KR20140096412A - Method to recommend digital contents based on search log and apparatus therefor - Google Patents

Abstract

Disclosed in the present specification is a content recommendation device. The device includes a search history analysis unit which extracts, from user′s content search history, either or both of search keywords and content information selected by the user from search results according to the search keywords; a recommendation score calculation unit which calculates recommendation scores for one or more recommendation candidate content based on either or both of the search keywords and the selected content information; and a recommendation content determination unit which determines the recommendation content to be recommended for the user based on the calculated recommendation scores.

Description

[0001] METHOD TO RECOMMEND DIGITAL CONTENTS BASED ON SEARCH LOG AND APPARATUS THEREFOR [0002]

The present invention relates to a content recommendation apparatus and a content recommendation service method, and more particularly, to a method for recommending a content closest to a user's taste based on information obtained from a user's content search history and an apparatus used therein .

2. Description of the Related Art [0002] Recently, as a communication network has dramatically developed, the utilization rate of communication services using mobile terminals has increased not only in communication services using existing fixed computers, but also in various types of communication services.

Among the communication services for such computers and / or mobile terminals, there are a content providing service and an app store for purchasing various contents and application programs such as music, movies, and coupons.

In recent years, since the environment in which such contents or applications can be produced in a self sufficient manner by general users other than professional producers has been developed, the number and kinds of contents existing on the online are in myriad.

As such, As the size of the market providing content (app, video, audio, eBook) grows, the choice of the user increases, but it becomes difficult to find contents that are related to the user's interest or taste. Therefore, the importance of selection and recommendation of content based on the user's usage pattern becomes more important. In the case of the existing recommendation system based on the user's purchase / evaluation history, if the user does not have a purchase / evaluation history, Or the number of recommended contents is limited. Therefore, as a supplementary means when the purchase / evaluation information is insufficient and it is difficult to provide the recommended service, the interest keyword and interest category are estimated from the search history of the user and the selection of the search result, and based on the content, There is a need for a way to recommend content that is relevant to your interests.

It is an object of the present invention to provide a content recommendation apparatus and method. More specifically, the present invention estimates interest keywords and interest categories from a user's search history and selection of search results, and recommends contents highly relevant to the user's interest among all the contents based on the contents, A method and an apparatus. Extracting at least one of the search keyword and the selection content information selected by the user from the search result according to the search keyword from the content search history of the user and selecting at least one recommendation based on at least one of the search keyword and the selected content information, It is an object of the present invention to improve a user's satisfaction by providing a content recommendation apparatus for calculating a recommendation score for a candidate content and determining a recommended content to recommend to the user according to the calculated recommendation score.

According to an embodiment of the present invention, a content recommendation apparatus is provided. The apparatus comprising: a search history analyzing unit for extracting at least one of a search keyword from a content search history of a user and selected content information selected by the user from search results according to the search keyword; A recommendation score calculation unit for calculating a recommendation score for at least one recommended candidate content based on at least one of the search keyword and the selected content information; And a recommended content determination unit for determining a recommended content to be recommended to the user according to the calculated recommended score.

The selected content information may include a category of the content selected by the user and an associated keyword of the selected content.

The recommendation score calculation unit may calculate a recommendation score based on the first score in accordance with whether the search keyword is included in the detailed information of the recommendation candidate content and the first score based on whether the related keyword of the search keyword is included in the detailed information of the recommendation candidate content A score of 2, a third score based on the degree of association between the category of the recommended candidate content and the search keyword, a fourth score based on the degree of association between the category of the selected content and the category of the recommended candidate content, The recommendation score can be calculated based on at least one of the fifth score according to the degree of association between the keyword and the associated keyword of the recommended candidate content.

The recommendation score calculator may calculate the recommendation score by assigning a weight determined in proportion to the number of related keywords of the search keyword included in the detailed information of the candidate content to the second score.

The related keyword of the search keyword is an associated keyword whose degree of association with the search keyword is equal to or greater than a predetermined criterion and the weight may be determined by reflecting the number of the related keywords and the degree of association of each of the related keywords with the search keyword .

