KR102096474B1 - Apparatus and method for recommending contents - Google Patents

Abstract

Regarding a content recommendation device that recommends appropriate content to one multimedia device when used by multiple users, according to an embodiment, the content recommendation device calculates similarity between channels based on attribute information of the channel, and calculates the channel A channel group generating unit for generating one or more channel groups using similarity between channels, a viewing history replacement unit for replacing content viewing history information for each user group with viewing history information for each channel group, and a channel group based on viewing history information for each channel group It may include a recommendation information generation unit for generating content recommendation information for each user group.

Description

Content recommendation device and method {APPARATUS AND METHOD FOR RECOMMENDING CONTENTS}

It relates to a content recommendation apparatus and method, and more specifically, to create a channel group according to channel properties in one multimedia device that can be used by a plurality of users, such as a TV, and analyzes the viewing history form of users for each channel group. It relates to a device and method for recommending suitable content for each group.

Unlike devices that are continuously used by specific individuals, such as mobile terminals such as smartphones, in the case of services using TV as a medium, such as cable TV, IPTV, satellite broadcasting, etc., various members constituting one generation use one device Is done. In this situation, each member gathers all at the same time to watch the same content, but in general, different kinds of content are watched at different times because the preferences of the preferred content are different.

In the case of general personalization / recommendation algorithms, such as association rule mining, it analyzes an individual's usage patterns or patterns and provides content suitable for individual preferences. However, when a large number of users, such as a TV, use the same service in common through one device, it is necessary to specify each individual first. To this end, in general, a method of receiving a profile of each individual who wants to use the TV is currently being tried through a separate user interface, but this causes inconvenience in terms of usability, and in fact, users specify their profile every time. It is also difficult to induce people to use TV.

 In addition, when the preference information is generated using the viewing information generated by one household and the content is recommended accordingly, there is a problem in that the content that is not related to the household member who actually uses the TV is also generated. For example, when a large number of users watch a single device, such as a TV, viewing patterns may vary for each time zone even in one day. In other words, in the morning, the patterns of watching beauty or drama-related programs usually watched by housewives are mainly watched for children's programs in the afternoon, and watching sports or movies in the evening and night. When the related rules are mined by integrating them into the usage history of one device, problems such as recommending an adult program to users watching children's programs such as animation may occur.

Published Korean Patent No. 10-2011-0071715, released on June 29, 2011 (Name: IPTV service system providing community service)

By categorizing each channel broadcast on a device such as a TV by channel group, and recommending suitable content for each channel group, the user can recommend content suitable for individual users without the need for multiple users to input a profile or the like. This is to provide an apparatus and method for recommending content.

According to an aspect, the content recommendation apparatus calculates the similarity between channels based on the attribute information of the channel, and a channel group generation unit that generates one or more channel groups by using the calculated similarity between channels, content viewing history information for each user group It may include a viewing history replacement unit for substituting the viewing history information for each channel group and a recommendation information generating unit for generating content recommendation information for each user group for each channel group based on the viewing history information for each channel group.

In addition, the content recommendation device may include one or more of electronic program guide information for each channel, meta information for each channel, meta information of content broadcasted on each channel, and web information. It may further include a broadcast information collection unit for collecting broadcast information.

In addition, the content recommendation device may further include a channel attribute generator that generates attribute information of each channel using the collected broadcast information.

The channel group generation unit may check content duplication information between channels using the collected broadcast information, and calculate similarity between channels based on the content duplication information.

At this time, the content duplication information may include one or more of whether content between channels is duplicated and the number of duplicate content.

In addition, the content recommendation apparatus further includes a content affiliation calculation unit that calculates a degree of belonging for each channel of content broadcasted in each channel, and a channel group update unit that updates a relationship between channel groups generated based on the affiliation of the calculated content. can do.

The content affiliation calculation unit may calculate the affiliation for each channel of the content based on the association between the attribute of each channel and the type of content broadcasted in each channel.

The channel group generation unit may compare the calculated similarity between channels and a preset threshold, and generate channels as one or more channel groups based on the result.

As a result of the creation of the channel group, the channel group generation unit may generate an upper channel group for the two or more channel groups when any one channel belongs to more than one channel group.

The recommendation information generating unit may calculate content preferences for each channel group based on viewing history information for each channel group, and generate content recommendation information for each channel group based on the calculated content preferences.

In addition, the content recommendation apparatus may further include a recommendation information providing unit that provides content recommendation information for each user group for each user group to each user group.

The recommendation information providing unit may provide content recommendation information of the channel group to which the currently selected channel belongs from each user group to each user group.

The recommendation information providing unit may provide each user group with content recommendation information of a channel group broadcasting content similar to the content currently being viewed by each user group based on the degree of belonging to each channel of the content.

According to an aspect, the content recommendation method includes calculating the similarity between channels based on the attribute information of the channel, generating one or more channel groups using the calculated similarity between channels, and viewing the content viewing history information for each user group. The method may include substituting the viewing history information for each group and generating content recommendation information for each user group for each channel group based on the viewing history information for each channel group.

