KR20170114673A - Contents recommendation system and method - Google Patents

Abstract

A content recommendation system, comprising: a first recommendation unit configured to generate a first recommended content group based on at least one of statistical analysis data and association analysis data on a plurality of contents; information on a recommended person; A second recommendation unit for analyzing the degree of similarity based on at least one of the user information and the plurality of content information and generating a second recommended content group based on the analyzed degree of similarity, And a controller for displaying at least some contents included in the recommended content group and the second recommended content group on the display screen as recommended contents.

Description

{CONTENTS RECOMMENDATION SYSTEM AND METHOD}

The present invention relates to a content recommendation system and method.

As personal terminal devices become more widespread, users connect to a predetermined content service server through a network using their terminal devices, and use content services such as playing or downloading corresponding contents as they select desired contents .

At this time, as the number and type of contents become wider, conventional general contents services classify contents according to categories such as latest contents, popular contents and the like for the convenience of users. For example, when the user selects the latest content category, the user can receive contents corresponding to the latest content as the result value for the selection.

In order to provide other contents (for example, popular contents) by providing only the contents belonging to one category on one screen, the conventional general contents services require the user to separately click the category corresponding to the other contents There is a need to perform input of "

In addition, since conventional general content services do not take into account the characteristics of various users or types of contents, there is a problem in that it can not provide optimum contents customized to the user preference in consideration of the entire contents to be continuously changed and added .

The background technology of the present application is disclosed in Korean Patent Laid-Open Publication No. 10-2013-0026567 (published on March 31, 2013).

SUMMARY OF THE INVENTION The present invention has been made in order to solve the above-mentioned problems of the prior art, and it is an object of the present invention to provide a content service providing method, a content providing method, (E.g., click) of the content recommendation system and method of the present invention.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems of the related art, and it is an object of the present invention to provide a content providing apparatus capable of providing optimal contents customized to a user by considering content recommendation based on statistics, user characteristics, And to provide a recommendation system and method.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a content recommendation system and method capable of providing optimal contents customized to user preferences in consideration of the entire contents to be continuously changed and added .

It is to be understood, however, that the technical scope of the embodiments of the present invention is not limited to the above-described technical problems, and other technical problems may exist.

According to an aspect of the present invention, there is provided a content recommendation system including: a content database storing a plurality of contents; and a plurality of types of analysis performed based on the plurality of contents, Wherein the processor includes a first recommendation unit configured to generate a first recommended content group based on at least one of statistical analysis data and association analysis data on the plurality of contents, A second recommendation unit for analyzing the degree of similarity based on at least one of the recommended person information, the plurality of pieces of user information, and the plurality of contents information, and generating a second recommended content group based on the analyzed degree of similarity, The first recommended content group and the second recommended content group, And a control unit for displaying at least some contents included in the second recommended content group on the display screen as the recommended content.

In addition, the first recommended content group may include a popular content group, a latest content group, an associated content group, and a sequential association content group, and the second recommended content group may include a user-based content group and an item-based content group .

The first recommended content group includes master recommended content, and the control unit displays the contents of the popular content group, the latest content group, the related content group, and the sequentially related content group associated with the master recommended content as the recommended content can do.

In addition, the controller may display the contents of the popular contents group, the latest contents group, the related contents group, and the sequentially related contents group on one page.

The processor may further include an acquisition unit that acquires content related history information of the recommendation target person, and the second recommendation unit generates the second recommended content group according to the analysis result performed based on the content related history information can do.

The processor may further include a user input identification unit for identifying a user input for selecting any one of the recommended contents displayed on the display screen, and the second recommendation unit may further include: the identified user input and the content related history information , The second recommended content group can be generated.

The acquiring unit may subdivide the plurality of pieces of user information according to customer characteristics based on a multi-dimensional analysis (Online Analytical Processing), and the first recommending unit and the second recommending unit may group the first recommended content group and the second recommended content group, The second recommended content group can be generated.

The content recommendation system includes a content database for storing a plurality of contents according to an exemplary embodiment of the present invention, and a processor for providing recommendation contents tailored to a recommendation target through a plurality of types of analysis performed based on the plurality of contents. The content recommendation method of the present invention includes the steps of generating a first recommended content group based on at least one of statistical analysis data and association analysis data on the plurality of contents, Analyzing the degree of similarity based on at least one of the first recommendation content and the second recommendation content and generating a second recommended content group based on the analyzed similarity, At least some of the content contained in the group As recommended contents may comprise a step of displaying on the display screen.

In addition, the first recommended content group may include a popular content group, a latest content group, an associated content group, and a sequential association content group, and the second recommended content group may include a user-based content group and an item-based content group .

In addition, the first recommended content group may include master recommended content, and the displaying may include comparing the content of the popular content group, the latest content group, the related content group, and the sequentially related content group associated with the master recommended content, As shown in FIG.

