KR100792701B1 - Method for targeting web advertisement clickers based on click pattern by using a collaborative filtering system and system thereof - Google Patents

Abstract

The present invention collects user click information on an advertisement, calculates a correlation between advertisements using the collected advertisement click information, and recommends an advertisement highly correlated with the advertisement clicked by the user. It is about. The present invention provides a method of collecting advertisement click information for a plurality of advertisements, analyzing correlation between advertisements using the collected advertisement click information, and correlation between the analyzed advertisements for advertisements previously clicked by a user. It provides an advertisement recommendation method comprising the step of recommending an advertisement having a high correlation with the clicked advertisement using a diagram. According to the present invention, there is provided a method and system for calculating a correlation between advertisements using advertisement click information, and recommending an advertisement highly correlated with an advertisement previously clicked by a user with reference to the calculated correlation between the advertisements. Providing can increase the advertising effect.

Advertising, Referrals, Correlation, Collaborative filtering

Description

TECHNICAL FOR TARGETING WEB ADVERTISEMENT CLICKERS BASED ON CLICK PATTERN BY USING A COLLABORATIVE FILTERING SYSTEM AND SYSTEM THEREOF}

1 is a view showing an interlocking relationship between a user terminal and an advertisement recommendation system according to the present invention.

2 is a diagram illustrating a flow of an advertisement recommending method according to an embodiment of the present invention.

3 is a diagram illustrating an example of a procedure of selecting a higher advertisement having a high correlation in the advertisement recommendation method.

4 illustrates whether a user clicks on an advertisement.

FIG. 5 is a diagram illustrating an example of a procedure for recommending an advertisement based on a result of comparing the correlation between advertisements in the advertisement recommending method.

6 is a diagram illustrating an example of an advertisement recommendation.

7 is a diagram showing the configuration of an advertisement recommendation system according to another embodiment of the present invention.

<Explanation of symbols for the main parts of the drawings>

700: Ad Referral System

710: advertising information collection unit

720: advertising pattern extraction unit

730: correlation calculation unit

740: comparison unit

750: recommendation

The present invention relates to a method and system for recommending advertisements, and more particularly, to collect user click information on an advertisement, calculate the correlation between the advertisements using the collected advertisement click information, The present invention relates to a method and system for recommending an advertisement highly correlated with an advertisement.

Based on the internet web, various services such as e-commerce, electronic advertisement, and internet phone are provided. Internet service is rapidly emerging as the most important business as users can easily receive services through the internet. have. A service for searching for specific data or contents on the Internet, which is a sea of information, is recognized as an essential service for efficiently using the Internet.

That is, when a user wants to search for specific data or content, the user can access a site that provides an Internet search service using an information retrieval system even if he / she does not know the URL or IP address of the site and inputs a keyword of the data or content to search for. Result data can be serviced.

The main revenue model of sites that provide such Internet search service is to expose banner advertisements to users who want to receive the search service. The profit model through the banner advertisement may increase the advertising revenue when the click-through rate of the exposed banner advertisement is clicked on by the user is higher than the banner advertisement is exposed to the user.

However, the conventional banner advertising method is closely related to the advertising revenue because a banner advertisement is not provided to the user because the user can not grasp the user's taste or interest when the search keyword is not input by the user in the search service. There is a problem that does not increase the clickthrough rate of banner ads.

Therefore, there is an urgent need for an advertisement recommendation method for increasing the clickthrough rate when a user clicks on a banner advertisement.

The present invention has been made to improve the above problems of the prior art, and provides a method and system for recommending an advertisement highly correlated with an advertisement previously clicked by a user to increase the click rate of the banner advertisement. For the purpose of

In order to achieve the above object and to solve the problems of the prior art, the present invention collects the advertisement click information for a plurality of advertisements, analyzing the correlation between the advertisements using the collected advertisement click information and And recommending an advertisement having a high correlation with the clicked advertisement by using a correlation between the analyzed advertisement with respect to an advertisement previously clicked by a user.

According to another aspect of the present invention, an advertisement recommendation system includes an advertisement information collection unit configured to collect clicked advertisement information and cookie information of a user who clicks the advertisement, and an advertisement pattern using the advertisement information and the cookie information of the user. A pattern extracting unit to extract, a correlation calculating unit calculating a correlation between advertisements using the extracted advertisement pattern, a comparison unit comparing the correlation between the previous advertisement clicked by the user and the calculated advertisement; And a recommendation unit recommending an advertisement highly correlated with the previous advertisement clicked by the user according to the comparison result.

