200 may select an item with a matching score higher than the reference score as the user's recommended item. Here, the standard score is a preset score and may vary depending on the embodiment.

That is, the device 200 may select a sales item whose matching score output through the first artificial neural network is higher than a preset standard score as the user's recommended item.

For example, when the reference score is 0.7, the device 200 generates a matching score of the first sale item as 0.8, a matching score of the second sale item as 0.1, and a third sale item through the first artificial neural network. The matching score of the sale item can be created as 0.75, the matching score of the fourth sale item can be created as 0.3, the matching score of the fifth sale item can be created as 0.9, and the matching score of the first sale item is higher than the standard score of 0.7. A sale item, a third sale item, or a fifth sale item can be selected as the user's recommended item.

Due to this, the device 200 has the effect of selecting a sales item that the user is likely to purchase, that is, a sales item with the highest matching score with the user, as the user's recommended item and recommending it to the user among the sales items being sold on the platform.

Figure 4 is a flowchart for explaining the process of selecting a selection item according to an embodiment.

Referring to FIG. 4, first, in step S401, the device 200 can check the number of items selected as recommended items by the user.

That is, when there are too many items selected as recommended items for the user, the device 200 may check the number of items selected as recommended items for the user in order to select more items suitable for the user and recommend them to the user.

In step S402, the device 200 may select a selection item by selecting the user's recommended items based on whether the number of items is greater than the standard number. Here, the standard number is a preset number and may vary depending on the embodiment.

Specifically, if the device 200 determines that the number of items selected as the user's recommended items is greater than a preset standard number, the device 200 may select the user's recommended items and select a selection item.

At this time, the device 200 may acquire the user's location information from the user's terminal 100 in order to select a selection item among the user's recommended items, and based on the acquired user's location information, the device 200 and Weather information about the user's location can be obtained through the Korea Meteorological Administration database, which can be communicated via wired or wireless communication. Additionally, the device 200 may select an item corresponding to the weather information as a selection item based on weather information about the user's location. At this time, the device 200 may be provided with a database storing information on sale items sold through the platform in order to select items corresponding to weather information as selection items, and the information on sale items sold through the platform may include the sale items. The name of the sale item, the image of the sale item, the category of the sale item, the characteristics of the sale item, the size of the sale item, the color of the sale item, and the brand of the sale item may be included. Additionally, the sale item information may include the ability to wear the sale item. Weather information including the season may be further included.

In addition, the device 200 may select a selection item from among the user's recommended items by using a related image extraction artificial neural network. A detailed description regarding this will be referred to with reference to FIG. 5 .

For this reason, when there are many items selected as recommended items for the user, the device 200 can select a selection item by selecting an item more suitable for the user among the recommended items, and by recommending the selection item to the user, the user can You can receive recommendations for items that are more suitable for you.

Figure 5 is a flow chart to explain the process of selecting a selection item through an artificial neural network for extracting related images according to an embodiment.

Referring to FIG. 5, first, in step S501, the device 200 may input each image of an item being sold on the platform into a related image extraction artificial neural network to generate an item-related image data set.

Specifically, the device 200 may be provided with a related image extraction artificial neural network or may communicate with a separate related image extraction artificial neural network wired or wirelessly. When an image is input, the related image extraction artificial neural network determines the color of the image and the corresponding image. It is an artificial neural network that analyzes the shape of and outputs images that are related to the input image. In addition, although it was not specifically written regarding the related image extraction artificial neural network, the related image extraction artificial neural network can be implemented in the same form as a commonly used image retrieval artificial neural network.

That is, the device 200 can obtain information about the sales item being sold on the platform through the platform, and can obtain an image of the sales item being sold on the platform through the acquired information about the sales item. In addition, the device 200 can obtain a related image associated with the image of the sales item by applying the image of the acquired sales item to a related image extraction artificial neural network, and match the acquired related image with the sales item to determine item-relatedness. An image data set can be created. Here, an item-related image data set can be created by matching 2 to 3 related images per sold item, but the number of related images matched per item is not limited to this. Additionally, the generated item-related image data set may be stored in a database provided in the device 200.

In step S502, the device 200 may select a selection item using an item-related image data set.

Specifically, the device 200 can generate an item-related image data set by applying the image of the sales item being sold on the platform to a related image extraction artificial neural network, and use the generated item-related image data set to create a recommended item. You can select selected items. The process of selecting a selection item using an item-related image data set will be described with reference to FIGS. 6 to 8.

Figure 6 is a flowchart illustrating a process of selecting a selection item through items and item-related image data sets owned by a user according to an embodiment.

Referring to FIG. 6, first, in step S601, the device 200 may obtain an image of the item the user has, which is an image of the item the user has, based on the user's equipment information.

Specifically, the device 200 checks the user's equipment information, which is information about the items the user has, obtained from the user's terminal 100, and checks the image of the item the user has, that is, the equipment the user has. You can obtain the image of the item that the user has, which is the image of the item.

In step S602, the device 200 may obtain the first image by inputting the image of the user's equipped item into a related image extraction artificial neural network.

Specifically, the device 200 may obtain a first image, which is an image related to the user's equipped item image, by inputting the user's equipped item image obtained through the user's equipped item information into a related image extraction artificial neural network.

In step S603, the device 200 may acquire a second image, which is a related image whose similarity to the first image is higher than the reference rate among the related images stored in the item-related image data set. Here, the reference ratio is a preset ratio and may vary depending on the embodiment.

Specifically, the device 200 may acquire the first image by inputting the image of the item equipped with the user into the related image extraction artificial neural network, and the device 200 may also input the acquired first image and the item-related image data set into the related image data set. By comparing the stored related images, a second image, which is a related image whose similarity to the first image is higher than the reference rate among the related images stored in the item-related image data set, may be obtained.

At this time, the device 200 compares the pixel values of the first image and the related image stored in the data set to identify a related image that has a higher similarity rate with the first image among the related images stored in the item-related image data set. The similarity between the first image and the related image may be calculated, and in this case, the process of calculating the similarity through pixel values can be calculated through a commonly used method. Additionally, the device 200 may check the latent vector value of the first image and the latent vector value of the related image and calculate the cosine similarity using the cosine angle between the vectors of the two images. In this case, the cosine similarity may be calculated through the vector value. The process of calculating can be calculated through commonly used methods. Meanwhile, the method of calculating the similarity between two images is not limited to this.