Also, the displaying step may display the contents of the popular contents group, the latest contents group, the related contents group, and the sequentially related contents group on one page.

The above-described task solution is merely exemplary and should not be construed as limiting the present disclosure. In addition to the exemplary embodiments described above, there may be additional embodiments in the drawings and the detailed description of the invention.

According to the above-mentioned task resolution means of the present invention, the present invention generates a first recommended content group based on at least one of statistical analysis data and association analysis data on a plurality of contents, And analyzing a degree of similarity based on at least any one of the plurality of contents information, generating a second recommended content group based on the analyzed similarity, and analyzing the first recommended content group and the second recommended content group, And a content recommendation system and method for displaying at least a part of contents included in a second recommended content group as the recommended content, thereby providing optimum contents customized to the user preference in consideration of the entire contents continuously added and added There is an effect that can be.

According to the present invention, there is provided a content recommendation system and method capable of displaying contents of a popular content group, a latest content group, an associated content group, and a sequential related content group on one page, Can be easily provided without any additional user input.

The present invention has the effect of providing a more diverse combination of contents on a single screen intuitively in an easy-to-understand format in consideration of the characteristics of the user.

1 is a schematic block diagram of a content recommendation system according to an embodiment of the present invention.
2 is a schematic block diagram of a content recommendation system according to another embodiment of the present application.
3 is a diagram illustrating an example of contents provided by association analysis and sequential association analysis in a content recommendation system according to an embodiment of the present invention.
4 is a diagram illustrating a display example of a main page in a content recommendation system according to an embodiment of the present invention.
FIG. 5 is a diagram schematically showing the overall concept of a content recommendation system 100 according to an embodiment of the present invention.
6 is a flowchart illustrating a content recommendation method according to an exemplary embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. It should be understood, however, that the present invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In the drawings, the same reference numbers are used throughout the specification to refer to the same or like parts.

Throughout this specification, when an element is referred to as being "connected" to another element, it is intended to be understood that it is not only "directly connected" but also "electrically connected" or "indirectly connected" "Is included.

It will be appreciated that throughout the specification it will be understood that when a member is located on another member "top", "top", "under", "bottom" But also the case where there is another member between the two members as well as the case where they are in contact with each other.

Throughout this specification, when an element is referred to as "including " an element, it is understood that the element may include other elements as well, without departing from the other elements unless specifically stated otherwise.

The present invention relates to a technology capable of providing recommended contents customized to a recommendation target person in consideration of the characteristics of the continuously changed and added contents and the recommendation target person and providing various combinations of contents in a form intuitively and easily understandable on one screen .

FIG. 1 is a schematic block diagram of a content recommendation system according to an embodiment of the present invention, and FIG. 2 is a schematic block diagram of a content recommendation system according to another embodiment of the present invention.

1, a content recommendation system 100 according to one embodiment of the present invention may include a content database 110 and a processor 120. The processor 120 may include a first recommender 121, A recommendation unit 122 and a control unit 123. [

2, the content recommendation system 100 according to another embodiment of the present invention includes a content database 110, a first recommendation unit 121, a second recommendation unit 122, and a control unit 123, And the processor 120 may further include an acquiring unit 124 and a user input identifying unit 125. The processor 120 may include an input unit 124,

1 and 2, the content database 110 may store a plurality of contents. Hereinafter, 'content' will be described as an example of a recommendation target, but the recommendation target may include all kinds of 'items' that can be provided to a user via the Internet or mobile. The content database 100 may store content and content attribute information together. For example, the content attribute (or type) information may include content content summary information, an entertainer name, total play time information, viewable age range information, content purchase form, and the like.

The processor 120 may provide recommendation contents tailored to the recommendation target through a plurality of kinds of analysis performed based on a plurality of contents stored in the content database 100. [ A plurality of types of analysis is performed to provide recommended contents, and includes a statistical analysis, association analysis, sequential association analysis, purchase cycle analysis, customer segmentation analysis, and collaborative filtering (CF) based similar members And similar item analysis, which will be described in more detail below.

The first recommendation unit 121 can generate the first recommended content group based on at least one of statistical analysis data and association analysis data on a plurality of contents. At this time, the association analysis data may include data on which association analysis has been performed or data on which sequential association analysis has been performed. The first recommended content group generated through the first recommendation unit 121 may include a popular content group, an up-to-date content group, an associated content group, and a sequentially related content group.

The first recommendation unit 121 may perform at least one of statistical analysis, association analysis, sequential association analysis, and purchase cycle analysis as a plurality of types of analysis in order to generate the first recommended content group.

Statistical analysis is a technique for statistically analyzing a plurality of contents, and popular contents and latest uploaded contents (new contents) can be derived from a plurality of contents through statistical analysis. For example, in the case of statistical analysis, popular contents can be identified by considering the number of times of viewing the contents and the like, and the latest contents can be identified by considering the registration date of contents and the like. The first recommendation unit 121 can generate the popular content group and the latest content group based on the statistical analysis.