The recommendation score calculator may calculate the recommendation score by summing the first score to the fifth score at a predetermined reflectance ratio, and the predetermined reflectance ratio may be updated according to the recommendation result of the user.

According to another embodiment of the present invention, a content recommendation method is provided. The method comprising: extracting at least one of a search keyword from a content search history of a user and a selection content information selected by the user from search results according to the search keyword; Calculating a recommendation score for at least one recommended candidate content based on at least one of the search keyword and the selected content information; And determining a recommended content to be recommended to the user according to the calculated recommendation score.

The selected content information may include a category of the content selected by the user and an associated keyword of the selected content.

The step of calculating the recommendation score may further include determining whether the detailed information of the recommendation candidate content includes a first score according to whether the search keyword is included or not, A third score based on the degree of association between the category of the recommended candidate content and the search keyword, a fourth score based on the degree of association between the category of the selected content and the category of the recommended candidate content, And calculating a recommendation score based on at least one of a fifth score according to an association degree between a related keyword of the content and a related keyword of the recommendation candidate content.

The step of calculating the recommendation score may include calculating a recommendation score by assigning a weight determined in proportion to the number of related keywords of the search keyword included in the detailed information of the candidate content to the second score.

The related keyword of the search keyword is an associated keyword whose degree of association with the search keyword is equal to or greater than a predetermined criterion and the weight may be determined by reflecting the number of the related keywords and the degree of association of each of the related keywords with the search keyword .

Wherein the step of calculating the recommendation score is a step of calculating the recommendation score by summing the first score to the fifth score at a predetermined reflectance ratio and the predetermined reflectance ratio is updated according to the user's recommendation response result .

A content recommendation service system is provided in accordance with another embodiment of the present disclosure. Wherein the system extracts at least one of the search keyword from the content search history of the user and the selection content information selected by the user from search results according to the search keyword, Calculating a recommendation score for the recommendation candidate content, and determining a recommendation content to be recommended to the user according to the calculated recommendation score; And a user terminal for displaying the recommended content recommended by the content recommendation apparatus.

The embodiments of the present invention estimate interest keywords and interest categories from a search history of a user and selection of a search result, and recommend contents having high relevance to user's interest among all the contents based on the contents. Furthermore, it is possible to increase the accuracy of analyzing customer preferences by steadily learning user's preference category and preference keyword from search history, and to capture volatility concern of user by providing real-time recommendation according to search and result selection time, Activity can be improved.

1 is a block diagram of a content recommendation apparatus according to an embodiment of the present invention.
2 is a configuration diagram illustrating a content recommendation service system according to an embodiment of the present invention.
3 is a flowchart illustrating a content recommendation method according to an exemplary embodiment of the present invention.

It is noted that the technical terms used herein are used only to describe specific embodiments and are not intended to limit the invention. It is also to be understood that the technical terms used herein are to be interpreted in a sense generally understood by a person skilled in the art to which the present invention belongs, Should not be construed to mean, or be interpreted in an excessively reduced sense. Further, when a technical term used herein is an erroneous technical term that does not accurately express the spirit of the present invention, it should be understood that technical terms that can be understood by a person skilled in the art are replaced. In addition, the general terms used in the present invention should be interpreted according to a predefined or prior context, and should not be construed as being excessively reduced.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings, wherein like reference numerals refer to like or similar elements throughout the several views, and redundant description thereof will be omitted. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail. It is to be noted that the accompanying drawings are only for the purpose of facilitating understanding of the present invention, and should not be construed as limiting the scope of the present invention with reference to the accompanying drawings. The spirit of the present invention should be construed as extending to all modifications, equivalents, and alternatives in addition to the appended drawings.

1 is a block diagram of a content recommendation apparatus according to an embodiment of the present invention.

Referring to FIG. 1, the content recommendation apparatus 100 may include a search history analyzer 101, a recommendation score calculator 102, and a recommendation content determiner 103.

The content recommendation apparatus 100 analyzes a digital content search history of a specific user, analyzes interest keywords and interest categories of the user, and can recommend personalized content using the result. In particular, the server analyzes a log (e.g., a search keyword, content viewed from search results, and the like) that a specific user has searched for in the content market or the like and grasped the content of interest, interest keyword, interest category, etc. . When the contents of interest, keywords of interest, and the like are analyzed by the method and apparatus disclosed in the present specification, recommendation of contents closest to the interest concerned can be provided.