Hereinafter, with reference to the accompanying drawings will be described in detail the advertising recommendation method and system according to an embodiment of the present invention.

1 is a view showing an interlocking relationship between a user terminal and an advertisement recommendation system according to the present invention.

Referring to FIG. 1, the advertisement recommendation system 100 provides a banner advertisement requested by an advertiser to a user terminal 120 connected to a web page of an advertisement recommendation site through the communication network 110. The advertisement recommendation system 100 collects click information of users of the banner advertisement and calculates a correlation between the advertisements using the collected click information.

That is, the advertisement recommendation system 100 recognizes the user information by using the cookie information of the user terminal 120 when the user terminal 120 is connected to the web page of the advertisement recommendation site through the communication network 110. The banner advertisement having a high correlation with the advertisement previously clicked by the user is provided to the user terminal 120 through the web page.

Accordingly, the advertisement recommendation system 100 recommends a banner advertisement having a high correlation with a banner advertisement previously clicked by the user to the user of the user terminal 120 connected through the communication network 110. To provide.

As such, the advertisement recommendation system 100 according to the present invention analyzes the advertisement click information of the user and analyzes the correlation between the advertisements to provide an advertisement having a high correlation with the advertisement previously clicked by the user. To increase advertising effectiveness.

The communication network 110 is a wired or wireless network that transmits various data transmitted and received between the user terminal 120 and the advertisement recommendation system 100.

The user terminal 120 may be connected to the advertisement recommendation system 100 through the communication network 110 as a terminal having a computing capability by including a microprocessor and a memory means such as a desktop PC, a notebook PC, a PDA, a mobile communication terminal, and the like. Collectively known as

The user terminal 120 connects to the advertisement recommendation system 100 through the communication network 110 and transmits the user's cookie information and information based on whether the user clicks on an advertisement provided from the advertisement recommendation system 100 to the advertisement recommendation system 100. send. In addition, the user terminal 120 receives an advertisement recommended from the advertisement recommendation system 100 through the communication network 110.

2 is a diagram illustrating a flow of an advertisement recommending method according to an embodiment of the present invention.

1 and 2, in step 210, the advertisement recommendation system 100 collects advertisement click information on a plurality of advertisements transmitted from the user terminal 120 through the communication network 110. The advertisement click information includes whether the advertisement is clicked and information on a user who clicks the advertisement.

That is, in step 210, the advertisement recommendation system 100 uses the cookie information of the user when the user clicks on a banner advertisement provided from the advertisement recommendation system 100 from the user terminal 120 through the communication network 110. The user identifier included in the cookie information may be collected and stored together with the banner advertisement identifier in an advertisement database. The cookie information includes a cookie ID (user ID) as the user identifier.

In step 220, the advertisement recommendation system 100 analyzes the correlation between advertisements using the collected advertisement click information. That is, in step 220, the advertisement recommendation system 100 extracts the advertisement pattern vector using the collected advertisement click information, and analyzes the correlation between the advertisements using the extracted advertisement pattern vector.

In operation 220, the advertisement recommendation system 100 analyzes the correlation between the advertisements through a collaborative filtering algorithm using the extracted advertisement pattern vectors when the number of extracted advertisement pattern vectors is sufficient. do.

The collaborative filtering according to the present invention is a technique for identifying advertisements having similar patterns of preferences and interests of users based on the extracted advertisement pattern vector, and crosses advertisements that have not yet been clicked on by users having similar tastes. It is used for recommending or recommending related advertisements according to the taste of classified users.

In operation 230, the advertisement recommendation system 100 recommends an advertisement having a high correlation with the clicked advertisement by using a correlation between the advertisements analyzed for the advertisement clicked by the user.

That is, in step 230, the advertisement recommendation system 100 stores the correlation between the advertisements, identifies the advertisement previously clicked by the user and the user, and compares the identified advertisements with the stored advertisements. Compare and recommend an advertisement highly correlated with the identified advertisement according to the comparison result.

Accordingly, the advertisement recommendation system 100 recommends an advertisement having a high correlation with the advertisement previously clicked by the user or having a similarity with the clicked advertisement and providing the advertisement to the user terminal 120 through the communication network 110. do.

For example, if the advertisement recommendation system 100 provides a second advertisement highly correlated with the first advertisement to a user who clicks the first advertisement, the advertisement recommendation system 100 may provide a third advertisement having low correlation with the first advertisement. The clickthrough rate for the second advertisement may be improved as compared with the recommendation advertisement.