The association analysis is a technique for analyzing a content having a high probability of purchasing (or viewing) a specific content together with a specific content, and related contents or the like can be derived from a plurality of contents through association analysis. In the case of association analysis, related contents can be identified by statistically considering contents purchase history or viewing history of a plurality of users. The first recommendation unit 121 can generate an associated content group based on this association analysis.

Sequential association analysis is an analytical technique in which sequence information is considered in association analysis, and sequential association analysis can be used to determine which content is to be purchased after specific content is purchased. The first recommendation unit 121 may generate an associated content group based on this sequential association analysis. The difference between the association analysis and the sequential association analysis can be more easily understood with reference to FIG.

3 is a diagram illustrating an example of contents provided by association analysis and sequential association analysis in a content recommendation system according to an embodiment of the present invention. In FIG. 3, A, B, C, and the like indicate the identification ID of the content.

FIG. 3 (a) shows an example of association analysis, in which user1 purchases content in the order of A? C? B? F, user2 purchases content in the order of C? A? B? I, and user3 purchases B? A? Suppose that you purchased content in the order C → E.

At this time, although users 1 through 3 are different in terms of order in purchasing the contents A, B, and C, users purchasing any one of the contents A, B, You can confirm that you have purchased the remaining content. This means that the A, B, and C contents are related to each other.

Accordingly, the first recommendation unit 121 can generate a group including the A, B, and C contents as an associated content group. Thereafter, when the recommended target person selects (or purchases, views) the A content, the content recommendation system 100 can recommend B and C content as related contents related to the A content.

FIG. 3B shows an example of the sequential association analysis, in which user1 purchases content in the order of A? B? C? D, user2 purchases content in the order of G? A? B? C, → B → F in that order.

At this time, the sequential association analysis considers the order information in the association analysis. Since the user 1 to user 3 commonly purchase the B content after the A content, the sequential association content of the A content may be B.

Accordingly, the first recommendation unit 121 can generate a group including A and B contents as a sequentially related content group. The first recommendation unit 121 may include sequence identification information in each of the A and B contents when generating the sequentially related content group. Thereafter, when the recommended target person selects the A content, the content recommendation system 100 can recommend the B content as the sequential related content related to the A content.

The purchase cycle analysis is a technique for analyzing a purchase cycle of contents based on a history of purchasing contents in the past by a recommendation target person or a search word history input when a content is searched. The first recommendation unit 121 analyzes The first recommended content group can be created. The content recommendation system 100 can use the content-related sales data when analyzing the purchase cycle.

For example, when the recommendation target purchases the domestic football game-related content on a 3-week basis, the first recommendation unit 121 acquires the latest recommendation content group related to the domestic football game It is possible to group contents related to domestic soccer games conducted in the last three weeks rather than grouping contents only. Also, the first recommendation unit 121 can extract and group some popular contents among domestic football-related contents for the last three weeks as a popular content group of the first recommended content group.

For example, assume that 80% of a plurality of users who watched A content watched B content after one month. At this time, the first recommendation unit 121 may generate A and B contents as associated content groups associated with each other in a one-month period. Thereafter, if the current recommendation target identified through the content recommendation system 100 has experienced purchase of the A content about one month ago, the content recommendation system 100 can recommend B content associated with A as recommendation content have.

Based on the statistical analysis, the association analysis, the sequential association analysis and the analysis of the purchase cycle, the first recommendation unit 121 acquires the first recommended content group including the popular content group, the latest content group, the related content group, Can be generated.

The first recommended content group generated through the first recommendation unit 121 may include master recommended content. Generally, in the case of a main page in which content is displayed, a high click rate can be obtained when a content to be noticed as a real-time issue is provided. Therefore, the content recommendation system 100 according to an exemplary embodiment of the present invention, The master recommendation content can be selected as the content. When the type of the page displayed on the user terminal is the main page, the controller 123, which will be described later, can select the master recommendation content from among the contents included in the first recommended content group. At this time, the master recommendation content may be selected from among the first recommended content group, particularly the popular content group or the latest content group, for a high click-through rate.

When the type of the page displayed on the user terminal is a main page, the control unit 123, which will be described later, determines whether or not the content of the popular content group, the latest content group, The content of the sequentially related content group can be displayed on the screen of the user terminal as the recommended content.

