KR102626741B1 - Method of recommanding products based on user activities - Google Patents

Abstract

A method for recommending products based on user activity is disclosed. A product recommendation method based on user activity according to an embodiment of the present invention includes collecting user-defined data, selecting at least one variable from the user-defined data to determine a characteristic value of the user, and selecting a characteristic value of the user. Creating a group that clusters users with similar values, generating a recommended product list using the characteristic values used to form the group for users included in the group, and selecting individual items belonging to the group from the list of selected products. It includes providing users with a list of products excluding products that overlap with products purchased by the user.

Description

{METHOD OF RECOMMANDING PRODUCTS BASED ON USER ACTIVITIES}

The present invention relates to a method for recommending products based on user activity, and more specifically, to a method for recommending products based on user activity that can recommend products optimized for users by analyzing the activities of users consuming content.

The consumption of content such as sound sources, videos, images, etc. is changing from the past method of purchasing physical media and playing it on electronic devices to playing content through streaming without physical media.

For example, this involves paying a subscription fee in a music playback application and downloading and streaming content files in real time.

This new form of content consumption enables services that specify users who reproduce content and recommend related products that take into account the tastes of individual users.

For example, when a specific user repeatedly plays music from a specific singer, this service provides information about products related to that singer or recommends similar content to encourage purchases.

However, in order for products recommended to users to lead to actual purchases, information in the desired form must be appropriately provided to the user at the desired time, and the need for an advanced recommendation method has emerged.

The present invention was created to solve the above-mentioned problems, and the purpose of the present invention is to provide a product recommendation method based on user activity that can provide related products with a high probability of purchase by considering the user consuming content and the user's tastes. I'm doing it.

The problems of the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

A user activity-based product recommendation method according to an embodiment of the present invention to achieve the above-described purpose includes collecting user-defined data, selecting at least one variable from the user-defined data, and determining the characteristic value of the user. A step of creating a group that clusters users with similar feature values to the user, generating a list of recommended products for users included in the group using the feature values used to form the group, and recommended products. It includes providing users with a list of products excluding products that overlap with products purchased by individual users belonging to the group from the list.

According to the above-described method of providing a recommended list, it is possible to provide a list of recommended products that can lead to actual purchases, which achieves the effect of providing useful information from the user's point of view and promoting user convenience from the seller's point of view. there is.

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description of the claims.

1 is a diagram illustrating a user activity-based product recommendation system according to an embodiment of the present invention.
Figure 2 is a flowchart illustrating a method for recommending products based on user activity according to an embodiment of the present invention.
Figure 3 is a diagram for explaining user-defined data according to an embodiment of the present invention.
FIG. 4 is a diagram illustrating a process of clustering users by selecting at least one variable from a plurality of user-defined data according to an embodiment of the present invention.
FIG. 5 is a diagram illustrating a process of counting the number of data items falling within a preset distance based on arbitrary data according to an embodiment of the present invention.
FIG. 6 is a diagram illustrating a process of determining data falling within a preset distance based on arbitrary data into one cluster according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings. However, the attached drawings are merely explained to more easily disclose the content of the present invention, and the scope of the present invention is not limited to the scope of the attached drawings. Those skilled in the art will easily understand this. You will find out.

In addition, in describing the embodiments of the present invention, it is clear in advance that the same names and the same symbols are used for components having the same function, but that they are not substantially the same as the components of the prior art.

Additionally, the terms used in this application are only used to describe specific embodiments and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly dictates otherwise. In this application, 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.

1 is a diagram illustrating a user activity-based product recommendation system according to an embodiment of the present invention.

The user activity-based product recommendation system 100 according to an embodiment of the present invention may include a user terminal device 200, an authentication device 300, and a management server 400.

The user terminal device 200 is an electronic device such as a smartphone or tablet PC, and has a dedicated application installed thereon, so that it can download and play content files through the management server 400.

In order to download a content file from the management server 400, the user transmits the authentication information stored in the authentication device 300 to the user terminal device 200, and the user terminal device 200 transmits it back to the management server 400. do. The management server 400 transmits a content file corresponding to the authentication information received from the user terminal device 200 to the user terminal device 200 .

Additionally, dedicated applications can provide fandom community services related to specific singers. Users can search for information about a specific artist or upload related information through the fandom community service. Alternatively, you can purchase products related to a specific artist from the fandom community.

The management server 400 collects information about artists identified by the user through a dedicated application and can recommend content or products that the user may be interested in based on this information.

Below, we will explain in detail how to provide a list of recommended products to the user based on the type of content played by the user and activity information in the dedicated application.

Figure 2 is a flowchart illustrating a method for recommending products based on user activity according to an embodiment of the present invention.

To provide a list of recommended products, user-defined data is first collected (S210). User-defined data relates to the user's activity history and may include purchase history of specific artist products, content playback history, search terms searched on external sites, etc.

Afterwards, at least one variable is selected from the user-defined data to determine the user's characteristic value (S220). The feature value is a variable needed to create a recommendation list and may be determined to have a different value depending on the type and purpose of creation of the recommendation list.

Once the feature value is determined, a group is created by clustering similar users using this (S230). A cluster refers to a group of users with similar tastes. For example, this may include a cluster of users who like a specific artist or a cluster of users who like specific content.

A specific method of creating a cluster of users using feature values will be described in detail in FIGS. 3 to 6.

When a cluster is created, a list of recommended products is created using the feature values used to create the cluster (S230). Here, the recommended product list may include a list of artists that the user may be interested in, product information related to the artist, performance information, etc. However, of course, the recommended product list is not limited to this and may include various information.

When providing a list of recommended products to a user, if a user belonging to the group has already purchased a product, the product with a purchase history is excluded from the recommended product list (S240).

