KR102442097B1 - User equipment, method, and recording medium for creating recommendation products - Google Patents

Abstract

The present disclosure relates to an apparatus, method and recording medium for recommending a product, and a product recommendation apparatus, method, and method for recommending a product using an optimal product recommendation logic by selecting a specific product recommendation logic from among a plurality of product recommendation logic and recommending a product; A recording medium may be provided. In particular, it is possible to provide a product recommendation device capable of maximizing profits while minimizing search by increasing product exposure by the product recommendation logic with the best performance in real time.

Description

Product recommendation device, method and recording medium {USER EQUIPMENT, METHOD, AND RECORDING MEDIUM FOR CREATING RECOMMENDATION PRODUCTS}

The present embodiments provide a product recommendation apparatus, a method, and a recording medium.

The size of the e-commerce service (online shopping mall) market is also gradually expanding through the development of information and communication technology, the spread of smart phones, and the development of safe and convenient payment systems. Accordingly, as the amount of products and information handled by the e-commerce service increases, there is a problem that the user has difficulty in selecting a product or feels tired. In addition, the operator who sells the product is also making an effort to increase the probability of selling the product by inducing the purchase conversion of the customers. Accordingly, when a customer accesses a specific site, operators of the e-commerce service market present products through a “product recommendation” function.

However, as the competition to expand customer share in the e-commerce service market becomes more intense, the product recommendation algorithm for providing the “product recommendation” function includes various types of algorithms, from simple rules-based algorithms to algorithms optimized with complex formulas. There is this. Therefore, there is a need for a technology that selects and provides a product recommendation algorithm that is expected to perform high among various product recommendation algorithms in order to minimize the customer's purchase conversion loss and maximize the conversion opportunity.

Against this background, the present embodiments may provide a product recommendation apparatus, method, and recording medium for recommending a product by selecting an optimal product recommendation logic from among various product recommendation logic.

In order to achieve the above object, in one aspect, in the present embodiment, in a product recommendation device, a logic selection unit for selecting a specific product recommendation logic from among a plurality of product recommendation logic preset using a logic selection algorithm, a specific An update that selects a product using product recommendation logic, and updates the logic selection algorithm using feedback information including product exposure unit for exposing the selected product, product exposure number information, and product conversion number information It provides a product recommendation device comprising a unit.

In another aspect, in the present embodiment, in the product recommendation method, a logic selection step of selecting a specific product recommendation logic from among a plurality of product recommendation logic preset using a logic selection algorithm, a product is selected using a specific product recommendation logic and an update step of updating a logic selection algorithm using feedback information including a product exposure step of exposing the selected product and information on the number of times of exposure of the product and information on the number of times of conversion of the product. to provide.

In another aspect, in the present embodiment, in a recording medium recording a program for executing a product recommendation method, a logic selection function for selecting a specific product recommendation logic from among a plurality of product recommendation logic preset using a logic selection algorithm; A product exposure function that selects products using specific product recommendation logic and exposes the selected product, and an update function that updates the logic selection algorithm using feedback information including product exposure information and product conversion number information A program to implement is recorded and a computer-readable recording medium is provided.

According to the present embodiments, it is possible to provide a product recommendation apparatus, method, and recording medium for recommending a product by selecting an optimal product recommendation logic from among various product recommendation logic.

1 is a diagram exemplarily illustrating a system configuration to which the present disclosure can be applied.
2 is a diagram illustrating a configuration of a product recommendation apparatus according to an embodiment of the present disclosure.
3 is a flowchart illustrating an operation of selecting a specific product recommendation logic in a product recommendation device according to an embodiment of the present disclosure.
4 is a diagram illustrating an example for explaining a plurality of product recommendation logic in the product recommendation apparatus according to an embodiment of the present disclosure.
5 is a diagram illustrating an example for explaining an initial operation of a logic selection algorithm in a product recommendation apparatus according to an embodiment of the present disclosure.
6 is a diagram illustrating an example for explaining a sampling operation of a logic selection algorithm in the product recommendation apparatus according to an embodiment of the present disclosure.
7 is a diagram illustrating an example of an operation in which a logic selection algorithm selects a specific product recommendation logic in the product recommendation device according to an embodiment of the present disclosure.
8 is a diagram illustrating an example of an operation in which a logic selection algorithm selects a specific product recommendation logic in a product recommendation device according to another embodiment of the present disclosure.
9 is a diagram illustrating an example of an operation in which a logic selection algorithm selects a specific product recommendation logic in a product recommendation apparatus according to another embodiment of the present disclosure.
10 is a diagram illustrating an example of an operation of selecting a specific product recommendation logic and exposing a product in the product recommendation device according to an embodiment of the present disclosure.
11 is a flowchart of a product recommendation method according to another embodiment of the present disclosure.

The present disclosure relates to a product recommendation apparatus, a method, and a recording medium.

Hereinafter, some embodiments of the present disclosure will be described in detail with reference to exemplary drawings. In adding reference numerals to components of each drawing, the same components may have the same reference numerals as much as possible even though they are indicated in different drawings. In addition, in describing the present embodiments, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present technical idea, the detailed description may be omitted. When "includes", "having", "consisting of", etc. mentioned in this specification are used, other parts may be added unless "only" is used. When a component is expressed in a singular, it may include a case in which the plural is included unless otherwise explicitly stated.

In addition, in describing the components of the present disclosure, terms such as first, second, A, B, (a), (b), etc. may be used. These terms are only for distinguishing the elements from other elements, and the essence, order, order, or number of the elements are not limited by the terms.