More specifically, when the type of the page displayed on the user terminal is the main page, the control unit 123 determines whether the type of the page displayed in the top of the popular content group is one of the first recommended content groups generated through the first recommendation unit 121 The content can be selected as the master recommended content. Then, based on the statistical analysis data and the association analysis data on the master recommendation contents, the first recommendation section 121 acquires the popularity recommendation content from the popularity recommendation content, the latest content group associated with the master recommendation content, It is possible to generate a sequential association content group associated with the linked related content group and the master recommendation content as a first recommended content group. In this case, for example, the popular content group associated with the master recommendation content may be one in which popular content is grouped among related contents associated with the master recommendation content, or related contents associated with the master recommendation content are grouped , Which is applicable to the same logic for the other groups. Then, the control unit 123 may display at least a part of the content among the contents in the first recommended content group associated with the master recommended content as recommended content on the screen. Examples of the recommended content displayed on the screen of the user terminal will be described later in more detail.

The second recommendation unit 122 may analyze the degree of similarity based on at least one of the information on the recommended person, the plurality of pieces of user information, and the plurality of pieces of content information, and generate the second recommended content group based on the analyzed similarity .

The recommendation target person means a target person (user) who receives the recommendation of the content provided through the content recommendation system 100. [ For example, the recommended person information may include at least one of a content usage tendency such as a sex, an age, a content purchase history, a search history related to content search, a content use time slot and a content usage amount, a favorite content genre, One can be included. The plurality of pieces of user information is information of users other than the recommended target, and may include information such as sex, age, and content purchase history as well as the recommended target information. The information on the recommended person and the plurality of pieces of user information may be stored in a user information database (not shown).

The user information database may store recommendation target information or information directly input from a plurality of users (for example, age, sex, favorite entertainer, etc.). In addition, the user information database may store information (e.g., purchase history information, usage time zone, etc.) analyzed based on the recommendation target person or content information purchased by a plurality of users.

According to an embodiment of the present invention, the second recommended content group generated through the second recommendation unit 122 may include a user-based content group, an item-based content group, and a customer segmentation-based content group. Also, the second recommendation unit 122 may generate a second recommended content group based on association analysis, sequential association analysis, collaborative filtering analysis, and customer segmentation analysis.

Collaborative filtering analysis is a technique for recommending an item that the recommendation target user is likely to like using the evaluation value of the recommendation target of the content (or item) and the evaluation value of other users, and includes user-based collaboration filtering and item-based collaboration filtering can do.

The user-based collaborative filtering refers to a technique of selecting recommended content based on what content the user who has a tendency similar to the recommended target purchases. In the case of user-based collaborative filtering, weights are given to the opinions of the users who have similar preference to the recommendation target, and a low weight is given to opinions of users who do not have similar preference, so that the recommendation contents to be provided to the recommendation target Can be selected.

Item-based collaborative filtering refers to a technique of selecting, as recommendation contents, contents having relevance to the purchased contents based on the contents previously purchased by the recommendation target person. In the case of item-based collaborative filtering, it is possible to predict what content the recommendation target person would prefer by measuring the similarity between the existing content that the recommendation target person has already experienced and the preference degree input, and the recommendable content. In the case where a high score is given to a content having a high degree of similarity to the content to be predicted by the recommended target person, the content to be predicted is highly evaluated and can be recommended as a recommended content to the recommendation target person. On the other hand, when the recommendation target person gives a low score to the content highly similar to the content to be predicted, the content to be predicted can be excluded from the recommendation content target. The second recommendation unit 122 may perform similarity analysis using a cosine similarity algorithm as an example.

According to one embodiment of the present invention, the second recommender 122 may generate an item-based content group based on user-based content group and item-based collaborative filtering based on user-based collaborative filtering, Based similar member analysis and collaborative filtering based similar item analysis can be performed. At this time, the second recommendation unit 122 may perform a process of real-time matching of similar members having a tendency similar to the recommended target person according to user input (for example, selection input of recommended content). Then, the second recommendation unit 122 can generate the second recommended content group based on the content history or the like that the similar member has purchased.

Further, the second recommendation unit 122 may generate a second recommended content group based on the collaborative filtering for the similar members matched (for example, a plurality of members (representative similar members) having a high degree of similarity with the recommended target members) The likelihood of similar members for new additional content collected in the stream may be reflected and a Rating Matrix may be generated based thereon. At this time, the preference weight reflected in the new additional content can be determined through a separate setting.

In addition, the second recommender 122 may generate the customer segmentation based content group based on the customer segmentation analysis. The customer segmentation analysis means an analysis technique for recommending contents differentiated according to the characteristics of a recommendation target person when creating a second recommendation content group. To this end, the content recommendation system 100 may classify a plurality of pieces of user information by customer characteristics by applying learning based on online analytical processing (OALP) to a plurality of pieces of user information. For example, the term 'by customer characteristic' means that a plurality of pieces of user information are classified according to characteristics of customers such as gender, age, content usage amount, mania layer, issue follower layer and family center layer. The segmentation by customer characteristic can be subdivided by the acquisition unit 124, which will be described later.