In step S604, the device 200 may select a recommended item that matches the second image as a selection item.

Specifically, when the device 200 identifies a second image that is a related image whose similarity to the first image is higher than the standard rate among the related images stored in the item-related image data set, the device 200 selects the second image through the item-related image data set. The items matched with can be checked, and the item selected as the user's recommended item among the items matched with the second image can be selected as a selection item.

At this time, the categories of the item provided by the user and the user's recommended item may be the same category, but may also be different categories. For example, if the item possessed by the user is blue jewelry, the device 200 may acquire the first sea image by inputting the blue jewelry, which is the item possessed by the user, into the related image extraction artificial neural network, and item-related. Among the related images stored in the image data set, a second sea image, which is a related image with a higher similarity rate to the first sea image, can be obtained, and the item matched with the second sea image is selected as the user's recommended item. If confirmed, dark blue pants can be selected as a selection item.

Due to this, the device 200 selects a recommended item having a similar related image to an item owned by the user as a selection item, thereby having the effect of recommending an item that is more suitable for the user.

Meanwhile, the device 200 checks the user's information acquired from the user's terminal 100, and if it is confirmed that the user has multiple items, the device 200 detects the user's information from the user's terminal 100. You can obtain the wearing time corresponding to each of the multiple items you have. Here, the wearing time is the time to check how long and how often the item is being worn. If the item is not a wearable item, it may be the time to check how long and how often the item is being used. there is. That is, the wearing time may be the time the item is worn or used during a preset period. For example, if the user has item A, item B, and item C, and the set period is a week from July 14, 2023 to July 20, 2023, the device 200 stores the user's From the terminal 100, 20 hours can be obtained as the wearing time of item A, 25 hours can be obtained as the wearing time of item B, and 10 hours can be obtained as the wearing time of item C.

Additionally, the device 200 may set a wearing weight corresponding to the item owned by the user based on the wearing time corresponding to the item owned by the user obtained from the user's terminal 100. For example, if the user has item A, item B, and item C, the wearing time of item A is 20 hours, the wearing time of item B is 25 hours, and the wearing time of item C is 10 hours. When the time is acquired, the device 200 may set the wearing weight to 4 in response to item A based on 20 hours, which is the wearing time of item A, and 25 hours, which is the wearing time of item B. The wearing weight can be set to 5 in response to the item B, and the wearing weight can be set to 2 in response to the item C based on the 10 hours of wearing time of the item C.

In addition, when the wearing weight corresponding to the item owned by the user is set, the device 200 provides a second image corresponding to the item in the second image, which is a related image whose similarity to the first image output based on the item is higher than the reference rate. Wearing weights can be applied, and selection items from recommended items can be selected based on the second image to which the wearing weights have been applied.

That is, when the user has a plurality of items, the device 200 can further obtain the wearing time corresponding to the item owned by the user, generate a wearing weight through the obtained wearing time, and apply the wearing weight. By selecting selected items, it is possible to recommend items that are highly related to items that the user wears for a long time and frequently.

Figure 7 is a flow chart to explain the process of selecting a selection item through the user's main outing time and item-related image data set according to an embodiment.

Referring to FIG. 7, in step S701, the device 200 may obtain confirmed location information, which is the user's location information during the reference period, from the user's terminal 100. Here, the reference period is a preset period and may vary depending on the embodiment.

Specifically, the device 200 may obtain the user's location information during a preset reference period from the user's terminal 100 and generate the user's location information during the reference period as confirmation location information. Here, the user's location information may include information about where the user is located by time, that is, the time and coordinates that are the user's location may be matched and stored in the user's location information. Additionally, for this purpose, the user's terminal 100 may generate the user's location information at preset intervals.

For example, if the reference period is from July 10, 2023 to July 20, 2023, the device 200 receives information from the user's terminal 100 from July 10, 2023 to July 20, 2023. The user's location information, which is information about where the user was located by time, can be obtained, and the acquired user's location information from July 10, 2023 to July 20, 2023 can be generated as confirmed location information. .

In step S702, the device 200 may extract location information that deviates from the reference location by more than a reference range from among the confirmed location information as out-of-home location information. Here, the reference location may be a location obtained in advance from the user's terminal 100, and may be the coordinates of the place where the user resides. Additionally, the reference range is a preset range and may vary depending on the embodiment.

Specifically, the device 200 may obtain confirmed location information, which is the user's location information during the reference period, from the user's terminal 100, and location information that deviates from a preset reference location among the confirmed location information by more than a preset reference range. You can extract outgoing location information by checking . In other words, the device 200 can extract out-of-town location information, which is location information that is outside the place where the user resides (reference location) (more than the standard range), that is, location information where the user has gone out beyond the place where the user resides, from among the confirmed location information. there is.

In step S703, the device 200 may obtain the exit time, which is the time corresponding to the exit location information.

Specifically, when going out location information is extracted from the confirmed location information, the device 200 can obtain the going out time, which is the time included in the going out location information. In this process, the device 200 determines that the going out time acquired through the going out location information is 7:00 on July 10, 2023, 7:01 on July 10, 2023, 7:02 on July 10, 2023, If it is 7:03 on July 10, 2023, ,,, , 11:58 on July 10, 2023, 11:49 on July 10, 2023, and 12:00 on July 10, 2023, the time to go out is It can be changed in time zones from 07:00 to 12:00 on July 10, 2023.

In step S704, the device 200 may select the user's main outing time based on the outing time.

Specifically, the device 200 can confirm the going out time through the going out location information, and can generate a distribution degree of the going out time through the confirmed going out time. In addition, based on the generated distribution of going out times, the device 200 can identify times when going out is concentrated, that is, the density is higher than the preset standard density, and select the corresponding time as the user's main going out time. there is.

In step S705, the device 200 may generate a daytime outing rate and a night outing rate based on the user's main outing time zone.

Specifically, the device 200 may generate a daytime outing rate and a nighttime outing rate based on the user's main outing time zone, where the daytime range is determined by the device manager to generate the daytime outing rate and the nighttime outing rate. and the night range may be set in advance, and the day range and night range may vary depending on the embodiment.

For example, if the day range is set to 07:00 to 18:00, the night range is set to 18:00 to 07:00, and the main outing time is 08:00 to 17:00, the device ( 200) can generate the daytime outing rate as 1 and the nighttime outing rate as 0.