In the description of the positional relationship of components, when two or more components are described as being "connected", "combined" or "connected", two or more components are directly "connected", "coupled" or "connected" ", but it will be understood that two or more components and other components may be further "interposed" and "connected," "coupled," or "connected." Here, other components may be included in one or more of two or more components that are “connected”, “coupled” or “connected” to each other.

In the description of the temporal flow relation related to the components, the operation method or the manufacturing method, for example, the temporal precedence relationship such as "after", "after", "after", "before", etc. Alternatively, when a flow precedence relationship is described, it may include a case where it is not continuous unless "immediately" or "directly" is used.

On the other hand, when numerical values or corresponding information (eg, level, etc.) for a component are mentioned, even if there is no explicit description separately, the numerical value or the corresponding information is based on various factors (eg, process factors, internal or external shock, It may be interpreted as including an error range that may be caused by noise, etc.).

Hereinafter, the present disclosure will be described in detail with reference to the accompanying drawings.

1 is a diagram exemplarily illustrating a system configuration to which the present disclosure can be applied.

Referring to FIG. 1 , the present disclosure relates to a system for providing a product recommendation method, and may be implemented in the product recommendation device 110 and the database 100 .

The product recommendation device 110 may include a general PC such as a general desktop or notebook computer, and may include a mobile terminal such as a smart phone, a tablet PC, a personal digital assistant (PDA) and a mobile communication terminal, and the like. However, it should be broadly interpreted as any electronic device capable of communicating with the database 100 .

The product recommendation device 110 has the same configuration as a typical web server (Web Server), web application server (Web Application Server), or web server (WAP Server) in terms of hardware. However, in terms of software, as will be described in detail below, it includes program modules that are implemented through any language such as C, C++, Java, PHP, .Net, Python, and Ruby and perform various functions. can do.

In addition, the product recommendation device 110 is a server program that is provided in various ways according to the operating system such as DOS, Windows, Linux, UNIX, Macintosh on general server hardware. It can be implemented using a web site (Website) used in a Windows environment, IIS (Internet Information Server) used in a Windows environment, Apache, Nginx, Light HTTP, etc. used in a Unix environment can be used.

The database 100 may be connected to an unspecified number of clients (including the device 110) and/or other servers through a network. Accordingly, the database 100 receives a request for performing a task from the client or other server and receives the request for it. It may refer to a computer system that derives and provides work results or computer software (server program) installed for such a computer system.

In addition, the database 100 is understood as a broad concept including a series of application programs operating on the database 100 and, in some cases, various databases built inside or outside, in addition to the above-described server program. it should be

Here, the database 100 may mean a collection of data in which data such as information or data is structured and managed for the purpose of being used by a server or other device, and may mean a storage medium for storing the aggregate of data. have.

In addition, the database 100 may include a plurality of databases classified according to a data structure method, a management method, a type, and the like. In some cases, the database 100 may include a database management system (DBMS), which is software that allows information or data to be added, modified, deleted, and the like.

Meanwhile, the network 120 is a network that connects the database 100 and the product recommendation device 110 , and may be a closed network 120 such as a local area network (LAN) or a wide area network (WAN). However, it may be an open network 120 such as the Internet. Here, the Internet includes the TCP/IP protocol and various services existing in its upper layers, namely HTTP (HyperText Transfer Protocol), Telnet, FTP (File Transfer Protocol), DNS (Domain Name System), SMTP (Simple Mail Transfer Protocol), It refers to a worldwide open computer network structure that provides Simple Network Management Protocol (SNMP), Network File Service (NFS), and Network Information Service (NIS).

The product recommendation apparatus, method, and recording medium according to an embodiment of the present disclosure briefly described above will be described in more detail below.

2 is a diagram illustrating a configuration of a product recommendation apparatus according to an embodiment of the present disclosure.

Referring to FIG. 2 , although the product recommendation device 110 providing the product recommendation method according to an embodiment of the present disclosure has been described on the premise that it is implemented in a server, the product recommendation device 110 may be a server or a terminal. and hardware for the product recommendation device 110 may not be limited. For example, the product recommendation device 110 selects a product using a logic selector 210 that selects a specific product recommendation logic from among a plurality of product recommendation logic preset using a logic selection algorithm, and selects a product using a specific product recommendation logic. , a product exposure unit 220 for exposing the selected product, and an update unit 230 for updating the logic selection algorithm using feedback information including information on the number of times of exposure of the product and information on the number of times of conversion of the product. can

The logic selection unit 210 may select a specific product recommendation logic from among a plurality of product recommendation logic preset by using a logic selection algorithm. For example, the logic selection unit 210 selects a specific product recommendation logic that maximizes conversion while minimizing the search for a plurality of product recommendation logic through an Exploration and Exploitation process using a logic selection algorithm. can be utilized. Here, details on a plurality of preset product recommendation logics will be described later with reference to FIG. 5 .

The logic selection algorithm may form a distribution having a random variable based on the exposure number information and the conversion number information, and select a specific product recommendation logic from among a plurality of product recommendation logic through sampling based on the distribution. For example, the exposure number information may mean that when a preset number of recommended products are disclosed and exposed at a specific location on the screen, the number of exposures of each of the exposed products recommended by specific product recommendation logics is counted. Also, the conversion number information may mean that the number of times the specific action has occurred is counted when a specific action designated according to a preset conversion goal matches the user action. As a specific example, the logic selection algorithm may form a distribution having a binomial random variable using information on the number of exposures and information on the number of conversions. The distribution can be changed through exposure information and conversion number information, and since the random variable is a real value between 0 and 1, it can be a distribution that can represent the reliability of the conversion rate. However, if the formed distribution has random variables such as ^t distribution, chi-squared distribution, F distribution, beta distribution, and Gamma distribution, , but not limited thereto.