For example, when a first user among a plurality of users purchases mainly family drama related contents among various kinds of contents based on a plurality of pieces of user information (for example, contents purchase history information of a user) , And the first user may be classified as a family-centered layer. Thereafter, the content recommendation system 100 can perform association analysis, sequential association analysis, and the like on the contents purchased by a plurality of users, targeting a plurality of users classified as a family-centered layer. Then, when the identified recommended person purchases mainly the family drama related content, the second recommendation unit 122 searches the second recommended content group based on association analysis, sequential association, and the like performed on the users in the family- Can be generated. The content recommendation system 100 according to an exemplary embodiment of the present invention recommends different combinations of contents for each characteristic (e.g., family-centered floor, gender, and the like) of the recommendable object based on the customer segmentation analysis So that it is possible to provide more optimized customized contents to the recommendation target person.

Also, the second recommendation unit 122 can generate the second recommended content group based on the real-time user input of the recommended person. For this, the second recommendation unit 122 can monitor in real time the behavior taken by the recommendation target person in relation to the content (for example, an action of selecting any one of the recommendation contents displayed on the screen of the user terminal). When the user selects any one of the recommended contents, the second recommendation unit 122 can generate a second recommended contents group related to the selected recommended contents with respect to the selected recommended contents. At this time, the second recommendation unit 122 may generate the second recommended content group by performing collaborative filtering or the like based on the attribute information of the selected recommended content and the characteristics of the recommended person.

In addition, the second recommendation unit 122 can perform association analysis, sequential association analysis, and the like on the added content in real time, and can generate a second recommended content group based on the association analysis. For example, assume that A, B, and C are related contents in the first recommended content group generated through the first recommendation unit 122. [ In this case, when the additional content Z is obtained as a stream through the acquiring unit 124, the second recommending unit 122 may perform the similarity analysis (for example, based on the existing contents A, B, C, For example, association analysis, etc.). At this time, when the similarity degree (i.e., the degree of association) of Z is higher than C as the related contents associated with A and B, the second recommendation section 122 can group A, B and Z as the second recommended content group . Thereafter, when the recommended target person selects the A content, the B and Z contents can be recommended as the recommended content. For example, if the real-time analysis is not performed on the added content or if the user input is not considered in real time, the second recommendation unit 122 selects B and C as recommended contents of the A content based on the first recommended content group Content can be recommended.

Meanwhile, the content recommendation system 100 according to an exemplary embodiment of the present invention may include an acquisition unit 124, and the acquisition unit 124 may acquire content-related history information of a recommended person. At this time, for example, the content related history information may include history information of purchasing the content, search history information used when searching for the content, genre history information of the purchased contents, content history information of the favorite entertainer, and the like . In addition, the acquiring unit 124 may acquire additional content in a stream.

The second recommendation unit 122 may generate a second recommendation content group according to a result of analysis performed on the basis of the content related history information of the recommendation target person (for example, a purchase cycle analysis, an association analysis, an associative sequential analysis, a collaboration filtering, Can be generated. For example, the second recommendation unit 122 can perform the purchase cycle analysis based on the purchase history information of the recommendation target person, and based on the purchase cycle analysis, Can be identified. In addition, the second recommender 122 may generate a second recommended content group based on user-based collaborative filtering for identified similar users.

In addition, the acquiring unit 124 may subdivide a plurality of pieces of user information according to customer characteristics based on the multidimensional analysis, and the first recommending unit 121 and the second recommending unit 122 may group the first recommended content group And a second recommended content group. More specifically, as the acquiring unit 124 subdivides a plurality of pieces of user information by customer characteristics, a combination of recommended content or recommended sequential related content for each customer characteristic may be changed. Based on the combination of the recommended contents that varies depending on the customer characteristics, the first recommendation unit 121 and the second recommendation unit 122 may generate the first recommendation content group < RTI ID = 0.0 > And a second recommended content group.

The content recommendation system 100 according to an embodiment of the present invention may further include a user input identification unit 125. The user input identification unit 125 may identify any one of the recommended contents displayed on the display screen of the user terminal Can be identified. Then, the second recommendation unit 122 refers to the second recommended content group on the basis of the user input identified through the user input identification unit 125, attribute information of the content selected by the user input, Can be generated.

For example, when the recommended person selected from the recommended contents displayed on the display screen selects the popular contents related to the gag program, the second recommendation unit 122 stores attribute information of the selected contents (e.g., guest artist information ) And the content-related history information of the recommendation target person. In this case, if a large number of entertainment-related contents appearing in the gag program are found in the content-related history information of the recommended target person, the second recommendation unit 122 may determine that the guest performer Related other contents as a second recommended content group. The generated second recommended content group may be an item-based content group based on collaborative filtering.

The control unit 123 may display at least some contents included in the first recommended content group and the second recommended content group as recommended contents on the display screen according to the type of the page displayed on the user terminal.