In order to efficiently select the recommended content according to the characteristics of the user analyzed from the search history, the content recommendation apparatus 100 can perform the pre-analysis work on the available content (recommended candidate content). That is, the content recommendation apparatus 100 may precede the task of analyzing / classifying the characteristics of the entire recommendation candidate contents so as to determine the recommendation contents by mapping with a search history (e.g., a specific search keyword) of the user have.

In one embodiment, the content recommendation apparatus 100 may generate and store association information between a category-keyword, a keyword-keyword, and a keyword-category of the recommendation candidate content.

In general, the digital content market includes not only information such as a content ID, a content name, a content price, and a content provider, but also detailed information (e.g., kind, category, have. The content recommendation apparatus 100 can generate category information, associated keyword information, and the like for each content from the corresponding detailed information. At this time, the category information may have a system allowing multi-level or redundancy depending on the nature of the content market. On the other hand, the related keyword information can extract the set of nouns through the morpheme analyzer and extract the significant keywords in the corresponding category through the TF-IDF (Term Frequency-Inverse Document Frequency) method have.

In this way, the content recommendation apparatus 100 can generate / classify information such as a content ID, a category ID, a keyword ID, and the like of each content and store the corresponding information. Category-keyword " information according to the number of keywords extracted from the content, and the degree of association between each keyword and the category, the degree of association between the keywords, and the like can be determined and stored. Table 1 below is an example of association information between keyword-categories generated for arbitrary contents.

keyword Related Categories Category - Keyword relevance
K1
C1 0.9 C3 0.1 C4 0.4
K2 C2 0.5 C3 0.6
K3 C1 0.7 C5 0.9

Referring to Table 1, the related keywords of arbitrary contents are K1, K2, and K3, and each keyword is associated with a plurality of categories, and the degree of association is analyzed. These analytical information are used in the calculation of recommendation fitness for future content recommendation.

Table 2 is an example of association information between keywords and keywords generated for arbitrary contents.

Keyword - Keyword relevance K1 K2 K3 K1 - 0.95 0.12 K2 0.95 - 0.55 K3 0.12 0.55 -

In Table 2, the relevance of each of the related keywords K1, K2, and K3 of arbitrary contents is analyzed. Based on the analysis, if the representative keyword of the content is K1, the keyword K3 can be expected to have a smaller degree of association with the content than K2. These analytical information are used in the calculation of recommendability for future recommendation of content.

The search history analyzing unit 101 of the content recommendation apparatus 100 may extract one or more of the search keyword from the content search history of the specific user and the selected content information selected by the user from the search results according to the search keyword.

That is, a history of a user searching for a content in a specific Internet site (e.g., a content market) is analyzed to determine (1) information about the search keyword (or a search word) And extracts the information about the selected content. Hereinafter, (2) will be referred to as 'selection content information'. At this time, the selected content information may include an ID of the corresponding content, a category of the corresponding content, detailed information of the corresponding content, a related keyword of the corresponding content, a related keyword, and a relation between categories.

Next, the recommended score calculation unit 102 of the content recommendation apparatus 100 calculates a recommendation score for one or more recommended candidate contents based on at least one of the search keyword and the selected content information. That is, the recommendation score (recommended goodness score) is calculated for each of the recommended candidate content (or the content that can be provided by the apparatus) based on the degree of association with the search keyword and / or the selected content information extracted from the search history of the user.

One embodiment for calculating the recommendation score is as follows.

The recommendation score calculator 102 determines whether or not the search keyword is included in the detailed information of the recommendation candidate content, (2) whether detailed information of the recommendation candidate content is included in the detailed information of the recommendation candidate content, (3) the degree of association between the category of the recommended candidate content and the search keyword, (4) the degree of association between the category of the selected content and the category of the recommended candidate content, and 5) The degree of association between the associated keyword of the selected content and the associated keyword of the recommended candidate content can be used. The scores calculated by the above (1) to (5) are referred to as first to fifth scores, respectively.