In addition, if the day range is set from 07:00 to 18:00, the night range is set to 18:00 to 07:00, and the main outing time is from 19:00 to 04:00, the device 200 The daytime outing rate can be created as 0, and the night outing rate can be created as 1.

Additionally, if the day range is set from 07:00 to 18:00, the night range is set to 18:00 to 07:00, and the main outing time is from 12:00 to 20:00, the device (200 ) can generate the daytime outing ratio as 0.75, and the nighttime outing rate as 0.25.

In step S706, the device 200 may determine whether the rate of going out during the day is greater than the rate of going out at night.

If it is confirmed in step S706 that the rate of going out during the day is greater than the rate of going out at night, in step S707, the device 200 assigns weight to the related images containing the first color more than the target rate among the related images stored in the item-related image data set. You can select selected items by granting them. Here, the target ratio is a preset ratio and may vary depending on the embodiment.

Specifically, if the device 200 determines that the rate of going out during the day is greater than the rate of going out at night, the device 200 may assign weight to the related image in which the first color is included at a target rate or higher among the related images in which the item-related image data set is stored. , Based on the assigned weight, a selection item can be selected from among the recommended items. In this process, the device 200 may further perform an operation of selecting the first color, and the device 200 may select the item purchase history of users whose rate of going out during the day is greater than the rate of going out at night to select the first color. You can check the color of items purchased by users whose daytime outing rate is greater than the nighttime outing rate through the item purchase history of users whose daytime outing rate is greater than their nighttime outing rate, and select the first color based on the most included color. can be selected. Meanwhile, in order to obtain the user's item purchase history, the device 200 matches the information of the user purchasing the item through the platform operated through the device 200 and the information of the item purchased by the user to the device 200. It can be saved in the provided database.

If it is confirmed in step S706 that the rate of going out during the day is not greater than the rate of going out at night, in step S708, the device 200 may determine whether the rate of going out at night is greater than the rate of going out during the day.

If it is confirmed in step S708 that the rate of going out at night is greater than the rate of going out during the day, in step S709, the device 200 assigns weight to the related images containing the second color more than the target rate among the related images stored in the item-related image data set. You can select selected items by granting them. Here, the target ratio is a preset ratio and may vary depending on the embodiment.

Specifically, if the device 200 determines that the rate of going out at night is greater than the rate of going out during the day, the device 200 may assign weight to the related image containing the second color at a target rate or higher among the related images in which the item-related image data set is stored. , Based on the assigned weight, a selection item can be selected from among the recommended items. In this process, the device 200 may further perform an operation of selecting a second color, and the device 200 may select the item purchase history of users whose rate of going out at night is greater than the rate of going out during the day to select the second color. You can check the color of items purchased by users whose night outing rate is greater than their daytime outing rate through the item purchase history of users whose night outing rate is greater than their daytime outing rate, and select a second color based on the most included color. can be selected. Meanwhile, in order to obtain the user's item purchase history, the device 200 matches the information of the user purchasing the item through the platform operated through the device 200 and the information of the item purchased by the user to the device 200. It can be saved in the provided database.

If it is confirmed in step S708 that the rate of going out at night is not greater than the rate of going out during the day, in step S710, the device 200 applies weight to the related images that contain the third color at a target rate or more among the related images stored in the item-related image data set. You can select a selection item by giving . Here, the target ratio is a preset ratio and may vary depending on the embodiment.

Specifically, if the device 200 determines that the daytime outing rate is not greater than the nighttime outing rate and the nighttime outing rate is not greater than the daytime outing rate, and thus the daytime outing rate and the nighttime outing rate are the same, the device 200 selects the item-related image data set. Among the stored related images, a weight may be assigned to related images containing the third color at a target ratio or higher, and based on the given weight, a selection item may be selected from among the recommended items. In this process, the device 200 may further perform an operation of selecting a third color, and the device 200 may select the item purchase history of users who have the same rate of going out during the day and going out at night to select the third color. You can check the color of items purchased by users with the same ratio of going out during the day and going out at night through the item purchase history of users with the same ratio of going out during the day and going out at night, and then select a third color based on the most included color. can be selected. Meanwhile, in order to obtain the user's item purchase history, the device 200 matches the information of the user purchasing the item through the platform operated through the device 200 and the information of the item purchased by the user to the device 200. It can be saved in the provided database.

Figure 8 is a flowchart for explaining the process of selecting a selected item through a visited brand and item-related image data set according to an embodiment.

Referring to FIG. 8, first, in step S801, the device 200 may obtain information on the visited brand, which is the brand visited by the user during the reference period, from the user's terminal 100. Here, the reference period is a preset period and may vary depending on the embodiment.

Specifically, the device 200 may obtain information about the visited brand, which is a brand visited by the user during a preset reference period, from the user's terminal 100. At this time, the information on the visited brand may include the brand name of the visited brand, the brand logo of the visited brand, and the location of the visited brand. Additionally, visited brands may include various types of brands, such as food brands and market brands, as well as brands related to clothing and items.

In step S802, the device 200 may obtain the brand logo of the visited brand based on information about the visited brand.

Specifically, the device 200 may obtain information on the visited brand, which is the brand visited by the user during the reference period, from the user's terminal 100, and obtain the brand logo of the visited brand based on the acquired information on the visited brand. can do.

In step S803, the device 200 may obtain a third image by inputting the brand logo of the visited brand into a related image extraction artificial neural network.

Specifically, when the device 200 obtains the logo of the visited brand through information on the visited brand, which is the brand visited by the user, it can input the logo of the visited brand into the related image extraction artificial neural network, and through the related image extraction artificial neural network. A third image, which is an image related to the logo of the visited brand, can be obtained.

In step S804, the device 200 may acquire a fourth image, which is a related image whose similarity to the third image is higher than the reference rate among the related images stored in the item-related image data set. Here, the reference ratio is a preset ratio and may vary depending on the embodiment.

Specifically, the device 200 may obtain a third image by inputting the brand logo of the visited brand into a related image extraction artificial neural network, and the device 200 may also input the acquired third image and the item-related image data set. By comparing the stored related images, a fourth image, which is a related image whose similarity to the third image is higher than the reference rate among the related images stored in the item-related image data set, may be obtained.