As described above, the advertisement recommendation method according to the present invention collects advertisement information clicked by users, analyzes the collected advertisement information, extracts correlation between advertisements, and uses the correlation between the extracted advertisements. The advertisement effect may be increased by recommending an advertisement having a high correlation with an advertisement previously clicked by the user and providing the advertisement to the user.

3 is a diagram illustrating an example of a procedure of selecting a higher advertisement having a high correlation in the advertisement recommendation method.

1 and 3, in step 310, the advertisement recommendation system 100 receives a cookie ID of a user transmitted through the communication network 110 from a user terminal 120 accessing an advertisement recommendation site. When the user terminal 120 accesses the advertisement recommendation site, the user terminal 120 transmits the user ID and password input from the user to the advertisement recommendation system 100 through the communication network 110 as cookie information to log in to the advertisement recommendation site.

The advertisement recommendation system 100 provides various advertisements to the user terminal 120 connected through the web page of the advertisement recommendation site. Then, when the user interested in the advertisement provided in the advertisement recommendation site is clicked by the user, the user terminal 120 transmits the click information on the clicked advertisement to the advertisement recommendation system 100 through the communication network 110. do. The click information indicates whether the user clicks on the advertisement provided by the user.

In operation 320, the advertisement recommendation system 100 receives click information on an advertisement transmitted from the user terminal 120 connected to the advertisement recommendation site.

In operation 330, the advertisement recommendation system 100 extracts an advertisement click pattern using the input cookie ID and the click information. The advertisement click pattern is implemented as a matrix including the cookie ID (user information) and the click information (clicked advertisement information).

In step 340, the advertisement recommendation system 100 determines whether the number of advertisement click vectors according to the extracted advertisement click pattern is sufficient to analyze the correlation between advertisements. That is, in step 340, if the number of the advertisement click vectors is small, the advertisement recommendation system 100 does not have a rule for analyzing the correlation between the advertisements, so that the number is enough to analyze the correlation. Determine whether or not.

In step 350, the advertisement recommendation system 100 selects an appropriate advertisement click vector from the extracted advertisement click pattern. That is, in operation 350, the advertisement recommendation system 100 selects an advertisement click vector suitable for analyzing the correlation between the advertisements from the extracted advertisement click pattern.

In step 360, the advertisement recommendation system 100 analyzes the correlation between the advertisements using the selected advertisement click vector. That is, in step 360, the advertisement recommendation system 100 analyzes the correlation between the advertisements using the selected advertisement click vector using a collaborative filtering algorithm.

Therefore, the advertisement recommendation method according to the present invention compares the selected advertisement click vectors through a collaborative filtering algorithm, and if the types are similar, analyzes that the correlation between the advertisements is high and compares the selected advertisement click vectors. If the types are not similar, it may be analyzed that the correlation between the advertisements is low.

The input data for performing the collaborative filtering algorithm according to the present invention is represented by an m * n (user-advertisement) matrix as a set of clicks of m user IDs and n advertisement IDs. As shown in FIG. 4, the matrix has a value of '1' when a user clicks on an advertisement and a value of '0' when the user has not clicked on the advertisement.

The collaborative filtering algorithm according to the present invention, for example, based on the advertisement ID shown in Figure 4 to determine the click pattern between the advertisement "{1,0,1, ...}, {0,1,0, ...}, {1,0,1, ...}, {0,0,0, ...}, ... and analyze the correlation between the ads by comparing the click patterns of the ads can do.

In step 360, the ad recommendation system 100 cross-corresponds with the advertisement 'a' and the advertisement 'a' if, for example, users who click on the advertisement 'a' by clicking on the advertisement 'a' based on the advertisement ID It can be analyzed that the degree is high, and if the user who clicks the 'A' does not click the 'A', the 'A' and the 'I' are not highly correlated.

In operation 370, the advertisement recommendation system 100 stores the correlation between the analyzed advertisements. That is, in step 370, the advertisement recommendation system 100 stores the correlation between advertisements analyzed for the plurality of advertisement click information.

In operation 380, the advertisement recommendation system 100 selects the top N advertisements having a high correlation between the analyzed advertisements. That is, in step 380, the advertisement recommendation system 100 selects the top N advertisements having high correlation between the analyzed advertisements in order to recommend advertisements having a high correlation with the advertisement previously clicked by the user.

As described above, the advertisement recommendation method according to the present invention extracts click information on a plurality of advertisements, analyzes the correlation between advertisements based on the extracted click information, and selects the top N advertisements having high analyzed correlation. Can be.

FIG. 5 is a diagram illustrating an example of a procedure for recommending an advertisement based on a result of comparing the correlation between advertisements in the advertisement recommending method.