At this time, in the above (1) to (5), 'relevance map' may use association information between keyword-category analyzed prior and / or association information between keyword-keywords.

The recommendation score may be calculated by setting appropriate weights for the variables (first to fifth scores). For example, a recommendation score of the content can be calculated by putting the largest weight in the above (1) (for example, 2) and assigning a smaller weight (for example, 0.5) to (2) This method intuitively calculates the recommendation score by increasing the weight of the search keyword itself, which is expected to be the most relevant.

Alternatively, it is possible to calculate the recommendation score of the contents by putting the largest weight in the above (5) (for example, 1.5) and assigning a smaller weight (for example, 1.2) to (1) to (4). This compares the entire set of keywords for each of the selected content-candidate content and places a greater weight on the potential interests of the user who are not directly exposed.

Alternatively, the recommendation score may be calculated by adding a weight determined in proportion to the number of associated keywords of the search keyword included in the detailed information of the candidate content to the second score. That is, in the comparison between the contents in which the search keyword is not directly included, it is determined that the content including more keywords of the search keyword is closer to the potential interest of the user.

Here, the related keyword of the search keyword is an associated keyword whose association degree with the search keyword is equal to or greater than a predetermined reference, and the weight is determined by reflecting the number of the related keywords and the degree of association of each of the related keywords with the search keyword It is possible. For example, even if the related keywords of the search keywords include five identical keywords, the number of related keywords having a relation with the predetermined content (for example, 0.7) or more may be different. In this case, The content including the above keywords can be determined as the content suitable for the recommendation. Also, even if the number of related keywords having the relevance level equal to or higher than the reference level is the same, if the relevance degree value is reflected in the weight value, the higher the degree of association, the closer the content to the user's interest.

The recommendation score calculator 102 calculates the recommendation score by summing the first score to the fifth score at a predetermined reflectance ratio, and the predetermined reflectance ratio may be updated according to the user's recommendation response result have. For example, the recommended score calculation unit 102 calculates

W1 * first score + W2 * second score + W3 * third score + W4 * fourth score + W5 * fifth score, where W1, W2, W3, W4, W5 is the weight of each score)

, And the recommendation score can be calculated using the same formula. In this case, W1 to W5 can be continuously learned by tracking and analyzing the recommendation response history of the user.

When the recommendation score for each candidate content is calculated as described above, the recommendation content determination unit 103 of the content recommendation apparatus 100 determines recommendation content to be recommended to the user according to the calculated recommendation score.

For example, you can recommend the top 10 content in the order of recommended scores, recommend all of the content with a recommended score above a certain threshold, or recommend the content out of the top 20 recommended by the user, The recommended content can be determined.

In this way, the content recommendation apparatus 100 can generate a recommendation result even when the user does not have a history of purchasing a content. The content recommendation apparatus 100 can analyze the customer taste by steadily learning the user's preference category and preferred keyword from the search history Accuracy can be increased.

2 is a configuration diagram illustrating a content recommendation service system according to an embodiment of the present invention.

The content recommendation service system may include a content recommendation apparatus 100 and a user terminal 10. At this time, the content recommendation apparatus 100 extracts at least one of the search keyword and the selection content information selected by the user from the search result according to the search keyword from the content search history of the user, and selects one of the search keyword and the selected content information Based on the above, a recommendation score for one or more recommended candidate contents may be calculated, and a recommendation content to be recommended to the user may be determined according to the calculated recommendation score. Also, the user terminal 10 can display the recommended content recommended by the content recommendation apparatus.

The user terminal 10 may be a terminal capable of receiving contents provided from the content recommendation apparatus 100 by accessing a service provided by the content recommendation apparatus 100, for example, a mobile terminal or a computer. The user terminal 10 can access a service provided by the content recommendation apparatus 100 through a series of login processes by a specific subscriber.

Accordingly, the user terminal 10 can request a content recommendation for a specific category, and thereby display the recommended content provided from the content recommendation apparatus 100. [ That is, the user terminal 10 can request the content recommendation apparatus 100 for a content recommendation for this particular category (e.g., movie) when a specific content category (e.g., movie) is input or selected by the user.