At this time, the device 200 compares the pixel values of the third image and the related image stored in the data set to identify a related image that has a higher similarity rate with the third image among the related images stored in the item-related image data set. The similarity between the third image and the related image may be calculated, and in this case, the process of calculating the similarity through pixel values can be calculated through a commonly used method. Additionally, the device 200 may check the latent vector value of the third image and the latent vector value of the related image and calculate the cosine similarity using the cosine angle between the vectors of the two images. In this case, the cosine similarity may be calculated through the vector value. The process of calculating can be calculated through commonly used methods. Meanwhile, the method of calculating the similarity between two images is not limited to this.

In step S805, the device 200 may select an item matching the fourth image as a selection item.

Specifically, when the device 200 identifies a fourth image that is a related image whose similarity to the third image is higher than the standard rate among the related images stored in the item-related image data set, the device 200 selects the fourth image through the item-related image data set. The items matched with can be checked, and the item selected as the user's recommended item among the items matched with the fourth image can be selected as a selection item.

Due to this, the device 200 selects a recommended item having a similar related image to the brand logo of a brand visited by the user as a selection item, thereby having the effect of recommending an item that is more suitable for the user.

Figure 9 is a flowchart illustrating a process for recommending styling by combining selected items according to an embodiment.

Referring to FIG. 9, first, in step S901, the device 200 may classify the selected items according to categories. Here, the category may include tops, bottoms, jewelry, shoes, perfume, etc., and other categories may also be included.

That is, the device 200 may check the category of the sales item selected as the selection item and classify the selection item according to the category of the sale item.

For example, if the items selected as selection items include A-sale items, B-sale items, C-sale items, D-sale items, E-sale items, F-sale items, G-sale items, H-sale items, and I-sale items, The device 200 checks the categories of A-sale item, B-sale item, C-sale item, D-sale item, E-sale item, F-sale item, G-sale item, H-sale item, and I-sale item to determine A-sale item and B-sale item. Classify sale items, C sale items, and D sale items into items with a category of tops, classify E sale items, and F sale items into items with a category of bottoms, and classify G sale items, H sale items, and I sale items into categories. It can be classified as an item that is shoes.

In step S902, the device 200 may select one of the selected items and extract it as a representative item based on color, texture, thickness, material, pattern, and size for each category.

Specifically, the device 200 can classify selected items according to categories, and the device 200 can classify the color of the item classified by category, the texture of the item, the thickness of the item, the material of the item, the pattern of the item, and the item. Based on the size of , a representative item can be extracted by selecting one of the selected items for each category. Here, the device 200 may select a representative item by considering all of the color, texture, thickness, material, pattern, and size in order to select one of the selected items for each category and extract it as a representative item. , a representative item may be selected by considering at least one of material, pattern, and size. At this time, the device 200 selects one of the selected items for each category, extracts a representative item, and recommends styling by selecting a color combination of items by category, a texture combination of items by category, a thickness combination of items by category, and a combination of thicknesses of items by category. A styling combination database in which material combinations of items, pattern combinations of items by category, and size combinations of items by category are stored may be provided. The styling combination database may store styling combinations according to input from an administrator who manages the device 200, and the number of styling combinations stored in the styling combination database may gradually increase as the embodiment is performed. For example, the color combination of items included in the styling combination database may store a combination of matching an item with a white color category as tops and an item with a dark blue color category as bottoms, and a combination with a navy color category as tops. A combination may be stored to match an item with a beige color category as bottoms, and a combination may be stored to match an item with a beige color category as tops and a black color category as bottoms. Other color combinations may be matched. In addition, the material combination of the items included in the styling combination database may store a combination of matching an item with a linen material category as tops and an item with a cotton material category as bottoms, and an item with a cotton material category as tops. A combination that matches an item whose denim material category is bottoms may be stored, and other material combinations may be matched. Meanwhile, in the above example, only examples of color combinations and material combinations were created, but additional texture combinations, thickness combinations, pattern combinations, and size combinations may be stored in the styling combination database, and in the above example, the items and categories whose categories are tops Although we only created examples of combinations of items with the category of bottoms, there are more combinations of items with the category of tops, items with the category of bottoms, items with the category of jewelry, items with the category of shoes, items with the category of perfume, and items with other categories. It may be stored.

That is, the device 200 can arbitrarily select one category, can arbitrarily select one of the selected items included in the selected category as a representative item, and provide information and styling of the representative item of the randomly selected category. Through the combination database, one of the selected items for each category can be selected and extracted as a representative item.

For example, the items selected as selection items include A sale item, B sale item, C sale item, D sale item, E sale item, F sale item, G sale item, H sale item, I sale item, and A sale item. Sale items, B sale items, C sale items, and D sale items were classified as tops items, E sale items, and F sale items were classified as bottoms items, G sale items, H sale items, and I If the sale item is classified as an item whose category is shoes, the device 200 can select the top by randomly selecting one of tops, bottoms, and shoes, and the A sale item and B sale item classified as an item whose category is tops. , one of the C sale items and the D sale items can be randomly selected to select the A sale item as the representative item of the tops category, and the device 200 may select the A sale item as the representative item of the bottoms category through the information and styling combination database of the A sale item. An item can be selected as an F sales item, and a representative item in the shoe category can be selected as a G sales item.

In step S903, the device 200 may generate recommended styling by combining representative items extracted for each category.

Specifically, the device 200 can generate recommended styling by combining representative items extracted for each category, and provide the generated recommended styling to the user's terminal 110 .

For example, the items selected as selection items include A sale item, B sale item, C sale item, D sale item, E sale item, F sale item, G sale item, H sale item, I sale item, and A sale item. Sale items, B sale items, C sale items, and D sale items were classified as tops items, E sale items, and F sale items were classified as bottoms items, G sale items, H sale items, and I If the sale item is classified as an item whose category is shoes, the device 200 can select the top by randomly selecting one of tops, bottoms, and shoes, and the A sale item and B sale item classified as an item whose category is tops. , one of the C sale items and the D sale items can be randomly selected to select the A sale item as the representative item of the tops category, and the device 200 may select the A sale item as the representative item of the bottoms category through the information and styling combination database of the A sale item. An item can be selected as an F sales item, and a representative item in the shoe category can be selected as a G sales item. Additionally, the device 200 can generate recommended styling by combining representative items extracted for each category, that is, A-sale items, F-sale items, and G-sale items, and combining A-sale items, F-sale items, and G-sale items. The generated recommended styling can be provided to the user's terminal 110.