1 and 5, in step 510, the advertisement recommendation system 100 receives a cookie ID of a user from a user terminal 120 connected to a web page of an advertisement recommendation site through the communication network 110. That is, in step 510, the advertisement recommendation system 100 receives cookie information transmitted through the communication network 110 from the user terminal 120 accessing the advertisement recommendation site, and extracts the cookie ID from the cookie information. Identifies the user information.

In step 520, the advertisement recommendation system 100, when a banner advertisement provided by the advertisement recommendation site is clicked by a user of the user terminal 120 accessing the web page of the advertisement recommendation site, the advertisement recommendation system 100 is added to the clicked banner advertisement. Enter the click information for That is, in step 520, the advertisement recommendation system 100 receives the click information including the clicked banner advertisement ID when the banner advertisement provided on the web page of the mall advertisement recommendation site is clicked by the user.

Accordingly, the advertisement recommendation system 100 may identify the user information clicked on the banner advertisement and the advertisement information clicked by the user through steps 510 and 520.

In operation 530, the advertisement recommendation system 100 extracts a click vector using the input cookie ID and the click information.

In operation 540, the advertisement recommendation system 100 compares the degree of correlation with the advertisement using the extracted click vector. That is, in operation 540, the advertisement recommendation system 100 may select an advertisement having a high correlation with the advertisement clicked by the user by comparing the extracted click vector with the stored correlation.

In operation 550, the advertisement recommendation system 100 recommends the selected advertisement, and provides the recommended advertisement to the user terminal 120 connected through the web page of the advertisement recommendation site. That is, in step 550, the advertisement recommendation system 100 provides an advertisement highly correlated with the advertisement clicked by the user to the user terminal 120 connected through the web page of the advertisement recommendation site as a recommendation advertisement. . Therefore, after checking the recommendation advertisement provided on the web page of the advertisement recommendation site through the user terminal 120, the user clicks the advertisement in the case of the interest interest advertisement, and if the advertisement is not his interest interest advertisement, You may not click.

In operation 560, the advertisement recommendation system 100 determines whether the user of the user terminal 120 connected through the web page of the advertisement recommendation site clicks on the provided recommendation advertisement.

If the user does not click on the recommended advertisement provided through the web page of the ad recommendation site, the ad recommendation system 100 performs step 550 again to correlate with the clicked advertisement through the web page of the ad recommendation site. Recommend another high degree ad. At this time, the advertisement recommendation system 100 sequentially sends the advertisement having a high correlation with the clicked advertisement to the user terminal 120 as a recommendation advertisement until the user clicks on the recommendation advertisement provided to the web page of the advertisement recommendation site. Can provide.

Meanwhile, when the user clicks on a recommended advertisement provided through the web page of the advertisement recommendation site, the advertisement recommendation system 100 ends the advertisement recommendation process.

As described above, the advertisement recommendation method according to the present invention can improve the advertisement effect by increasing the click rate of the recommended advertisement by the user by recommending an advertisement highly correlated with the advertisement previously clicked by the user.

6 is a diagram illustrating an example of an advertisement recommendation.

Referring to FIG. 6, a first advertisement 610 is a banner advertisement for advertising a genuine article shopping mall of Renault Samsung Motors, and a second advertisement 620 is a banner advertisement for advertising a Santa Fe launch of a modern automobile. As such, since the first advertisement 610 and the second advertisement 620 are both advertisements associated with automobiles, the correlation between each other is high.

Therefore, when the user clicks on the first advertisement 610, the advertisement recommendation system 100 recommends the second advertisement 620 having a high correlation with the first advertisement 610. That is, the second advertisement 620 is an advertisement that users who click on the first advertisement 610 generally click a lot, and since the second advertisement 620 has a high correlation with the first advertisement 610, the second advertisement 610 is previously clicked. Providing a user with a recommended advertisement increases the probability of being clicked by the user. Therefore, the advertisement effect due to the advertisement recommendation method according to the present invention can be increased.

7 is a diagram showing the configuration of an advertisement recommendation system according to another embodiment of the present invention.

Referring to FIG. 7, the advertisement information collecting unit 710 collects click information of an advertisement and user information of clicking the advertisement. That is, the advertisement information collecting unit 710 collects click information indicating whether a plurality of advertisements are clicked and cookie information of a user who clicks on the advertisements in order to recommend an advertisement.

The pattern extractor 720 extracts an advertisement pattern by using the clicked advertisement information and the cookie information of the user. That is, the pattern extracting unit 720 extracts an advertisement pattern vector from a matrix having the user's cookie information and the advertisement information as rows or columns as shown in FIG. 4.