The logic selection algorithm may select a specific product recommendation logic in which a maximum conversion ratio occurs among a plurality of product recommendation logic through sampling based on distribution. However, if there are a plurality of selected specific product recommendation logic, the exposure ratio by each specific product recommendation logic may be determined in proportion to the statistical value of the distribution.

In addition, the logic selection algorithm may select a specific product recommendation logic by comparing the increase in the number of times of exposure information versus the information on the number of conversions in the sampling result by distribution. For example, in the logic selection algorithm, even if there is a product recommendation logic corresponding to the convergence level, if the increase in the conversion number information is larger than the exposure number information of other specific product recommendation logic in the search process, the specific product recommendation logic is changed. may be selected.

In addition, when information on the number of exposures and information on the number of conversions is accumulated according to an increase in traffic, the logic selection algorithm may select a product recommendation logic in which a sampling result by distribution converges to a specific value as the specific product recommendation logic. For example, in the logic selection algorithm, the more the user's action value, the more clearly the distribution is formed, and the probability that the distribution shape converges with respect to a specific product recommendation logic may increase. Details of the logic selection algorithm will be described later with reference to FIGS. 5 to 9 .

The product exposure unit 220 may select a product by using the selected specific product recommendation logic and expose the selected product. For example, when an access request signal for requesting access to a specific page providing product recommendation is input from the user, the product exposure unit 220 may expose the selected product using the selected specific product recommendation logic. For example, the access request signal may be a request signal for requesting a specific web page through a browser of a device connected to the corresponding web site. This is to provide a product recommendation function to a user accessing a preset specific page without a separate request from the user. Alternatively, when a recommendation request signal for requesting product recommendation is input from the user, the product exposure unit 220 may expose the selected product using the selected specific product recommendation logic. For example, the recommendation request signal may be a signal generated by accessing a specific menu requesting product recommendation or clicking a button. This is to provide a product recommendation function according to a user's product recommendation request.

The update unit 230 may update the logic selection algorithm by using feedback information including information on the number of times of exposure of the product and information on the number of times of conversion of the product. For example, the updater 230 may designate an action of clicking or ordering an exposed product according to a preset conversion target as a conversion standard, and may determine that the conversion is performed when the designated click or ordering action occurs. Accordingly, when the exposed product is clicked or ordered, the update unit 230 may determine that the product recommendation logic for recommending the product is converted.

However, the update unit 230 may set a conversion target differently according to a performance to be obtained through product recommendation, and may also set a conversion standard designated according to a preset conversion target differently. For example, the update unit 230 may set a conversion target according to whether the performance that the operator wants to obtain through product recommendation is to increase product views (product exposure, click), purchase conversion, or the like. That is, the criteria for determining that the conversion is performed can be variously set, such as conversion of membership, adding to shopping cart, etc. according to the conversion goal, in addition to clicking or ordering. Through this, automatic optimization of the product recommendation logic that meets the preset conversion goal may be performed.

When a product is exposed, the update unit 230 may update the exposure number information by increasing the exposure number of a specific product recommendation logic for recommending the exposed product by one. Also, when the exposed product is converted, the update unit 230 may update the conversion number information by increasing the number of conversions of the specific product recommendation logic for recommending the converted product by one. However, the update unit 230 may set different increases in the number of exposures and the number of conversions according to the conversion target, and may also set different increases in the number of conversions according to clicks and ordering actions.

3 is a flowchart illustrating an operation of selecting a specific product recommendation logic in a product recommendation device according to an embodiment of the present disclosure.

Referring to FIG. 3 , the logic selection unit 210 of the product recommendation apparatus according to an embodiment of the present disclosure may select a specific product recommendation logic ( S310 ). For example, the logic selection unit 210 may select a specific product recommendation logic from among a plurality of product recommendation logic preset using a logic selection algorithm to perform real-time optimization. Here, the logic selection algorithm may be to select a specific product recommendation logic through sampling by applying the product recommendation result to a distribution having a random variable.

The product exposure unit 220 may select a product using the selected specific product recommendation logic and expose the selected product ( S320 ). For example, the product exposing unit 220 may select and expose a preset number of products in a designated area of the screen. Also, the product exposing unit 220 may select and expose a product by selecting a specific product recommendation logic by a preset number of products in order to expose recommended products as much as a preset number of products. However, when there are a plurality of selected specific product recommendation logic, the product exposure unit 220 adjusts the number of products selected by each specific product recommendation logic based on the preset number of products, or adjusts each product posting time. The selected product can be exposed by applying the exposure ratio determined by the method or the like.

The update unit 230 may update the feedback information including the exposure number information and the conversion number information ( S330 ). For example, the update unit 230 may update the feedback information including the exposure number information and the conversion number information to reflect the recommendation result of the selected product recommendation logic. For example, when a user accesses a specific page, the update unit 230 exposes a recommended product using a specific product recommendation logic, and updates the exposure number information by increasing the number of exposures for the specific product recommendation algorithm by 1. can do. Also, when a click occurs on the exposed product, the update unit 230 may update the conversion number information by increasing the number of conversions of the specific product recommendation logic that recommends the clicked product by one. Of course, when an order is generated for the clicked product, the update unit 230 may update the conversion number information by incrementing the number of conversions of the specific product recommendation logic that recommends the ordered product again by one.

The update unit 230 may update the logic selection algorithm by using the feedback information (S340). As an example, the updater 230 may update the logic selection algorithm by reshaping a distribution having a random variable according to the updated information on the number of exposures and information on the number of conversions. Therefore, the update unit 230 continuously exposes products using the selected product recommendation logic and updates and uses the logic selection algorithm by observing the results, thereby providing an effect of quickly applying an online reaction.