Meanwhile, if the device 200 confirms that the number of items selected as the user's items is not greater than the standard number and does not select a selection item by selecting the user's recommended items, the device 200 selects the user instead of the selection items. Recommended styling may be generated based on the recommended item, and recommended styling generated based on the user's recommended item may be provided to the user's terminal 110.

Through this, the device 200 can generate recommended styling combined with items suitable for the user and provide the generated recommended styling to the user.

Figure 10 is a flow chart to explain the process of selecting recommended sales items and providing them to the seller according to an embodiment.

Referring to FIG. 10, first, in step S1001, the device 200 may select a recommended sale item based on the history of items being sold through the platform.

Specifically, the device 200 can check the user's history of purchasing items through the platform, that is, the history of items being sold through the platform, and select recommended sales items based on the history of items being sold through the platform. can do. At this time, the process of selecting a recommended sales item based on the history of item sales through the platform will be described with reference to FIG. 11.

In step S1002, the device 200 may transmit information on the recommended sales item to the seller's terminal. At this time, the device 200 can communicate wired or wirelessly with the seller's terminal, where the seller may be a seller who sells items through a platform.

Specifically, the device 200 may transmit information on a recommended sales item selected based on the item's sales history through the platform to the seller's terminal. At this time, the information on the recommended sales item may include the name of the recommended sales item, the brand of the recommended sales item, the category of the recommended sales item, and the color of the recommended sales item.

Due to this, the device 200 may recommend an item suitable for the user among the items for sale on the platform to a user who purchases an item through the platform, or among new items to be sold on the platform to a seller who sells an item through the platform. You can recommend items that are likely to sell well. In other words, the device 200 has the effect of providing necessary information to not only the user purchasing the item but also the seller selling the item.

Figure 11 is a flowchart for explaining the process of selecting a recommended sales item according to an embodiment.

Referring to FIG. 11 , in step S1101, the device 200 may obtain a sales history, which is a history of items being sold through the platform during a first period. Here, the first period is a preset period and may vary depending on the embodiment.

Specifically, the device 200 may obtain a sales history, which is a history of items being sold through the platform during a first period. To this end, when a user purchases an item through the platform, the device 200 may match the user's information and the information of the item the user purchased and store the information in a database provided in the device 200. That is, the device 200 may obtain sales history, which is the history of items being sold through the platform during the first period, through the database.

In step S1102, the device 200 may check the popular category, which is the most sold category, based on sales history.

Specifically, the device 200 checks information on items sold during the first period based on sales history, which is the history of items sold through the platform during the first period, and determines information on items sold during the first period. Through this, you can check the most sold category during the first period and select that category as a popular category.

In step S1103, based on sales history, the device 200 selects items that have sold more than the target number as popular items and can check the popular color, which is the color most included among the popular items. Here, the target number is a preset number and may vary depending on the embodiment.

Specifically, the device 200 checks information on items sold during the first period based on sales history, which is the history of items sold through the platform during the first period, and determines information on items sold during the first period. Through this, it is possible to check which items were sold more than the target number during the first period and select the items as popular items. Additionally, the device 200 can check information on the item selected as a popular item to check the color included most frequently among the colors of the popular item, and select the color included most often as the popular color.

In step S1104, the device 200 can check the most popular brand, which is the most sold brand, based on sales history.

Specifically, the device 200 checks information on items sold during the first period based on sales history, which is the history of items sold through the platform during the first period, and determines information on items sold during the first period. Through this, you can check the most sold brand during the first period and select that brand as a popular brand.

In step S1105, the device 200 may check the popular search keyword, which is the most included search keyword during the first period, through the platform's search engine.

Specifically, the device 200 may obtain search keywords used by users to search for products on the platform during the first period through the search engine of the platform, and determine the search keyword that contains the most of the obtained search keywords. Thus, the corresponding search keyword can be selected as a popular search keyword.

In step S1106, the device 200 may determine whether a new item that satisfies all of the popular categories, popular colors, popular brands, and popular search keywords exists.

Specifically, the device 200 can communicate with websites, including SNS, wired or wirelessly and access websites, and the device 200 can access websites of multiple brands to purchase new items released by the brands. In addition, the device 200 can check the name of the new item, the image of the new item, the category of the new item, the characteristics of the new item, the size of the new item, the color of the new item, and the new item confirmed through the brand's website. You can obtain information about new items, including the brand of the item. Here, the new item may be an item newly released by the brand during a preset critical period, or may be an item that has not been registered on the platform operated by the device 200.

That is, the device 200 determines whether the category of the new item is a popular category, the color of the new item is a popular color, or the brand of the new item is a popular brand, based on information about the new item acquired through the brand's website. You can determine whether a new item that satisfies everyone exists by checking whether the features of the new item include popular search keywords.

If it is confirmed in step S1106 that a new item that satisfies all of the popular categories, popular colors, popular brands, and popular search keywords exists as the first item, in step S1107, the device 200 selects the first item as a recommended sales item. You can.

Specifically, based on the information on the new item, the device 200 performs a popular search based on whether the category of the new item is a popular category, the color of the new item is a popular color, the brand of the new item is a popular brand, and the characteristics of the new item. As a result of checking whether the keyword is included, based on the information of the new item, the device 200 determines that the category of the new item is a popular category, the color of the new item is a popular color, the brand of the new item is a popular brand, and the new item's If it is confirmed that a new item that includes a popular search keyword in the features, that is, a new item that satisfies all of the popular brand, popular category, popular color, and popular search keyword exists as the first item, the device 200 sends the first item to the seller on the platform. You can select recommended sales items to recommend for sale through .

If it is determined in step S1106 that there is no new item that satisfies all of the popular categories, popular colors, popular brands, and popular search keywords, in step S1108, the device 200 determines a new item that satisfies at least one as a candidate sales item. You can.