In addition, the content recommendation method includes one or more of electronic program guide information for each channel, meta information for each channel, meta information of content broadcasted on each channel, and web information. The method may further include collecting broadcast information, and generating attribute information of each channel using the collected broadcast information.

The step of generating a channel group may include checking content duplication information between channels using the collected broadcast information, and calculate similarity between the channels based on the identified content duplication information.

In addition, the content recommendation method may further include calculating a degree of belonging for each channel of content broadcasted in each channel and updating a relationship between channel groups generated based on the calculated degree of belonging of the content.

The step of generating a channel group includes comparing the similarity between the calculated channels and a preset threshold, generating channels into one or more channel groups based on the comparison result, and generating a group of channels. If it belongs to a channel group, it may include generating a higher channel group for the two or more channel groups.

In addition, the content recommendation method may further include providing content recommendation information for each user group generated for each user group to each user group.

When multiple users use the same device, such as TV, IPTV, CSTV, OTT service, etc., each user recommends suitable content for the group of their preferred channel according to the viewing history type of each group of channels broadcast on the device Can receive

In addition, each user does not need to perform unnecessary interaction to input profile information or the like in order to recommend his / her favorite content, thereby improving user convenience.

1 shows a content recommendation system according to an embodiment.
2 is a block diagram of a content recommendation apparatus according to an embodiment.
3 is an example of broadcast information collected by a content recommendation apparatus according to an embodiment.
4 is a diagram for explaining that the content recommendation apparatus according to an embodiment calculates similarity between channels.
5 to 8 are diagrams for explaining that the content recommendation apparatus according to an embodiment generates a channel group.
9 and 10 are views for explaining viewing history information collected by the content recommendation apparatus according to an embodiment.
11 and 12 are views for explaining viewing history information for each channel group substituted by the content recommendation apparatus according to an embodiment.
13 is a flowchart of a content recommendation method according to an embodiment.

If described in detail with reference to the accompanying drawings the present invention. Here, repeated descriptions and detailed descriptions of well-known functions and configurations that may unnecessarily obscure the subject matter of the present invention are omitted. Embodiments of the present invention are provided to more fully describe the present invention to those skilled in the art.

The terms or words used in the present specification and claims described below should not be construed as being limited to ordinary or dictionary meanings, and the inventor is appropriate as a concept of terms for explaining his or her invention in the best way. Based on the principle that it can be defined, it should be interpreted as meanings and concepts consistent with the technical spirit of the present invention. Therefore, the configuration shown in the embodiments and drawings described in this specification is only one of the most preferred embodiments of the present invention, and does not represent all of the technical spirit of the present invention, and thus can replace them at the time of application. It should be understood that there may be equivalents and variations. In addition, terms such as first and second are used to describe various components, and are used only to distinguish one component from other components, and are not used to limit the components.

Hereinafter, various embodiments of a content recommendation apparatus and method will be described with reference to the drawings.

1 is a block diagram of a content recommendation system according to an embodiment.

Referring to FIG. 1, the content recommendation system 1 according to an embodiment may include a content recommendation device 100 and one or more user groups 200.

Each user group 200 is composed of a plurality of users using one multimedia device. Here, the multimedia device is installed in a specific place (for example, one home) and used by multiple members, such as cable TV, IPTV, CSTV (Cloud Streaming TV), and OTT (Over The Top) services such as on-demand VOD It may include a service providing device. At this time, the CSTV means that the server processes the change of the operation screen or image without a set-top box, and the client receives only the screen generated by the server in real time through an Internet connection. In addition, OTT service means an internet video service.

The content recommendation device 100 may collect broadcast information from an electronic program guide (EPG), web and channel information, and the like. For example, the broadcast information may collect program names collected from the electronic program guide, program description, meta information provided for each channel, broadcast information published on the web, and meta information of content broadcasted on each channel.

When the broadcast information is collected in this way, the content recommendation apparatus 100 may define the properties of each channel using the collected broadcast information. In addition, when attributes of each channel are defined, channels having similar attributes may be classified into one channel group to generate content recommendation information for each channel group.

In addition, the content recommendation device 100 may collect viewing history information from each user group 200. At this time, the collected viewing history information may include channel information in each user group 200 and content information viewed for each channel.

When the viewing history information is collected from each user group, the content recommendation device 100 replaces the collected viewing history information with the viewing history for each channel group, and generates the recommended content information in units of each channel group to generate a user group 200 Can be provided on.

At this time, the content recommendation device 100 may provide the recommended content information of the channel group corresponding to the selected channel when a specific user in the specific user group 200 selects a channel to watch. Alternatively, the content recommendation apparatus 100 may provide recommendation content generated for a channel group different from the current specific channel based on the similarity of content currently viewed by a specific user on a specific channel.