The display screen on which the recommended content is displayed may be a screen of all display devices on which contents such as a smart phone, a mobile phone, a PDA, a notebook, a desktop PC, and a TV can be displayed.

Examples of the types of pages displayed on the user terminal include a main page and an end page.

If the displayed page is the main page, the control unit 123 may display some of the contents in the first recommended content group as recommended contents. When the recommended content is displayed on the main page, the control unit 123 displays at least some of the contents included in each of the popular content group, the latest content group, the related content group, and the sequentially related content group on one page can do.

At this time, the number of contents to be displayed on the main page, that is, the number of contents to be recommended as recommended contents, for each of the popular, latest, association, and sequential related content groups may be set based on a predetermined number. For example, the number of contents to be displayed on the main page may be set in advance such as one in the popular content group, five in the latest content group, two in the related content group, and two in the sequentially related content group.

Also, the number of recommended contents calculated in each of the popularity, latest, association, and sequential association content groups may be determined by assigning weights to the respective content groups based on the content-related history information of the recommended persons.

For example, if the recommendation target person does not see any content corresponding to the latest content among a plurality of kinds of contents displayed on the main page (i.e., popularity, latest, association, and sequential association contents) Suppose you watched content at a ratio of 5: 3: 2, respectively. At this time, weights of 5: 0: 3: 2 may be respectively assigned to each of the popularity, latest, association, and sequential association content groups. Thus, the main page can display 5 pieces of popular content, 3 pieces of related contents, and 2 pieces of related contents sequentially.

Also, when the recommended content is displayed on the main page, the process of selecting the master recommended content may be performed for the high click rate of the recommended content as described above. The master recommended content may be classified into the popular content group or the latest content group Lt; / RTI > If the master recommendation content is selected, the contents of the popular, newest, association, and sequential association content groups associated with the master recommendation content may be displayed as recommended content on the main page. A display example of the main page can be more easily understood with reference to Fig.

4 is a diagram illustrating a display example of a main page in a content recommendation system according to an embodiment of the present invention.

FIG. 4 (a) shows a configuration example of content for each of popular, latest, association, and sequential related content groups generated through the first recommendation unit 121. FIG. The first recommendation unit 121 can generate the respective groups based on statistical analysis data and association analysis data. The popular content group may include contents such as A1, A2, and A3. The latest contents group may include contents such as B1, B2, and B3. The related content group may include contents such as C1, C2, C3, and the like. The sequential association content group may include contents such as D1, D2, and D3.

When the main page is displayed on the user terminal, the first recommendation unit 121 can select A1 content as the master recommendation content in FIG. 4 (a). Then, the first recommendation unit 121 may generate popular, newest, associative, and sequential related content groups associated with the A1 content, which is the master recommended content.

Referring to FIG. 4B, the popular content group associated with the A1 content may include content such as A2 ', A3', and the like. The latest content group associated with the A1 content may include content such as B1 ', B2', B3 ', and the like. The related content group associated with the A1 content may include content such as C1 ', C2', C3 ', and the like. Contents of D1 ', D2', D3 ', etc. may be included in the sequentially related content group associated with the A1 content.

Thereafter, the control unit 123 can select at least a part of the contents included in each group of FIG. 4B as the recommended contents, and display the selected recommended contents on the screen. For example, the control unit 123 may display A1, B1 ', B2', C1 'and D1' as recommended content.

4 (c) shows an example of a main page displayed on a display screen of a TV as an example. The A1 popular content that is the master recommended content on the main page can be displayed in a larger area than the other contents. In addition, the latest contents B1 'and B2', the related contents C1 'and the sequentially related contents D1' can be displayed in one area on the screen, and can be displayed in a narrow area compared to the master recommended contents.

The content recommendation system 100 according to an exemplary embodiment of the present invention provides a variety of combinations of contents associated with master contents in a form intuitively and easily understandable on one screen, Optimized customized content can be recommended. That is, the content recommendation system 100 according to an exemplary embodiment of the present invention can provide recommendable personalized contents with improved diversity and accuracy.

Also, the control unit 123 can display the recommended content in consideration of the characteristics (e.g., broadcasting company) of the main page. For example, if the screen displayed on the user terminal among the A, B, and C broadcasters is the main page of the A channel, the controller 123 determines whether the content included in A The contents and the sequentially related contents can be displayed on the screen as recommended contents.

On the other hand, the end page may be a page displayed when the recommendation target person selects any one of a plurality of contents displayed on the screen, and may be a page displayed based on a user input of the recommendation target person. In the end page, when the user input is identified from the recommended person, based on at least one of information or characteristics of the recommended person, information of a plurality of users, customer characteristics, and attribute information of the selected contents, Content can be provided.

When the user selects a specific content from a plurality of contents, it is highly likely that the selected specific content is a content field (for example, entertainment, documentary, etc.) preferred by the recommendation target person or a content in which the interested star appears.