Specifically, based on the information on the new item, the device 200 performs a popular search based on whether the category of the new item is a popular category, the color of the new item is a popular color, the brand of the new item is a popular brand, and the characteristics of the new item. As a result of checking whether keywords are included, the category of the new item is a popular category, the color of the new item is a popular color, the brand of the new item is a popular brand, and the features of the new item include a popular search keyword, that is, all are satisfied. If it is confirmed that a new item does not exist, the device 200 determines whether the category of the new item is a popular category, the color of the new item is a popular color, the brand of the new item is a popular brand, or a new item is selected based on the information of the new item. If the characteristics of the item include a popular search keyword, that is, a new item that satisfies at least one of a popular brand, a popular category, a popular color, and a popular search keyword can be checked and the new item can be selected as a candidate sales item.

In step S1109, the device 200 may generate a list including candidate sales items by sequentially sorting them in order of the number of items being satisfied.

Specifically, when a candidate sales item is selected based on the information on the new item, the device 200 selects several of the popular brands, popular categories, popular colors, and popular search keywords through the information on the candidate sales items. You can check whether you are satisfied with the number of candidate sales items that are satisfied. Additionally, the device 200 may generate a list including the candidate sales items by sorting them in order, starting with the candidate sales items with the highest number of satisfied items, based on the number of candidate sales items that are satisfied.

That is, based on the information on the new item, if there is a new item that satisfies all of the popular brands, popular categories, popular colors, and popular search keywords, the device 200 recommends the new item to the seller to sell through the platform. You can.

In addition, based on the information on the new item, if there is no new item that satisfies all of the popular brands, popular categories, popular colors, and popular search keywords, the device 200 selects the new items as popular brands, popular categories, popular colors, and popular search keywords. By checking how many of the search keywords are satisfied and sorting them in order, starting with the number of new items that are satisfied, a list is created and the created list is provided to the seller, who then selects new items to sell through the platform based on the list. You can select an item.

Meanwhile, the device 200 may receive a purchase request for the user's recommended item from the user's terminal 100, and upon receiving a purchase request for the user's recommended item from the user's terminal 100, the device 200 may send the user a purchase request for the user's recommended item. You can sell recommended items.

Figure 12 is a flow chart to explain the process of selling a user's recommended item to a user according to an embodiment.

Referring to FIG. 12 , first, in step S1201, the device 200 may obtain the user's purchase history of the same category as the user's recommended item, which is the history of the user purchasing items of the same category as the user's recommended item. Here, the same category purchase history is the history of when and how much the user has purchased items in the same category as the recommended item that the user wants to purchase in the past.

Specifically, when the device 200 receives a purchase request for a user's recommended item from the user's terminal 100, the device 200 may sell the user's recommended item to the user, and may store the user's recommended item in a warehouse to sell the user's recommended item to the user. In order to select an item to be delivered to the user among the user's recommended items provided in , the user's purchase history of the same category, which is the history of the user purchasing items of the same category as the user's recommended item in the past, can be obtained. At this time, the device 200 may obtain the user's purchase history of the same category as the item the user currently wants to purchase from the user's terminal 100, including when and how much the user purchased the same category of items, and the device 200 may Through the user database provided by 200, the user's purchase history of the same category, which is the history of purchasing items of the same category as the recommended item of the user that the user currently wants to purchase, may be obtained. The user database may store purchase history, which is the history of the user purchasing items in the past through the platform by matching the user. That is, the user database may store when and how much a user purchased an item of a certain name in a certain category in the past. Information about whether a purchase was made is stored.

That is, the device 200 obtains the user's purchase history of the same category, which is information about when and how much the user purchased items of the same category as the recommended item that the user currently wants to purchase, through the user's terminal 100 or the user database. can do.

In step S1202, the device 200 may create a purchase cycle of the same category based on the user's purchase history of the same category.

Specifically, the device 200 can check the date when the user purchased an item of the same category as the user's recommended item, based on the user's purchase history of the same category, and use the confirmed date to create a purchase cycle of the same category. You can.

That is, based on the user's purchase history of the same category, the device 200 can determine at what cycle the user repurchases items of the same category as the user's recommended item and generate a purchase cycle of the same category.

For example, if the user wants to purchase a first perfume, which is the user's recommended item, and the first perfume is included in the perfume category, the device 200 selects the first perfume and the category through the user's terminal 100 or the user database. It is possible to obtain the perfume purchase history of a user who has purchased the same perfume, and through the user's perfume purchase history, the date the perfume was purchased is January 20, 2023, March 20, 2023, and 5, 2023, respectively. You can see that it is the 10th of the month and the current date is July 20, 2023, and based on the purchase date, you can create a perfume purchase cycle of two months.

In step S1203, the device 200 may check the first date, which is the next purchase date of the item in the same category as the user's recommended item, based on the current date and the purchase cycle of the same category.

For example, if the user wants to purchase a first perfume, which is the user's recommended item, the first perfume is included in the perfume category, the current date is July 20, 2023, and the purchase cycle of the perfume is two months, the device (200) can confirm September 20, 2023 as the first date, which is the next purchase date of the item in the same category as the first perfume, which is the recommended item for the user to purchase.

In step S1204, the device 200 may determine whether there is a user-recommended item whose expiration date is the first date.

Specifically, the device 200 may obtain warehouse data from the seller's terminal, and the warehouse data may include information about which items are available in the seller's warehouse and how many. That is, the device 200 may determine, through warehouse data, whether there is a user-recommended item whose expiration date is the first date among the user-recommended items that the user wishes to purchase available in the warehouse.

If it is confirmed in step S1204 that the user's recommended item whose expiration date is the first date exists as the first recommended item, the device 200 may sell the first recommended item to the user in step S1205.

Specifically, the device 200 determines, through warehouse data, whether there is a user-recommended item whose expiration date is the first date among the user-recommended items that the user wishes to purchase, which are available in the warehouse. If it is confirmed that the user's recommended item whose expiration date is the first date among the user's recommended items that the user wishes to purchase exists as the first recommended item, the first recommended item may be sold to the user. At this time, the process of selling and delivering the first recommended item to the user has not been specifically written, but it can be sold by a commonly used selling method or a commonly used delivery method.

For example, if the user wants to purchase a first perfume, which is the user's recommended item, the first perfume is included in the perfume category, the current date is July 20, 2023, and the purchase cycle of the perfume is two months, the device (200) can confirm September 20, 2023 as the first date, which is the next purchase date of perfume, which is an item of the same category as the first perfume, which is the recommended item of the user that the user wants to purchase. Among the first perfumes available, check whether there is a first perfume with an expiration date of September 20, 2023, and if it is confirmed that the first perfume with an expiration date of September 20, 2023 is stocked in the warehouse as the 1-1 perfume. , perfume 1-1 with an expiration date of September 20, 2023 can be sold to users.