Alternatively, the user terminal 10 may set recommendation request information for a time recommendation for a content recommendation request in advance. When the user terminal 10 reaches a specific time zone, a content recommendation request for a specific content category (e.g., entertainment) To the content recommendation apparatus 100. [ In addition, the manner in which the user terminal 10 requests content recommendation for a particular content category with respect to the subscriber may be realized in various ways.

The user terminal 10 can immediately display the content access information that allows the user to view the recommended content, for example, the recommended content, if the content recommending content is provided from the content recommendation apparatus 100 according to the content recommendation request, The user can display the content access information of the recommended content.

Meanwhile, the user terminal 10 may provide content search history information to the content recommendation apparatus 100. This can be done via the user terminal 10 or the content server. If necessary, the user terminal 10 may transmit the user's prior consent for providing the information.

The content recommendation apparatus 100 analyzes a digital content search history of a specific user, analyzes interest keywords and interest categories of the user, and can recommend personalized content using the result. In particular, the server analyzes a log (e.g., a search keyword, content viewed from search results, and the like) that a specific user has searched for in the content market or the like and grasped the content of interest, interest keyword, interest category, etc. . When the contents of interest, keywords of interest, and the like are analyzed by the method and apparatus disclosed in the present specification, recommendation of contents closest to the interest concerned can be provided.

In order to efficiently select the recommended content according to the characteristics of the user analyzed from the search history, the content recommendation apparatus 100 can perform the pre-analysis work on the available content (recommended candidate content). That is, the content recommendation apparatus 100 may precede the task of analyzing / classifying the characteristics of the entire recommendation candidate contents so as to determine the recommendation contents by mapping with a search history (e.g., a specific search keyword) of the user have.

In one embodiment, the content recommendation apparatus 100 may generate and store association information between a category-keyword, a keyword-keyword, and a keyword-category of the recommendation candidate content.

In general, the digital content market includes not only information such as a content ID, a content name, a content price, and a content provider, but also detailed information (e.g., kind, category, have. The content recommendation apparatus 100 can generate category information, associated keyword information, and the like for each content from the corresponding detailed information. At this time, the category information may have a system allowing multi-level or redundancy depending on the nature of the content market. On the other hand, the related keyword information can extract the set of nouns through the morpheme analyzer and extract the significant keywords in the corresponding category through the TF-IDF (Term Frequency-Inverse Document Frequency) method have.

In this way, the content recommendation apparatus 100 can generate / classify information such as a content ID, a category ID, a keyword ID, and the like of each content and store the corresponding information. Category-keyword " information according to the number of keywords extracted from the content, and the degree of association between each keyword and the category, the degree of association between the keywords, and the like can be determined and stored. Tables 1 and 2 described in FIG. 1 are examples of association information between keyword-category and keyword-keyword generated for arbitrary contents.

The content recommendation apparatus 100 may extract one or more of the search keyword from the content search history of the specific user and the selected content information selected by the user from the search results according to the search keyword.

That is, a history of a user searching for a content in a specific Internet site (e.g., a content market) is analyzed to determine (1) information about the search keyword (or a search word) And extracts the information about the selected content. Hereinafter, (2) will be referred to as 'selection content information'. At this time, the selected content information may include an ID of the corresponding content, a category of the corresponding content, detailed information of the corresponding content, a related keyword of the corresponding content, a related keyword, and a relation between categories.

Next, the content recommendation apparatus 100 calculates a recommendation score for one or more recommended candidate contents based on at least one of the search keyword and the selected content information. That is, the recommendation score (recommended goodness score) is calculated for each of the recommended candidate content (or the content that can be provided by the apparatus) based on the degree of association with the search keyword and / or the selected content information extracted from the search history of the user.

One embodiment for calculating the recommendation score is as follows.

The content recommendation apparatus 100 may include (1) whether or not the search keyword is included in the detailed information of the recommendation candidate content, (2) whether the detail keyword of the recommendation candidate content is included in the detailed information of the recommendation candidate content (4) the degree of association between the category of the selected content and the category of the recommended candidate content; (5) the degree of association between the category of the selected candidate content and the recommended candidate content; ) Association degree between the related keyword of the selected content and the related keyword of the recommended candidate content can be used. The scores calculated by the above (1) to (5) are referred to as first to fifth scores, respectively.