The logic selection unit 210 may reselect a specific product recommendation logic using the updated logic selection algorithm (S350). For example, the logic selection unit 210 exposes a product recommended by a specific product recommendation logic selected from a plurality of product recommendation logic using a logic selection algorithm, and selects a product recommendation logic with the best conversion conversion rate according to the exposure to a specific product It can be reselected by recommendation logic. This may provide an effect of automatically changing a specific product recommendation logic for recommending a product to an optimal product recommendation logic by reflecting the exposure and conversion results.

For example, the logic selection unit 210 may repeat the process of updating the logic selection algorithm by reflecting the product recommendation result selected by the specific product recommendation logic and reselecting the specific product recommendation logic again. Also, when the reliability estimated using the Bayesian method corresponds to 95% or satisfies a preset condition, the logic selector 210 may end the process of reselecting the specific product recommendation logic.

4 is a diagram illustrating an example for explaining a plurality of product recommendation logic in the product recommendation apparatus according to an embodiment of the present disclosure.

Referring to FIG. 4 , a plurality of product recommendation logics that can be used to recommend products to a user in the logic selector 210 may be described.

For example, the product recommendation logic of the logic selector 210 may be configured in an artificial intelligence recommendation method and a statistical recommendation method. For example, the product recommendation logic of the artificial intelligence recommendation method may be a product recommendation based on deep learning, a collaborative filtering recommendation, a clustering recommendation, a clustering recommendation based on natural language processing, a personalized recommendation, or a search term recommendation logic. Here, the deep learning-based product recommendation logic automatically learns features of the input value by learning a large amount of data from a deep neural network consisting of a multi-layered network, and through this, the objective function, that is, the prediction accuracy It can be configured in the form of learning the network to minimize the error of . For a specific example, CNN (Convolutional Neural Network), DHN (Deep Hierachical Network), CDBN (Convolutional Deep Belief Network), DDN (Deconvolutional Deep Network), RNN (Recurrent Neural Network), GAN (Generative Adversarial Network), etc. And, the present disclosure is not limited thereto and may be composed of various deep learning algorithms that can be used now or in the future.

As another example, the product recommendation logic of the statistical recommendation method may be divided based on the data inquiry period. For example, the logic selector 210 selects a product recommendation logic of a statistical type recommendation method that recommends a product using statistical data based on the frequency of product viewing, clicking, or purchase of users per day (24h). can For a specific example, the product recommendation logic of the logic selector 210 is based on products viewed a lot, a lot of products, a lot of purchases, an order conversion rate, a recommended inflow rate, a recommended click-through rate, or a high discount rate by users per day (24h). statistical data can be used. For another example, the logic selection unit 210 reflects the aggregation result for the last one hour, and uses the product that users have purchased a lot in real time or the product that has been viewed a lot in real time to recommend a product recommendation logic of a statistical type recommendation method. Can be used. However, the last hour reflecting the counting result is an example and is not limited thereto.

Also, the logic selector 210 may select a specific product recommendation logic from among a plurality of product recommendation logics by using a separate logic selection algorithm. Here, the logic selection algorithm may be a separate algorithm for selecting a specific product recommendation logic by analyzing and testing customer performance for each of a plurality of product recommendation logics.

5 is a diagram illustrating an example for explaining an initial operation of a logic selection algorithm in a product recommendation apparatus according to an embodiment of the present disclosure.

Referring to FIG. 5 , an initial operation of a logic selection algorithm for the logic selection unit 210 to select a specific product recommendation logic from among a plurality of product recommendation logics may be described. As an example, the logic selection algorithm may form a continuous probability distribution defined in the [0,1] section according to parameters in order to select a specific product recommendation logic, and may perform sampling by the distribution. In this case, the parameter may be set using information on the number of exposures and information on the number of conversions. In addition, the X-axis of the formed distribution may be a conversion rate, the Y-axis may be a likelihood, and Recommendations 1,2,3 may mean product recommendation logic 1,2,3.

The initial operation of the logic selection algorithm may select a specific product recommendation logic in an initial state in which there is no existing accumulated result data. For example, the logic selection algorithm in the initial state can form a uniform distribution in which all probabilities from 0 to 1 are uniform when both parameters are 1. This may mean that both the exposure number information and the conversion number information are 0.

6 is a diagram illustrating an example for explaining a sampling operation of a logic selection algorithm in the product recommendation apparatus according to an embodiment of the present disclosure.

Referring to FIG. 6 , the sampling operation by the distribution of the logic selection algorithm in the logic selection unit 210 may be described. As an example, the logic selection algorithm may extract a sampling value by distribution as a changed value each time it is executed. This is because the sampling value is a value extracted from the distribution that continuously reflects the result. Therefore, the logic selection algorithm may change the product recommendation logic selected by sampling before the distribution converges to a specific value. For example, the shape of the formed distribution may have a maximum value in an expected value of a conversion rate. In addition, it can be confirmed that the sampling value is highly likely to be extracted near the average value of the distribution. Accordingly, the logic selection algorithm may select the product recommendation logic based on the sampling value extracted using the expected value of the conversion rate.

As a specific example, the logic selection algorithm may form a distribution based on the exposure number information and the conversion number information. In addition, the logic selection algorithm may confirm from the formed distribution that the sampling value of the product recommendation logic 1 (recommendation 1) is greater than that of the other product recommendation logics 2 and 3 (recommendation 2,3). Therefore, the logic selection algorithm may select the product recommendation logic 1 having the largest sampling value as the specific product recommendation logic.