Accordingly, when the type of the page displayed on the user terminal is the end page, the control unit 123 can display some of the contents included in the second recommended content group as recommended content. At this time, the second recommended content group may include an analysis (e.g., similarity analysis, association analysis, sequential association analysis, similar membership based on collaborative filtering, and the like) performed based on information of the recommended target, Similar item analysis and customer segmentation analysis, etc.).

In the end page, some content among user-based content based on collaborative filtering, item based content based on collaborative filtering, association analysis based content, sequential association analysis content and customer segmentation based content can be displayed as recommended content.

In the content recommendation system 100 according to an exemplary embodiment of the present invention, at least one of popular, latest, association, and sequential related contents is displayed on the main page, and the end page includes user- Based content can be displayed. However, the present invention is not limited to this, and the user-based content, the customer segmentation-based content, and the like may be displayed on the main page, and the popularity, contents, May be displayed.

In addition, the content recommendation system 100 according to an exemplary embodiment of the present invention may concurrently generate a first recommended content group or a second recommended content group based on real-time behavior of a recommended target. At this time, the real-time behavior of the recommended person may mean, for example, an action of selecting any one of a plurality of contents, an action of viewing an end page, and an action of viewing a main page.

FIG. 5 is a diagram schematically showing the overall concept of the content recommendation system 100 according to an embodiment of the present invention, and the entire concept will be briefly described based on the contents described above in detail.

Referring to FIG. 5, a content recommendation system 100 according to an embodiment of the present invention may store a plurality of contents in a content database 110 (Data Mart). The content database 110 may store content, content attribute information, and content purchase form.

The processor 120 may perform at least one of a plurality of analyzes 10 for a plurality of contents according to respective characteristics when a user input is received from a recommendation object or additional contents are received as a stream. The content recommendation system 100 according to an exemplary embodiment of the present invention may include an analysis algorithm corresponding to each of the plurality of analyzes 10 and may be customized to a recommended person based on various analysis results calculated through a plurality of analysis algorithms Recommendable content can be recommended.

The first recommendation unit 121 may perform rule matching in real time when user input is identified through the user input identifying unit 125 or additional content is received as a stream. In this way, the first recommendation unit 121 can generate the first recommended content group 121a. The first recommended content group 121a may include a popular, up-to-date, association, and sequential association content group. The first recommendation unit 121 may generate the first recommended content group 121a by analyzing at least one of statistics, association, sequential association, and purchase cycle analysis.

When the user input is identified through the user input identification unit 125 or the additional content is received as a stream, the second recommendation unit 122 transmits the identified user input information, the content of the recommended target acquired through the acquisition unit 124 Related history information, and the similarity matching based on collaborative filtering (CF) can be performed in real time. Accordingly, the second recommendation unit 122 can generate the second recommended content group 121b. The second recommended content group 121b may include a CF-based user-based content group, a CF-based item-based content group, and a customer segmentation-based content group. The second recommendation unit 122 generates the second recommended content group 121b by analyzing at least one of association, sequential association, CF-based similar member analysis, CF-based similar item analysis, and customer segmentation analysis .

The control unit 123 displays at least some of the contents included in the first recommended content group 121a and the second recommended content group 121b as recommended contents according to the type of page displayed on the user terminal .

The control unit 123 generates the first recommended content group 121a or the second recommended content group 121b based on the real-time behavior of the recommended target obtained through the acquisition unit 124 to the user input identification unit 125, And it is possible to provide the recommended contents optimized for the situation according to the real time behavior of the recommended person.

Hereinafter, the operation flow of the present invention will be briefly described based on the details described above.

6 is a flowchart illustrating a content recommendation method according to an exemplary embodiment of the present invention. The content recommendation method illustrated in FIG. 6 may be performed by the content recommendation system 100 described with reference to FIGS. 1 to 5. FIG. Accordingly, the content described with respect to the content recommendation system 100 through FIG. 1 through FIG. 5 may be similarly applied to FIG. 6 even if omitted below.

Referring to FIG. 6, a content recommendation system includes a content database for storing a plurality of contents, and a processor for providing recommendation contents tailored to a recommendation target through a plurality of types of analysis performed based on the plurality of contents. In operation S610, the first recommendation unit 121 may generate a first recommended content group based on at least one of statistical analysis data and association analysis data related to a plurality of contents.

In step S610, the first recommended content group generated by the first recommendation unit 121 may include a popular content group, an up-to-date content group, an associated content group, and a sequentially related content group.

In addition, in step S610, the first recommended content group may include master recommended content.

Next, in step S620, the degree of similarity is analyzed based on at least one of the recommended person information, the plurality of pieces of user information, and the plurality of pieces of content information through the second recommendation unit 122, and based on the analyzed degree of similarity, It is possible to create a recommended content group.

In step S620, the second recommended content group generated by the second recommendation unit 122 may include a user-based content group, an item-based content group, and a customer segmentation-based content group.