At this time, in the above (1) to (5), 'relevance map' may use association information between keyword-category analyzed prior and / or association information between keyword-keywords.

The recommendation score may be calculated by setting appropriate weights for the variables (first to fifth scores). For example, a recommendation score of the content can be calculated by putting the largest weight in the above (1) (for example, 2) and assigning a smaller weight (for example, 0.5) to (2) This method intuitively calculates the recommendation score by increasing the weight of the search keyword itself, which is expected to be the most relevant.

Alternatively, it is possible to calculate the recommendation score of the contents by putting the largest weight in the above (5) (for example, 1.5) and assigning a smaller weight (for example, 1.2) to (1) to (4). This compares the entire set of keywords for each of the selected content-candidate content and places a greater weight on the potential interests of the user who are not directly exposed.

Alternatively, the recommendation score may be calculated by adding a weight determined in proportion to the number of associated keywords of the search keyword included in the detailed information of the candidate content to the second score. That is, in the comparison between the contents in which the search keyword is not directly included, it is determined that the content including more keywords of the search keyword is closer to the potential interest of the user.

Here, the related keyword of the search keyword is an associated keyword whose association degree with the search keyword is equal to or greater than a predetermined reference, and the weight is determined by reflecting the number of the related keywords and the degree of association of each of the related keywords with the search keyword It is possible. For example, even if the related keywords of the search keywords include five identical keywords, the number of related keywords having a relation with the predetermined content (for example, 0.7) or more may be different. In this case, The content including the above keywords can be determined as the content suitable for the recommendation. Also, even if the number of related keywords having the relevance level equal to or higher than the reference level is the same, if the relevance degree value is reflected in the weight value, the higher the degree of association, the closer the content to the user's interest.

The content recommendation apparatus 100 may calculate the recommendation score by summing the first score to the fifth score at a predetermined reflectance ratio, and the predetermined reflectance ratio may be updated according to the recommendation result of the user . For example, the content recommendation apparatus 100 may include

W1 * first score + W2 * second score + W3 * third score + W4 * fourth score + W5 * fifth score, where W1, W2, W3, W4, W5 is the weight of each score)

, And the recommendation score can be calculated using the same formula. In this case, W1 to W5 can be continuously learned by tracking and analyzing the recommendation response history of the user.

When the recommendation score for each candidate content is calculated as described above, the content recommendation apparatus 100 determines a recommendation content to be recommended to the user according to the calculated recommendation score.

For example, you can recommend the top 10 content in the order of recommended scores, recommend all of the content with a recommended score above a certain threshold, or recommend the content out of the top 20 recommended by the user, The recommended content can be determined.

3 is a flowchart illustrating a content recommendation method according to an exemplary embodiment of the present invention.

Hereinafter, the content recommendation method described above may be performed by the content recommendation apparatus.

The content recommendation apparatus analyzes a digital content search history of a specific user, analyzes interest keywords and interest categories of the user, and can recommend personalized content using the result. In particular, the server analyzes a log (e.g., a search keyword, content viewed from search results, and the like) that a specific user has searched for in the content market or the like and grasped the content of interest, interest keyword, interest category, etc. . When the contents of interest, keywords of interest, and the like are analyzed by the method and apparatus disclosed in the present specification, recommendation of contents closest to the interest concerned can be provided.

In order to efficiently select the recommended content according to the characteristics of the user analyzed from the search history, the content recommendation apparatus can perform a preliminary analysis work on the available content (recommended candidate content). That is, the content recommendation apparatus may precede the task of analyzing / classifying the characteristics of all recommended candidate contents so as to determine the recommended contents by mapping with a search history (e.g., a specific search keyword) of the user.

In one embodiment, the content recommendation apparatus can generate and store association information between category-keyword, keyword-keyword, and keyword-category of the recommendation candidate content.

In general, the digital content market includes not only information such as a content ID, a content name, a content price, and a content provider, but also detailed information (e.g., kind, category, have. The content recommendation apparatus can generate category information, associated keyword information, and the like for each content from the detailed information. At this time, the category information may have a system allowing multi-level or redundancy depending on the nature of the content market. On the other hand, the related keyword information can extract the set of nouns through the morpheme analyzer and extract the significant keywords in the corresponding category through the TF-IDF (Term Frequency-Inverse Document Frequency) method have.