7 is a diagram illustrating an example for explaining an operation in which a logic selection algorithm selects a specific product recommendation logic in the product recommendation device according to an embodiment of the present disclosure.

Referring to FIG. 7 , an operation in which the logic selection algorithm in the logic selection unit 210 selects a specific product recommendation logic through sampling by distribution may be described. As an example, the logic selection algorithm is a method of continuously updating information on the number of times of exposure and information on the number of conversions of a selected product using each of a plurality of product recommendation logic, and is optimal through the process of Exploration and Exploitation. of product recommendation logic can be selected. In addition, the logic selection algorithm can form a distribution that converges to a specific value by repeatedly performing exploration and exploitation processes. In this case, even if the logic selection algorithm repeatedly extracts the sampling value, it can be confirmed that only the product recommendation logic in which the probability distribution prevails is selected as the specific product recommendation logic.

When a plurality of specific product recommendation logics are selected, the logic selection algorithm of the logic selection unit 210 may determine an exposure ratio of each product recommendation logic in proportion to a sampling result extracted from the formed distribution. As a specific example, in the logic selection algorithm, the sampling value of the product recommendation logic 3 (recommendation 3) is the largest value from the formed distribution, and the ratio of the sampling values of each product recommendation logic may also be confirmed. Therefore, the logic selection algorithm selects product recommendation logic 3 as the optimal product recommendation logic, but product recommendation logic 1 and 2 may also be selected. In this case, the logic selection algorithm may determine the exposure ratio of the products selected by each product recommendation logic by using the ratio of the sampling values.

8 is a diagram illustrating an example of an operation in which a logic selection algorithm selects a specific product recommendation logic in a product recommendation device according to another embodiment of the present disclosure.

Referring to FIG. 8 , an operation in which the logic selection algorithm in the logic selection unit 210 selects a specific product recommendation logic through sampling by distribution may be described. For example, the logic selection algorithm may select a specific product recommendation logic based on a sampling value extracted using an expected value of a conversion rate. However, the logic selection algorithm may select a specific product recommendation logic by comparing the increase in the number of conversion information to the exposure number information.

As a specific example, the logic selection algorithm may determine from the formed distribution that the sampling value of the product recommendation logic 3 is greater than that of the other product recommendation logics 1 and 2 (recommendation 1 and 2). Therefore, the logic selection algorithm may select the product recommendation logic 3 having the largest sampling value as the specific product recommendation logic. However, in the logic selection algorithm, it can be seen that the expected value of the conversion rate is greatly changed in the distribution of the product recommendation logic 2 (recommendation 2). Accordingly, the logic selection algorithm may select the product recommendation logic 2 as a specific product recommendation logic in addition to the product recommendation logic 3 .

9 is a diagram illustrating an example for explaining an operation in which a logic selection algorithm selects a specific product recommendation logic in a product recommendation apparatus according to another embodiment of the present disclosure.

Referring to FIG. 9 , an operation in which the logic selection algorithm in the logic selection unit 210 selects a specific product recommendation logic through sampling by distribution may be described. As an example, the logic selection algorithm may select, as the specific product recommendation logic, a product recommendation logic in which a sampling result by distribution converges to a specific value when exposure number information and conversion number information are accumulated according to an increase in traffic. For example, in the form of the formed distribution, as the exposure number information and the conversion number information increase, the dispersion of the distribution may be reduced. And it can be confirmed that the formed distribution has a difference in variance even though the expected value of the conversion rate is the same. Accordingly, when the logic selection algorithm reflects more results on the user's behavior due to the increase in traffic, the criteria for selecting a specific product recommendation logic may become clear.

As a specific example, the logic selection algorithm may clearly confirm that the product recommendation logic 3 with the best response is the product recommendation logic 3 from the formed distribution, and may select the product recommendation logic 3 as the specific product recommendation logic.

10 is a diagram illustrating an example of selecting a specific product recommendation logic and exposing a product in the product recommendation device according to an embodiment of the present disclosure.

Referring to FIG. 10 , the logic selection unit 210 of the product recommendation device according to an embodiment of the present disclosure selects a specific product recommendation logic to provide a recommended product when a signal requesting access to a specific page is input from the user. can be (S1010). For example, the logic selection unit 210 may select a specific product recommendation logic having the highest sampling value generated using a specific function in order to select the first exposed product. In addition, the logic selection unit 210 may select a specific product recommendation logic having the highest sampling value in a state before the result application even when a situation arises in which the next recommended product must be provided before the product exposure result is generated. Here, the specific function is not limited thereto as long as it is a function for extracting a sampling value from a distribution, such as a beta function, a gamma function, and a Weibull function.

The product exposure unit 220 may expose the selected product using a specific product recommendation logic (S1020). For example, the product exposure unit 220 may expose the selected product using a specific product recommendation logic having the highest sampling value.

When a result is generated as a result of exposing the selected product, the update unit 230 may update feedback information including information on the number of exposures of the product and information on the number of conversions ( S1030 ). For example, the update unit 230 may update the feedback information to select product recommendation logic using a sampling value generated using a specific function in a state in which product exposure and exposure results are reflected.

The logic selection unit 210 may select a specific product recommendation logic through sampling based on distribution (S1040). For example, the logic selection unit 210 may select a specific product recommendation logic by forming and sampling a distribution using the updated feedback information. Also, the logic selector 210 may reselect a specific product recommendation logic as a sampling value in a state in which product exposure and exposure results are reflected. In addition, the product exposure unit 220 may provide the product to the user by exposing the product selected by the reselected specific product recommendation logic.

Hereinafter, a product recommendation method that can be performed by the product recommendation apparatus described with reference to FIGS. 1 to 10 will be described.