Accordingly, it is possible to provide information about products that may actually be purchased, thereby achieving the effect of increasing satisfaction in usage.

Figure 3 is a diagram for explaining user-defined data according to an embodiment of the present invention.

User-defined data refers to a type of raw data related to the user's activity history when using content.

Below, as an example, a dedicated application for playing content is installed on the user terminal device (smartphone, etc.) used by the user, and in order to play the content, the user must connect to the application server through the dedicated application and download the content. Listen and explain.

User-defined data according to an embodiment of the present invention can be divided into internal data 320 that can be acquired as the user executes a dedicated application and external data 310 that can be collected by user activities outside the application. there is.

External data 310 that can be collected from external sites linked to dedicated applications may include information about videos played on external sites, posts posted on social media, product purchase history, news viewed on external sites, etc.

Internal data 320 that can be collected from a dedicated application for playing content include the playback frequency of a specific sound source, the tendency to listen to the sound source (playback time, characteristics of the playback sound source by time period), content purchase and use history, and information provided by the dedicated application. This may include the number of community activities, the number of levels and missions achieved by the community, whether the content creation function provided by the dedicated application is executed, and the user's sound characteristics recorded when the content creation function is activated.

However, the user-defined data that can be collected from external sites or dedicated applications is not limited to this, and may include other information that can identify the user's tendencies.

Meanwhile, there may be cases where not all of the collected custom data is needed to create a recommended product list based on the user's activity history on an external site or dedicated application.

Accordingly, the user activity-based product recommendation method according to an embodiment of the present invention can select only the data necessary to create a recommended product list from a plurality of user-defined data.

FIG. 4 is a diagram illustrating a process of clustering users by selecting at least one variable from a plurality of user-defined data according to an embodiment of the present invention.

Figure 4 shows data in which at least one variable among a plurality of user-defined data is determined as a feature value and displayed as visualized data. That is, the first to ninth data 410 to 490 shown in FIG. 4 visualize a user using at least one variable as a feature value.

In FIG. 4 , the first to ninth data 410 to 490 are shown as an example displayed on a two-dimensional plane, but they may also be displayed in three-dimensional space depending on the number of variables selected.

*To cluster users with similar feature values, first select random data and calculate the distance between the selected data and other data. For example, when the first data 410 is selected, the distance between the first data 410 and the remaining data is calculated.

Likewise, the same operation is repeated for other data. That is, the same operation is repeated for the second to ninth data 420 to 490. Next, the number of data items falling within a preset distance based on random data is counted.

FIG. 5 is a diagram illustrating a process of counting the number of data items falling within a preset distance based on arbitrary data according to an embodiment of the present invention.

As explained in FIG. 4, users displayed as feature values can be located on a two-dimensional plane or three-dimensional space, so the distance between data can be calculated.

Afterwards, a circle with a radius of ε is drawn centered on the location of random data, and the number of data belonging to the circle is counted. In FIG. 3, since the distance between arbitrary data and other data is calculated based on arbitrary data, it can be implemented by counting the number of data whose distance between data is ε or less.

At this time, if the number of data included in a circle with a radius of ε based on the random data is greater than or equal to a preset threshold, the random data and the data included in the circle are determined to be one cluster.

On the other hand, if the number of data included in a circle with a radius of ε based on random data is less than a preset threshold, this is treated as noise and determined not to be a cluster.

In FIG. 4 , since there is no data other than the first data 410 that falls within a circle with a radius of ε centered on the first data 410, it is determined that the data centered on the first data 410 is not a cluster.

FIG. 6 is a diagram illustrating a process of determining data falling within a preset distance based on arbitrary data into one cluster according to an embodiment of the present invention.

As shown in FIG. 6, when the seventh data 470 is determined as the reference data, the fourth data 440, the fifth data 450, the sixth data 460, and the eighth data ( 480) is included.

That is, a cluster containing a plurality of data is formed based on the seventh data 470. In this case, the fourth to eighth data 480 are determined as one cluster having the same feature value. At this time, the radius ε may vary depending on the scale of the entire data, the number of required clusters, etc.

As described above, the preferred embodiments according to the present invention have been examined, and the fact that the present invention can be embodied in other specific forms in addition to the embodiments described above without departing from the spirit or scope thereof is recognized by those skilled in the art. It is self-evident to them. Therefore, the above-described embodiments are to be regarded as illustrative and not restrictive, and thus the present invention is not limited to the above description but may be modified within the scope of the appended claims and their equivalents.

Claims (1)

In the user activity-based product recommendation method performed by the management server,
collecting user-defined data by the management server;
determining, by the management server, a characteristic value of the user by selecting at least one variable from the user-defined data;
creating, by the management server, a group clustering users with similar characteristic values to the user;
generating, by the management server, a list of recommended products for users included in the group using feature values used to form the group; and
Providing, by the management server, to users a product list excluding products that overlap with products purchased by individual users belonging to the group from among the recommended product list,
The step of creating a group in which users with similar characteristic values are clustered,
Visualizing a plurality of users with a plurality of data according to the determined feature values,
determining any data among the plurality of visualized data as reference data and calculating a distance between the determined reference data and other data;
Counting the number of other data items within a preset distance based on the reference data; and
Comparing the counted number of other data with a preset threshold to determine whether to create a cluster centered on the reference data,
The step of determining whether the cluster is created is,
If the counted number of other data is greater than or equal to the preset threshold, determining that a cluster has been created in the data centered on the reference data, and
If the counted number of other data is less than the preset threshold, determining that no clusters have been created in the data centered on the reference data,
The step of creating a group in which users with similar characteristic values are clustered,
When it is determined that the cluster has been created, further comprising determining the reference data and other data falling within the preset distance based on the reference data into one cluster,
Product recommendation method based on user activity.