Next, in step S630, at least some contents included in the first recommended content group and the second recommended content group are displayed on the display screen as recommended contents according to the type of the page displayed on the user terminal through the control unit 123 .

At this time, in step S630, the controller 123 may display the contents of the popular contents group, the latest contents group, the related contents group, and the sequentially related contents group associated with the master recommended contents as recommended contents.

Also, in step S630, the control unit 123 may display the contents of the popular contents group, the latest contents group, the related contents group, and the sequentially related contents group on one page.

In the above description, steps S610 to S630 may be further divided into additional steps or combined into fewer steps, according to embodiments of the present disclosure. Also, some of the steps may be omitted as necessary, and the order between the steps may be changed.

The content recommendation method according to an exemplary embodiment of the present invention may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the medium may be those specially designed and configured for the present invention or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media 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. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

The content recommendation method described above can also be implemented in the form of a computer program or an application executed by a computer stored in a recording medium.

It will be understood by those of ordinary skill in the art that the foregoing description of the embodiments is for illustrative purposes and that those skilled in the art can easily modify the invention without departing from the spirit or essential characteristics thereof. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. For example, each component described as a single entity may be distributed and implemented, and components described as being distributed may also be implemented in a combined form.

The scope of the present invention is defined by the appended claims rather than the detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalents should be construed as being included within the scope of the present invention.

100: Content recommendation system
110: Content database
120: Processor
121: First recommendation section
122: second recommendation section
123:

Claims (12)

A content database for storing a plurality of contents; And
A processor for providing recommendation contents tailored to a recommendation target through a plurality of types of analysis performed based on the plurality of contents,
, ≪ / RTI &
The processor comprising:
A first recommendation unit for generating a first recommended content group based on at least one of statistical analysis data and association analysis data on the plurality of contents;
A second recommendation unit for analyzing the degree of similarity based on at least one of the recommended person information, the plurality of pieces of user information, and the plurality of contents information, and generating a second recommended content group based on the analyzed similarity; And
A control unit for displaying at least some contents included in the first recommended content group and the second recommended content group on the display screen as the recommended content according to the type of page displayed on the user terminal,
And a content recommendation system. The method according to claim 1,
Wherein the first recommended content group includes a popular content group, a latest content group, an associated content group, and a sequentially related content group,
Wherein the second recommended content group comprises a user based content group and an item based content group. 3. The method of claim 2,
Wherein the first recommended content group includes master recommended content,
Wherein the controller displays the content of the popular content group, the latest content group, the related content group, and the sequentially related content group associated with the master recommended content as the recommended content. 3. The method of claim 2,
Wherein the controller displays content of the popular content group, the latest content group, the related content group, and the sequentially related content group on one page. The method according to claim 1,
The processor comprising:
Further comprising an acquiring unit for acquiring content-related history information of the recommended person,
Wherein the second recommendation unit generates the second recommended content group according to an analysis result performed based on the content related history information. 6. The method of claim 5,
The processor comprising:
Further comprising a user input identifying unit for identifying a user input for selecting one of the recommended contents displayed on the display screen,
Wherein the second recommendation unit,
And generates a second recommended content group based on the identified user input and the content related history information. 5. The method of claim 4,
Wherein the obtaining unit comprises:
And segmenting the plurality of pieces of user information by customer characteristics based on multi-dimensional analysis (online analytical processing)
Wherein the first recommendation unit and the second recommendation unit generate the first recommended content group and the second recommended content group for each of the customer characteristics. A content recommendation method of a content recommendation system, comprising: a content database storing a plurality of contents; and a processor for providing recommendation contents tailored to a recommendation target through a plurality of types of analysis performed based on the plurality of contents,
Generating a first recommended content group based on at least one of statistical analysis data and association analysis data on the plurality of contents;
Analyzing a degree of similarity based on at least one of the recommended person information, the plurality of pieces of user information, and the plurality of contents information, and generating a second recommended content group based on the analyzed similarity; And
Displaying at least some contents included in the first recommended content group and the second recommended content group on the display screen as the recommended content according to the type of page displayed on the user terminal,
And a content recommendation method. 9. The method of claim 8,
Wherein the first recommended content group includes a popular content group, a latest content group, an associated content group, and a sequentially related content group,
Wherein the second recommended content group comprises a user based content group and an item based content group. 10. The method of claim 9,
Wherein the first recommended content group includes master recommended content,
Wherein the displaying step displays the contents of the popular contents group, the latest contents group, the related contents group and the sequentially related contents group associated with the master recommended contents as the recommended contents. 10. The method of claim 9,
Wherein the displaying step displays the content of the popular content group, the latest content group, the related content group, and the sequentially related content group on one page. A computer-readable recording medium recording a program for executing the method of any one of claims 8 to 11 on a computer.

Images