11 is a flowchart of a product recommendation method according to another embodiment of the present disclosure.

Referring to FIG. 11 , the product recommendation method of the present disclosure may include a logic selection step of selecting a specific product recommendation logic ( S1110 ). As an example, the product recommendation device may select a specific product recommendation logic from among a plurality of product recommendation logic preset by using a logic selection algorithm. For example, the product recommendation device may use a logic selection algorithm to select a specific product recommendation logic that maximizes conversion while minimizing the search for a plurality of product recommendation logic through the process of Exploration and Exploitation. .

The logic selection algorithm used in the product recommendation device forms a distribution with a random variable based on the exposure number information and the conversion number information, and recommends a specific product among a plurality of product recommendation logic through sampling based on the distribution You can choose logic. For example, the exposure number information may mean that when a preset number of recommended products are exposed at a specific location on the screen, the number of exposures of specific product recommendation logics that recommend each of the exposed products is counted. In addition, the conversion number information may refer to counting the number of times that a specified user's action has occurred according to a preset conversion goal. As a specific example, the logic selection algorithm may form a distribution having a binomial random variable using information on the number of exposures and information on the number of conversions. Here, the distribution can change the shape of the distribution through the exposure number information and the conversion number information, and since the random variable is a real value between 0 and 1, it can be a distribution that can represent the reliability of the conversion rate.

The logic selection algorithm may select a specific product recommendation logic in which a maximum conversion ratio occurs among a plurality of product recommendation logic through sampling based on distribution. However, if there are a plurality of selected specific product recommendation logic, the exposure ratio by each specific product recommendation logic may be determined in proportion to the sampling result based on the distribution.

In addition, the logic selection algorithm may select a specific product recommendation logic by comparing the increase in the number of times of exposure information versus the information on the number of conversions in the sampling result by distribution. For example, in the logic selection algorithm, even if there is a product recommendation logic corresponding to the convergence level, if the increase in the conversion number information is larger than the exposure number information of other specific product recommendation logic in the search process, the specific product recommendation logic is changed. may be selected.

In addition, when information on the number of exposures and information on the number of conversions is accumulated according to an increase in traffic, the logic selection algorithm may select a product recommendation logic in which a sampling result by distribution converges to a specific value as the specific product recommendation logic. For example, in the logic selection algorithm, the more the user's action value, the more clearly the distribution is formed, and the possibility that the distribution shape converges with respect to a specific product recommendation logic may increase.

The product recommendation method may include a product exposure step of exposing a recommended product (S1120). For example, the product recommendation apparatus may select a product by using the selected specific product recommendation logic and expose the selected product. For example, when an access request signal for requesting access to a specific page providing product recommendation is input from the user, the product recommendation apparatus may expose the selected product using the selected specific product recommendation logic. Here, the access request signal may be a request signal for requesting a specific web page through a browser of a device connected to the corresponding web site. This is to provide a product recommendation to a user accessing a preset specific page without a separate request from the user. As another example, when a recommendation request signal for requesting product recommendation is input from the user, the product recommendation apparatus may expose the selected product using the selected specific product recommendation logic. Here, the recommendation request signal may be a signal generated by accessing a specific menu requesting product recommendation or clicking a button. This is to provide product recommendation according to the user's product recommendation request.

The product recommendation method may include an update step of updating the logic selection algorithm (S1130). For example, the product recommendation apparatus may update the logic selection algorithm by using feedback information including information on the number of times of exposure of the product and information on the number of times of conversion of the product. For example, the product recommendation apparatus may designate an action of clicking or ordering an exposed product according to a preset conversion target as a conversion, and may determine that the product has been converted when the designated click or ordering action occurs. That is, when the exposed product is clicked or ordered, the product recommendation device may determine that the product recommendation logic for recommending the product is converted. However, the product recommendation device may set a different conversion according to the conversion goal, and may set variously other than click or order actions, such as member registration conversion and shopping cart addition actions.

As another example, when a product is exposed, the product recommendation device may update the exposure number information by increasing the exposure number of a specific product recommendation logic for recommending the exposed product by one. Also, when the exposed product is converted, the product recommendation apparatus may update the conversion number information by increasing the number of conversions of a specific product recommendation logic that recommends the converted product by one. As a specific example, when a user accesses a specific page, the product recommendation device may expose a recommended product using a specific product recommendation logic and increase the number of exposures for the specific product recommendation algorithm by 1. Also, when a click occurs on the exposed product, the product recommendation apparatus may increase the number of conversions of the specific product recommendation logic for recommending the clicked product by one. In addition, when an order is generated for the clicked product, the product recommendation device may increase the number of conversions of the specific product recommendation logic for recommending the ordered product by 1 again.

In the above, it has been described that the product recommendation method according to the embodiment of the present disclosure is performed with the same procedure as in FIG. 11, but this is only for convenience of description, and within the scope not departing from the essential concept of the present disclosure, the implementation method Accordingly, the procedure for performing each step may be changed, two or more steps may be integrated, or one step may be performed separately into two or more steps. In addition, the product recommendation method may be performed in a server or a terminal, and arbitrary steps may be separately performed in each of the server and the terminal.

The above-described product recommendation method may be implemented as a program and recorded in a computer-readable recording medium. In addition, the program for executing the above-described product recommendation method selects a product using a logic selection function for selecting a specific product recommendation logic from among a plurality of product recommendation logic preset using a logic selection algorithm, and a specific product recommendation logic and executes an update function for updating the logic selection algorithm by using the product exposure function for exposing the selected product and the feedback information including information on the number of exposures of the product and information on the number of times of conversion of the product.