As described above, the content recommendation apparatus can generate / classify information such as a content ID, a category ID, a keyword ID, and the like of each content and store the corresponding information. Category-keyword " information according to the number of keywords extracted from the content, and the degree of association between each keyword and the category, the degree of association between the keywords, and the like can be determined and stored. Tables 1 and 2 described in FIG. 1 are examples of association information between keyword-category and keyword-keyword generated for arbitrary contents.

The content recommendation apparatus may extract one or more of the search keyword from the content search history of the specific user and the selected content information selected by the user from the search results according to the search keyword.

That is, a history of a user searching for a content in a specific Internet site (e.g., a content market) is analyzed to determine (1) information about the search keyword (or a search word) And extracts the information about the selected content. Hereinafter, (2) will be referred to as 'selection content information'. At this time, the selected content information may include an ID of the corresponding content, a category of the corresponding content, detailed information of the corresponding content, a related keyword of the corresponding content, a related keyword, and a relation between categories.

Next, the content recommendation apparatus calculates a recommendation score for one or more recommended candidate contents based on at least one of the search keyword and the selected content information. That is, the recommendation score (recommended goodness score) is calculated for each of the recommended candidate content (or the content that can be provided by the apparatus) based on the degree of association with the search keyword and / or the selected content information extracted from the search history of the user.

One embodiment for calculating the recommendation score is as follows.

(1) whether or not the search keyword is included in the detailed information of the recommendation candidate content, (2) whether the search keyword is included in the detailed information of the recommendation candidate content, (3) the degree of association between the category of the recommended candidate content and the search keyword, (4) the degree of association between the category of the selected content and the category of the recommended candidate content, (5) And the related keyword of the recommended candidate content can be used. The scores calculated by the above (1) to (5) are referred to as first to fifth scores, respectively.

At this time, in the above (1) to (5), 'relevance map' may use association information between keyword-category analyzed prior and / or association information between keyword-keywords.

The recommendation score may be calculated by setting appropriate weights for the variables (first to fifth scores). For example, a recommendation score of the content can be calculated by putting the largest weight in the above (1) (for example, 2) and assigning a smaller weight (for example, 0.5) to (2) This method intuitively calculates the recommendation score by increasing the weight of the search keyword itself, which is expected to be the most relevant.

Alternatively, it is possible to calculate the recommendation score of the contents by putting the largest weight in the above (5) (for example, 1.5) and assigning a smaller weight (for example, 1.2) to (1) to (4). This compares the entire set of keywords for each of the selected content-candidate content and places a greater weight on the potential interests of the user who are not directly exposed.

Alternatively, the recommendation score may be calculated by adding a weight determined in proportion to the number of associated keywords of the search keyword included in the detailed information of the candidate content to the second score. That is, in the comparison between the contents in which the search keyword is not directly included, it is determined that the content including more keywords of the search keyword is closer to the potential interest of the user.

Here, the related keyword of the search keyword is an associated keyword whose association degree with the search keyword is equal to or greater than a predetermined reference, and the weight is determined by reflecting the number of the related keywords and the degree of association of each of the related keywords with the search keyword It is possible. For example, even if the related keywords of the search keywords include five identical keywords, the number of related keywords having a relation with the predetermined content (for example, 0.7) or more may be different. In this case, The content including the above keywords can be determined as the content suitable for the recommendation. Also, even if the number of related keywords having the relevance level equal to or higher than the reference level is the same, if the relevance degree value is reflected in the weight value, the higher the degree of association, the closer the content to the user's interest.

The content recommendation apparatus may calculate the recommendation score by summing the first score to the fifth score at a predetermined reflectance ratio, and the predetermined reflectance ratio may be updated according to the recommendation response result of the user. For example, the content recommendation apparatus

W1 * first score + W2 * second score + W3 * third score + W4 * fourth score + W5 * fifth score, where W1, W2, W3, W4, W5 is the weight of each score)

, And the recommendation score can be calculated using the same formula. In this case, W1 to W5 can be continuously learned by tracking and analyzing the recommendation response history of the user.