The correlation calculator 730 calculates a correlation between advertisements using the extracted advertisement pattern. That is, the correlation calculator 730 calculates a correlation between the advertisements using the extracted advertisement pattern vector.

In addition, the correlation calculator 730 calculates a correlation between the advertisements through a collaborative filtering algorithm using the extracted advertisement pattern vector.

The comparison unit 740 compares the correlation between the previous advertisement clicked by the user and another advertisement using the calculated correlation between the advertisements. That is, the comparison unit 740 compares the correlation between the advertisement registered as the banner advertisement and the previous advertisement clicked by the user according to the advertiser's request by using the calculated correlation between the advertisements.

The recommender 750 recommends an advertisement having a high correlation with the previous advertisement clicked by the user among the registered advertisements according to the comparison result.

As such, the advertisement recommendation system according to the present invention collects click information of users for banner advertisements, calculates a correlation between advertisements using the collected click information, and uses a banner advertisement previously clicked by a user. By recommending and providing a banner advertisement having a high correlation with the advertisement, it is possible to increase the click-through rate for the recommended advertisement to increase the advertisement effect.

Embodiments of the invention also include computer-readable media containing program instructions for performing various computer-implemented operations. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. The medium or program instructions may be those specially designed and constructed for the purposes of the present invention, or they may be of the kind well-known and available to those having skill in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical media such as CD-ROMs, DVDs, and magnetic disks, such as floppy disks. Magneto-optical media, and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. The medium may be a transmission medium such as an optical or metal wire, a waveguide, or the like including a carrier wave for transmitting a signal specifying a program command, a data structure, or the like. 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.

As described above, the present invention has been described by way of limited embodiments and drawings, but the present invention is not limited to the above-described embodiments, which can be variously modified and modified by those skilled in the art to which the present invention pertains. Modifications are possible. Accordingly, the spirit of the present invention should be understood only by the claims set forth below, and all equivalent or equivalent modifications thereof will belong to the scope of the present invention.

According to the present invention, a method and system for calculating a correlation between advertisements using advertisement click information, and recommending an advertisement highly correlated with an advertisement previously clicked by a user with reference to the calculated correlation between the advertisements Providing can increase the advertising effect.

Claims (10)

In the ad recommendation method, Collecting advertisement click information for a plurality of advertisements; Analyzing a correlation between advertisements using the collected advertisement click information; Storing a correlation between the analyzed advertisements; Identifying a user and an advertisement previously clicked by the user; Comparing a correlation between the identified advertisement and the stored advertisement; And Recommending an advertisement highly correlated with the identified advertisement according to the comparison result Advertising recommendation method comprising a. The method of claim 1, The step of analyzing the correlation between the advertisements using the collected advertisement click information, Extracting an advertisement click vector by using the collected advertisement click information; And Analyzing the correlation between the advertisements using the extracted advertisement click vectors. Advertising recommendation method comprising a. The method of claim 2, The step of analyzing the correlation between the advertisements using the extracted advertisement click vector, And if the number of extracted advertisement click vectors is sufficient, analyzing the correlation between the advertisements through a collaborative filtering algorithm using the extracted advertisement click vectors. delete The method of claim 1, The recommending the advertisement having a high correlation with the identified advertisement according to the comparison result may include: The advertisement recommendation method of claim 1, wherein the advertisement is highly recommended in correlation with the identified advertisement until the user clicks on the advertisement. A computer-readable recording medium for recording a program for executing the method of any one of claims 1 to 3, 5 on a computer. An advertisement information collection unit for collecting clicked advertisement information and cookie information of a user who clicks the advertisement; A pattern extractor configured to extract an advertisement pattern using the advertisement information and the cookie information of the user; A correlation calculator for calculating a correlation between advertisements using the extracted advertisement pattern; A comparison unit identifying a user and a previous advertisement clicked by the user, and comparing the correlation between the previous advertisement clicked by the user and another advertisement using the calculated correlation between the advertisements; And A recommending unit recommending an advertisement highly correlated with a previous advertisement clicked by the user according to the comparison result Ad recommendation system comprising a. The method of claim 7, wherein The pattern extraction unit, And an advertisement pattern vector extracted from a matrix including rows and columns of the cookie information and the advertisement information of the user. The method of claim 8, The correlation calculation unit, The advertisement recommendation system of claim 1, wherein the correlation between the advertisements is calculated using the extracted advertisement pattern vector. The method of claim 9, The correlation calculation unit, And using the extracted advertisement pattern vector to calculate the correlation between the advertisements through a collaborative filtering algorithm.

Images