In the above-described product recommendation method according to an embodiment of the present disclosure, an application basically installed in the product recommendation device 110 (this may be a program included in a platform basically mounted on the terminal, included in an operating system, etc., or a compatible program) In addition, the user is compatible with the operating system of the product recommendation device 110 through an application providing server, such as an application store server, an application or a web server related to the service, and directly installed in the product recommendation device 110 . It may also be executed by an application (ie, a program). Here, the operating system of the product recommendation device 110 is an operating system such as Windows, Macintosh, Linux, etc. installed in a general PC such as a desktop, or a mobile terminal such as a smartphone or tablet PC. It may be a mobile-only operating system such as iOS, Android, or Windows Mobile to be installed.

In this sense, the product recommendation method according to the embodiment of the present disclosure described above is implemented as an application (ie, a program) installed by default in the product recommendation device 110 or directly installed by the user, and the product recommendation device 110, etc. may be recorded on a computer-readable recording medium of

The program implementing the product recommendation method according to an embodiment of the present disclosure executes a logic selection function, a product exposure function, an update function, and the like. In addition, all functions corresponding to the product recommendation method according to the embodiment of the present disclosure described above with reference to FIGS. 1 to 11 may be executed.

Such a program is recorded on a computer-readable recording medium and executed by the computer, so that the above-described functions can be executed.

In this way, in order for the computer to read the program recorded on the recording medium and execute the product recommendation method implemented as the program, the above-described program is a computer such as C, C++, JAVA, machine language, etc. that can be read by the computer's processor (CPU). It may include code (Code) coded in the language.

Such code may include a function code related to a function defining the above-mentioned functions, etc., and may include an execution procedure related control code necessary for the processor of the computer to execute the above-mentioned functions according to a predetermined procedure.

In addition, this code may further include additional information necessary for the processor of the computer to execute the above-described functions or code related to memory reference for which location (address address) in the internal or external memory of the computer should be referenced. .

In addition, when the computer's processor needs to communicate with any other computer or database in a remote location in order to execute the above-mentioned functions, the code is transmitted by the computer's processor to the computer's communication module (eg, wired and/or wireless communication module). ) may be used to further include communication-related codes for how to communicate with any other computer or database located remotely, and what information or media to transmit and receive during communication.

And, in consideration of the system environment of a computer that reads the recording medium and executes the program, the functional program and related codes and code segments for implementing the present disclosure, programmers in the technical field to which the present disclosure belongs may be easily inferred or changed by

In addition, the computer-readable recording medium in which the above-described program is recorded is distributed in a computer system connected by a network, so that the computer-readable code can be stored and executed in a distributed manner. In this case, one or more of the plurality of distributed computers may execute some of the functions presented above, and transmit the result of the execution to one or more of the other distributed computers, and receive the result The computer may also execute some of the functions presented above and provide the results to other distributed computers as well.

As described above, the computer-readable recording medium recording the program for executing the product recommendation method according to the embodiment of the present disclosure is, for example, ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical media storage devices, and the like.

In addition, the computer-readable recording medium recording the application, which is a program for executing the product recommendation method according to an embodiment of the present disclosure, is an application store server (Application Store Server), an application or a web server related to the corresponding service (Web Server) ), etc., may be a storage medium (eg, hard disk, etc.) included in an application providing server (Application Provider Server), or the application providing server itself, may be another computer or a storage medium having recorded a program.

A computer capable of reading a recording medium recording an application, which is a program for executing the product recommendation method according to an embodiment of the present disclosure, is not only a general PC such as a general desktop or notebook computer, but also a smart phone, a tablet PC, and a personal digital assistant (PDA). Digital Assistants) and mobile terminals such as mobile communication terminals, as well as this, should be interpreted as all computing devices.

If a computer capable of reading a recording medium recording an application, which is a program for executing the product recommendation method according to an embodiment of the present disclosure, is a smart phone, a tablet PC, a personal digital assistant (PDA), and a mobile terminal such as a mobile communication terminal In this case, the mobile terminal may download and install the corresponding application from an application providing server including an application store server, a web server, etc. It may be installed in the terminal.

In the above, even though it has been described that all components constituting the embodiment of the present disclosure are combined or operated as one, the present disclosure is not necessarily limited to this embodiment. That is, within the scope of the present disclosure, all the components may operate by selectively combining one or more. In addition, although all the components may be implemented as one independent hardware, some or all of the components are selectively combined to perform some or all of the functions of one or a plurality of pieces of hardware. It may be implemented as a computer program having Codes and code segments constituting the computer program can be easily inferred by those skilled in the art of the present disclosure. Such a computer program is stored in a computer-readable storage medium (Computer Readable Media), read and executed by the computer, thereby implementing the embodiment of the present disclosure. The storage medium of the computer program may include a magnetic recording medium, an optical recording medium, and the like.

In addition, terms such as "comprises", "comprises" or "have" described above mean that the corresponding component may be embedded, unless otherwise stated, excluding other components. Rather, it should be construed as being able to include other components further. All terms, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs, unless otherwise defined. Commonly used terms, such as those defined in the dictionary, should be interpreted as being consistent with the meaning of the context of the related art, and should not be interpreted in an ideal or excessively formal meaning unless explicitly defined in the present disclosure.

The above description is merely illustrative of the technical spirit of the present disclosure, and various modifications and variations may be made by those of ordinary skill in the art to which the present disclosure belongs without departing from the essential characteristics of the present disclosure. Accordingly, the embodiments disclosed in the present disclosure are for explanation rather than limiting the technical spirit of the present disclosure, and the scope of the technical spirit of the present disclosure is not limited by these embodiments. The protection scope of the present disclosure should be interpreted by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present disclosure.