When the recommendation score for each candidate content is calculated as described above, the content recommendation apparatus determines recommendation content to be recommended to the user according to the calculated recommendation score.

For example, you can recommend the top 10 content in the order of recommended scores, recommend all of the content with a recommended score above a certain threshold, or recommend the content out of the top 20 recommended by the user, The recommended content can be determined.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

100: content recommendation device
101: Search history analyzing section
102: recommended score calculation unit
103: Recommendation content determination unit

Claims (13)

A search history analyzer for extracting at least one of the search keyword from the user's content search history and the selected content information selected by the user from the search results according to the search keyword;
A recommendation score calculating unit for calculating a recommendation score for at least one recommended candidate content based on at least one of the search keyword and the selected content information; And
And a recommended content determiner for determining a recommended content to be recommended to the user according to the calculated recommendation score. The method according to claim 1,
Wherein the selected content information includes a category of the content selected by the user and a related keyword of the selected content. The method according to claim 1,
The recommendation score calculator calculates,
A first score according to whether or not the search keyword is included in the detailed information of the recommended candidate content, a second score based on whether or not the related keyword of the search keyword is included in the detailed information of the recommended candidate content, A third score based on the relevance degree between the category of the content and the search keyword, a fourth score based on the degree of association between the category of the selected content and the category of the recommendation candidate content, And the fifth score according to the degree of association with the related keyword of the content recommendation score. The method of claim 3,
The recommendation score calculator calculates,
And the recommendation score is calculated by assigning a weight determined in proportion to the number of related keywords of the search keyword included in the detailed information of the candidate content to the second score. 5. The method of claim 4,
Wherein the related keyword of the search keyword is a related keyword whose degree of association with the search keyword is equal to or greater than a predetermined reference and the weight is determined by reflecting the number of the related keywords and the degree of association of each of the related keywords with the search keyword Wherein the content recommendation apparatus comprises: The method of claim 3,
The recommendation score calculator calculates,
Wherein the recommendation score is calculated by summing the first score to the fifth score at a predetermined reflectance ratio, and the predetermined reflectance ratio is updated according to the recommendation response result of the user. Extracting at least one of the search keyword from the content search history of the user and the selected content information selected by the user from search results according to the search keyword;
Calculating a recommendation score for at least one recommended candidate content based on at least one of the search keyword and the selected content information;
And determining a recommended content to be recommended to the user according to the calculated recommendation score. 8. The method of claim 7,
Wherein the selected content information includes a category of the content selected by the user and a related keyword of the selected content. 8. The method of claim 7,
The step of calculating the recommendation score includes:
A first score according to whether or not the search keyword is included in the detailed information of the recommended candidate content, a second score based on whether or not the related keyword of the search keyword is included in the detailed information of the recommended candidate content, A third score based on the relevance degree between the category of the content and the search keyword, a fourth score based on the degree of association between the category of the selected content and the category of the recommendation candidate content, And calculating a recommendation score based on at least one of a fifth score according to a degree of association with a related keyword of the content recommendation score. 10. The method of claim 9,
The step of calculating the recommendation score includes:
And calculating the recommendation score by assigning a weight determined in proportion to the number of related keywords of the search keyword included in the detailed information of the candidate content to the second score. 11. The method of claim 10,
Wherein the related keyword of the search keyword is a related keyword whose degree of association with the search keyword is equal to or greater than a predetermined reference and the weight is determined by reflecting the number of the related keywords and the degree of association of each of the related keywords with the search keyword A content recommendation method characterized by: 10. The method of claim 9,
The step of calculating the recommendation score includes:
Calculating the recommendation score by summing the first score to the fifth score at a predetermined reflectance ratio, and the predetermined reflectance ratio is updated according to the recommendation result of the user. Extracting at least one of the search keyword and the search result according to the search keyword from the content search history of the user and selecting at least one of the search keyword and the selected content information, A content recommendation device for calculating a recommendation score for content and determining a recommendation content to be recommended to the user according to the calculated recommendation score; And
And a user terminal for displaying the recommended content recommended by the content recommendation apparatus.

Images