If it is confirmed in step S1204 that the user's recommended item whose expiration date is the first date does not exist, in step S1206, the device 200 generates second dates that fall before the preset range based on the first date. You can. Here, the preset range may vary depending on the embodiment, and the range may become wider depending on the purchase cycle of the same category as the purchase cycle of the same category becomes longer, and may become narrower as the purchase cycle of the same category becomes shorter.

Specifically, the device 200 determines, through warehouse data, whether there is a user-recommended item whose expiration date is the first date among the user-recommended items that the user wishes to purchase, which are available in the warehouse. If it is confirmed that among the user's recommended items that the user wishes to purchase in the warehouse, there is no user's recommended item whose expiration date is on the first date, it falls within the preset range from the first date. Second dates may be created.

For example, a user wants to purchase a first perfume, which is the user's recommended item, the first perfume is included in the perfume category, the current date is July 20, 2023, the perfume's purchase cycle is two months, and the preset If the range is 7 days, the device 200 may check September 20, 2023 as the first date, which is the next purchase date of the perfume, which is an item in the same category as the first perfume, which is the recommended item for the user that the user wants to purchase. , through warehouse data, check whether there is a first perfume with an expiration date of September 20, 2023 among the first perfumes in the warehouse, and determine that there is no first perfume with an expiration date of September 20, 2023 in the warehouse. If confirmed, the second dates within a preset range of 7 days based on the first date, September 20, 2023, are September 13, 2023, September 14, 2023, September 15, 2023, You can create September 16, 2023, September 17, 2023, September 18, 2023, and September 19, 2023.

In step S1207, the device 200 may determine whether there is a user-recommended item whose expiration date is the second date.

Specifically, the device 200 determines whether there is a user-recommended item whose expiration date is a second date that is one of the second dates among the user-recommended items that the user wishes to purchase stored in the warehouse through warehouse data. can be judged.

If it is confirmed in step S1207 that the user's recommended item whose expiration date is the second date exists as a second recommended item, the device 200 may sell the second recommended item to the user in step S1208.

Specifically, the device 200 determines, through warehouse data, whether there is a user-recommended item whose expiration date is the second date among the user-recommended items that the user wishes to purchase, which are available in the warehouse. If it is confirmed that the user's recommended item whose expiration date is the second date among the user's recommended items that the user wishes to purchase exists as a second recommended item, the second recommended item may be sold to the user.

For example, a user wants to purchase a first perfume, which is the user's recommended item, the first perfume is included in the perfume category, the current date is July 20, 2023, the perfume's purchase cycle is two months, and the preset If the range is 7 days, the device 200 may check September 20, 2023 as the first date, which is the next purchase date of the perfume, which is an item in the same category as the first perfume, which is the recommended item for the user that the user wants to purchase. , through warehouse data, check whether there is a first perfume with an expiration date of September 20, 2023 among the first perfumes in the warehouse, and determine that there is no first perfume with an expiration date of September 20, 2023 in the warehouse. If confirmed, the second dates within a preset range of 7 days based on the first date, September 20, 2023, are September 13, 2023, September 14, 2023, September 15, 2023, September 16, 2023, September 17, 2023, September 18, 2023, and September 19, 2023 can be created, and through warehouse data, the expiration date of the first perfume in the warehouse is 2023. One of September 13, September 14, 2023, September 15, 2023, September 16, 2023, September 17, 2023, September 18, 2023, September 19, 2023 1 Check whether perfume is present and the expiration date is September 13, 2023, September 14, 2023, September 15, 2023, September 16, 2023, September 17, 2023, September 2023 9 If it is confirmed that the first perfume is stocked in the warehouse as the 1-2 perfume, either on the 18th of September or the 19th of September 2023, the 1-2 perfume can be sold to the user. In this process, the device 200 selects the first perfume provided in the warehouse with an expiration date of September 13, 2023, September 14, 2023, September 15, 2023, September 16, 2023, and 2023. If it is confirmed that there are at least two first perfumes with expiration dates of September 17, 2023, September 18, 2023, and September 19, 2023, the expiration dates are September 13, 2023, September 14, 2023, and 2023. The first perfume with the earliest expiration date among the first perfumes of September 15, 2023, September 16, 2023, September 17, 2023, September 18, 2023, and September 19, 2023 can be provided to the user. However, it is not limited to this.

If it is confirmed in step S1207 that the user's recommended item whose expiration date is the second date does not exist, in step S1209, the device 200 determines whether the expiration date is limited among the user's recommended items that the user wants to purchase provided in the warehouse. The third recommended item, which is the user's recommended item with the earliest date after date 1, can be checked, and the third recommended item can be sold to the user.

Specifically, the device 200 determines, through warehouse data, whether there is a user-recommended item whose expiration date is the second date among the user-recommended items that the user wishes to purchase, which are available in the warehouse. If it is confirmed that there is no user-recommended item whose expiration date is on the second date among the recommended items that the user wants to purchase, the expiration date is on the first date among the recommended items that the user wants to purchase in the warehouse. The third recommended item, which is the user's recommended item with the earliest date after the date, can be checked, and the third recommended item can be sold to the user.

For example, a user wants to purchase a first perfume, which is the user's recommended item, the first perfume is included in the perfume category, the current date is July 20, 2023, the perfume's purchase cycle is two months, and the preset If the range is 7 days, the device 200 may check September 20, 2023 as the first date, which is the next purchase date of the perfume, which is an item in the same category as the first perfume, which is the recommended item for the user that the user wants to purchase. , through warehouse data, check whether there is a first perfume with an expiration date of September 20, 2023 among the first perfumes in the warehouse, and determine that there is no first perfume with an expiration date of September 20, 2023 in the warehouse. If confirmed, the second dates within a preset range of 7 days based on the first date, September 20, 2023, are September 13, 2023, September 14, 2023, September 15, 2023, September 16, 2023, September 17, 2023, September 18, 2023, and September 19, 2023 can be created, and through warehouse data, the expiration date of the first perfume in the warehouse is 2023. One of September 13, September 14, 2023, September 15, 2023, September 16, 2023, September 17, 2023, September 18, 2023, September 19, 2023 1 Check whether perfume is present, 13 September 2023, 14 September 2023, 15 September 2023, 16 September 2023, 17 September 2023, 18 September 2023 , if it is confirmed that the first perfume that is one of September 19, 2023 does not exist, the first perfume whose expiration date is the earliest of the dates after September 20, 2023 among the first perfumes in the warehouse By checking the 1-3 perfumes, the 1-3 perfumes can be sold to the user.