Claims (17)

a logic selection unit for selecting a specific product recommendation logic from among a plurality of product recommendation logic preset using a logic selection algorithm;
a product exposure unit that selects a product using the specific product recommendation logic and exposes the selected product; and
An update unit for updating the logic selection algorithm using feedback information including information on the number of times of exposure of the product and information on the number of times of conversion of the product,
The logic selection algorithm is
A distribution having a random variable is formed based on the exposure number information and the conversion number information, and the specific product logic is selected from among the plurality of product recommendation logic based on a sampling value extracted through sampling by the distribution. choose,
The distribution is
The product recommendation device, characterized in that the distribution is changed through the exposure number information and the conversion number information, and has a maximum value in the expected value of the conversion ratio before the distribution converges to a specific value. delete The method of claim 1,
The logic selection algorithm is
If the selected specific product recommendation logic is plural, the product recommendation apparatus characterized in that the exposure ratio is determined by each specific product recommendation logic in proportion to a sampling result based on the distribution. The method of claim 1,
The logic selection algorithm is
The product recommendation device, characterized in that the specific product recommendation logic is selected by comparing an increase in the number of times of exposure information to the information on the number of conversions in the sampling result based on the distribution. The method of claim 1,
The logic selection algorithm is
When the information on the number of exposures and the information on the number of conversions are accumulated according to an increase in traffic, the product recommendation logic in which the sampling result by the distribution converges to a specific value is selected as the specific product recommendation logic. The method of claim 1,
The product exposure part,
When an access request signal for requesting access to a specific page providing product recommendation or a recommendation request signal for requesting the product recommendation is input from the user, the selected product is exposed using the specific product recommendation logic Product recommendation device. The method of claim 1,
The update unit,
When the exposed product is clicked or ordered according to a preset conversion target, it is determined that the product is converted. 8. The method of claim 7,
The update unit,
When the product is exposed, the exposure number information is updated by increasing the exposure number of the specific product recommendation logic by 1, and when the product is converted, the conversion number information is updated by increasing the conversion number of the specific product recommendation logic by 1 Characterized product recommendation device. ◈Claim 9 was abandoned at the time of payment of the registration fee.◈ In the product recommendation method performed by the product recommendation device,
a logic selection step in which the product recommendation device selects a specific product recommendation logic from among a plurality of product recommendation logic preset using a logic selection algorithm;
a product exposure step in which the product recommendation device selects a product using the specific product recommendation logic and exposes the selected product; and
An updating step of updating, by the product recommendation device, the logic selection algorithm by using feedback information including information on the number of times of exposure of the product and information on the number of times of conversion of the product;
The logic selection algorithm is
A distribution having a random variable is formed based on the exposure number information and the conversion number information, and the specific product logic is selected from among the plurality of product recommendation logic based on a sampling value extracted through sampling by the distribution. choose,
The distribution is
A product recommendation method, characterized in that the shape of the distribution is changed through the exposure number information and the conversion number information, and the distribution has a maximum value in the expected value of the conversion ratio before the distribution converges to a specific value. delete ◈Claim 11 was abandoned when paying the registration fee.◈ 10. The method of claim 9,
The logic selection algorithm is
If the selected specific product recommendation logic is plural, the product recommendation method characterized in that the exposure ratio by each specific product recommendation logic is determined in proportion to a sampling result based on the distribution. ◈Claim 12 was abandoned when paying the registration fee.◈ 10. The method of claim 9,
The logic selection algorithm is
The product recommendation method, characterized in that the specific product recommendation logic is selected by comparing an increase in the number of times of conversion information to the information on the number of exposures in the sampling result based on the distribution. ◈Claim 13 was abandoned when paying the registration fee.◈ 10. The method of claim 9,
The logic selection algorithm is
When the exposure number information and the conversion number information are accumulated according to an increase in traffic, the product recommendation logic in which the sampling result by the distribution converges to a specific value is selected as the specific product recommendation logic. ◈Claim 14 was abandoned when paying the registration fee.◈ 10. The method of claim 9,
The product exposure step is
When an access request signal for requesting access to a specific page providing product recommendation or a recommendation request signal for requesting the product recommendation is input from the user, the selected product is exposed using the specific product recommendation logic How to recommend products. ◈Claim 15 was abandoned when paying the registration fee.◈ 10. The method of claim 9,
The update step is
When the exposed product is clicked or ordered according to a preset conversion target, it is determined that the product is converted. ◈Claim 16 has been abandoned at the time of payment of the registration fee.◈ 16. The method of claim 15,
The update step is
When the product is exposed, the exposure number information is updated by increasing the exposure number of the specific product recommendation logic by 1, and when the product is converted, the conversion number information is updated by increasing the conversion number of the specific product recommendation logic by 1 How to recommend a product that is characterized. ◈Claim 17 was abandoned when paying the registration fee.◈ A recording medium recording a program for executing a product recommendation method, the recording medium comprising:
A logic selection function that selects a specific product recommendation logic from among a plurality of product recommendation logic preset using a logic selection algorithm;
A product exposure function that selects a product using the specific product recommendation logic and exposes the selected product; and
Implementing an update function for updating the logic selection algorithm using feedback information including information on the number of times of exposure of the product and information on the number of times of conversion of the product,
The logic selection algorithm is
A distribution having a random variable is formed based on the exposure number information and the conversion number information, and the specific product logic is selected from among the plurality of product recommendation logic based on a sampling value extracted through sampling by the distribution. choose,
The distribution is
A computer-readable recording medium in which the form of a distribution is changed through exposure number information and conversion number information, and a program having a maximum value in an expected value of a conversion ratio before the distribution converges to a specific value is recorded and computer-readable.

Images