In addition, if the device 200 determines in this process that there is no user-recommended item whose expiration date is after the first date among the user-recommended items that the user wishes to purchase in the warehouse, the expiration date is before the first date. Among the user's recommended items, the user's recommended item with the longest expiration date can be sold to the user.

For this reason, the device 200 determines the date on which the user will next purchase an item in the same category as the user's recommended item through the purchase cycle of the item in the same category as the user's recommended item that the user wishes to purchase, that is, the same category purchase cycle. By confirming and providing the user's recommended items with an expiration date by that date, there is an effect of selling items with an expiration date suitable for the user.

Figure 13 is an exemplary diagram of the configuration of a device according to an embodiment.

Device 200 according to one embodiment includes a processor 210 and memory 220. The processor 210 may include at least one device described above with reference to FIGS. 1 to 12 or may perform at least one method described above with reference to FIGS. 1 to 12 . A person or organization using the device 200 may provide services related to some or all of the methods described above with reference to FIGS. 1 to 12 .

The memory 220 may store information related to the methods described above or store a program in which methods described later are implemented. Memory 220 may be volatile memory or non-volatile memory.

The processor 210 can execute programs and control the device 200. The code of the program executed by the processor 210 may be stored in the memory 220. The device 200 is connected to an external device (eg, a personal computer or a network) through an input/output device (not shown) and can exchange data through wired or wireless communication.

The embodiments described above may be implemented with 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, and a field programmable gate (FPGA). It may be implemented using one or more general-purpose or special-purpose computers, such as an array, programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions. A processing device may execute an operating system (OS) and one or more software applications that run on the operating system. Additionally, a processing device may access, store, manipulate, process, and generate data in response to the execution of software. For ease of understanding, a single processing device may be described as being used; however, those skilled in the art will understand that a processing device includes multiple processing elements and/or multiple types of processing elements. It can be seen that it may include. For example, a processing device may include a plurality of processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are possible.

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., singly or in combination. Program instructions recorded on the medium may be specially designed and configured for the embodiment or may be known and available to those skilled in the art of 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 optical media (magneto-optical media) and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, etc. Examples of program instructions include machine language code, such as that produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. 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 these, which may configure a processing unit to operate as desired, or may be processed independently or collectively. You can command the device. Software and/or data may be used on any type of machine, component, physical device, virtual equipment, computer storage medium or device to be interpreted by or to provide instructions or data to a processing device. , or may be permanently or temporarily embodied in a transmitted signal wave. Software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

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

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

Claims (3)

In a method of providing a user interaction-based styling recommendation and lifestyle item sales platform service performed by a device,
Obtaining the user's personal information including the user's name, the user's age, the user's gender, the user's occupation, and the user's body type from the user's terminal;
Obtaining the user's preference information including the user's preferred style and the user's preferred brand from the user's terminal;
Obtaining the user's equipment information including information on items the user has from the user's terminal;
Selecting a recommended item for the user from among items sold on the platform based on the user's personal information, the user's preference information, and the user's provided information; and
Generating recommendation information for the user's recommended item and providing it to the user's terminal,
Checking the number of items selected as recommended items by the user; and
Based on the fact that the number of items is greater than a preset standard number, selecting the user's recommended items to select a selection item further includes,
The step of selecting a selection item by selecting the user's recommended items is,
Inputting each image of an item sold on the platform into a related image extraction artificial neural network to generate an item-related image data set, and
A step of selecting a selection item using the item-related image data set,
The step of selecting a selection item using the item-related image data set is,
Based on the information provided by the user, obtaining an image of the item provided by the user, which is an image of the item provided by the user,
Obtaining a first image by inputting the image of the user's items into the related image extraction artificial neural network;
Obtaining a second image, which is a related image whose similarity to the first image is higher than a preset standard ratio among the related images stored in the item-related image data set, and
A step of selecting a recommended item matched with the second image as a selection item,
The step of selecting a selection item using the item-related image data set is,
Obtaining confirmed location information, which is location information of the user during a preset reference period, from the user's terminal,
Extracting a preset reference location among the confirmed location information into outing location information, which is location information outside a preset reference range,
Obtaining an outing time, which is a time corresponding to the outing location information,
Selecting the main outing time of the user based on the outing time,
Generating a daytime outing rate and a night outing rate based on the user's main outing time,
Comparing the daytime going out rate and the night going out rate,
When the rate of going out during the day is greater than the rate of going out at night, selecting a selection item by assigning weight to related images containing a first color more than a target rate among related images stored in the item-related image data set;
If the rate of going out at night is greater than the rate of going out during the day, selecting a selection item by assigning weight to related images containing a second color greater than the target rate among related images stored in the item-related image data set; and
When the rate of going out during the day and the rate of going out at night match, selecting a selection item by assigning weight to related images containing a third color greater than the target rate among related images stored in the item-related image data set. It further includes,
The step of selecting a selection item using the item-related image data set is,
Obtaining information on a visited brand, which is a brand visited by the user during the reference period, from the user's terminal,
Obtaining the brand logo of the visited brand based on the information of the visited brand,
Obtaining a third image by inputting the brand logo into the related image extraction artificial neural network,
Obtaining a fourth image, which is a related image whose similarity to the third image is higher than a preset standard ratio among the related images stored in the item-related image data set, and
Further comprising the step of selecting an item matching the fourth image as a selection item,
Method of providing user interaction-based styling recommendation and lifestyle item sales platform service. delete According to paragraph 1,
Combining the user's selected items to generate recommended styling, and providing the generated recommended styling to the user's terminal,
The step of generating recommended styling by combining the user's selected items is,
Classifying the selected items according to categories,
A step of selecting one of the selected items and extracting it as a representative item based on the color of the item, the texture of the item, the thickness of the item, the material of the item, the pattern of the item, and the size of the item for each category, and
Including the step of generating recommended styling by combining the representative items extracted for each category,
Method of providing user interaction-based styling recommendation and lifestyle item sales platform service.