The content recommendation apparatus 100 according to an embodiment will be described in more detail with reference to FIG. 2 below. Hereinafter, for convenience of description, a TV installed in each user group will be described as an example.

2 is a block diagram of a content recommendation apparatus according to an embodiment. 3 is an example of broadcast information collected by a content recommendation apparatus according to an embodiment. 4 is a diagram for explaining that the content recommendation apparatus according to an embodiment calculates similarity between channels.

Referring to FIG. 2, the content recommendation device 100 includes a broadcast information collection unit 110, a channel attribute generation unit 120, a channel group generation unit 130, a viewing history collection unit 140, and a viewing history replacement unit ( 150), a recommendation information generation unit 160, a recommendation information providing unit 170, a content belonging degree calculation unit 180, and a channel group update unit 190.

The broadcast information collection unit 110 may collect broadcast information from an electronic program guide, the web, channel meta information, content meta information, and the like. For example, as illustrated in FIG. 3, an electronic program guide, channel information such as broadcast content name, broadcast content description, genre, channel name, channel description from each channel's meta information, and broadcast information published on the web, each It is possible to collect broadcast information 10 such as a content name, actor, director, synopsis, and producer through meta information of content broadcasted on the channel.

The channel attribute generator 120 may define attributes of each channel using the collected broadcast information and generate attribute information of each channel. At this time, the properties of the channel may be classified into "specialized channel for movies", "specialized channel for drama", "children's channel", "education channel", "economic broadcasting channel", "news channel", and the like.

For example, referring to the broadcast information 10 illustrated in FIG. 3, channels # 1 and channel # 3 may be defined as a “movie channel” because the main content to be broadcast is a movie. Similarly, since channel # 2 mainly broadcasts content such as children's cartoons and early childhood education, it may be defined as a "children's channel", and channel #N may be defined as a "drama professional channel".

When the attribute of each channel is defined, the channel group generating unit 130 may calculate the similarity between each channel by using the attribute information of the generated channel. At this time, N (N-1) / 2 similarities may be calculated for N channels, and may be calculated using a jaccard similarity or cosine similarity calculation technique.

For example, referring to FIGS. 3 and 4, since channel # 1 and channel # 3 of FIG. 3 overlap only the movie # 3 among the broadcasted content, it can be said that the degree of content overlap is low. However, since channel # 1 and channel # 3 have both defined channel properties as "film specialized channels", the similarity can be calculated relatively high as shown in FIG. 4.

Additionally, the channel group generation unit 130 may check overlapping information of the content broadcasted in each channel, and calculate similarity based on the checked duplicated content information. At this time, the content duplication information may include information on whether or not the content is duplicated, and if duplicated, the number of duplicated content. If the content is duplicated, the higher the number of duplicated content, the higher the similarity can be calculated.

For example, the properties of channel # 3 and channel #N are differently defined as "film specialized channel" and "drama specialized channel", and according to the channel properties, similarity will be calculated low. However, since both channels # 3 and #N are broadcasting a drama, and drama # 2 in the broadcast drama is overlapped with each other, as shown in FIG. 4, the similarity can be calculated somewhat higher than when only channel attributes are used.

As such, the channel group generation unit 130 may calculate more accurate similarity by considering not only the channel attribute information generated for each channel, but also overlapping information of content broadcasted in each channel.

In addition, when the similarity between each channel group is calculated as illustrated in FIG. 4, the channel group generation unit 130 may generate one or more channel groups using the calculated similarity.

5 to 8 are diagrams for explaining that a content recommendation apparatus generates a channel group according to an embodiment, and will be described with reference to FIGS. 5 to 8 for example.

The channel group generating unit 130 compares the calculated similarity with a preset similarity threshold to generate channels having similarities exceeding the similarity threshold as one or more channel groups. At this time, if the similarity has a value between 0 and 1, the preset similarity threshold may be preset to an arbitrary value (for example, 0.5) so as to be generated as an appropriate number of channel groups compared to the number of channels.

For example, the similarity between channel # 1, channel # 2, channel # 3, and channel #N is greater than the similarity threshold (eg, 0.5), and the similarity between channel # 4, channel # 5 and channel # 6 Assuming that it is greater than the similarity threshold (eg, 0.5), it may be generated as channel group # 1 and channel group # 2, respectively, as shown in FIGS. 5 and 6.

At this time, each channel group may have a correlation with each other. That is, as shown, similarity exists between channels # 3 and # 4, but can be correlated with each other by having similarity not exceeding a preset similarity threshold.

In addition, channels whose similarity does not exceed the similarity threshold are generated as one channel group (eg, channel group # 3), and this channel group # 3 may not be correlated with other channel groups # 1 and # 2. have. In addition, a channel that is not similar to all other channels, for example, a channel having a similarity level of 0 between each channel may be generated as one channel group.

Meanwhile, the preset similarity threshold may have range information consisting of two or more values, for example, an upper limit (eg 0.7) and a lower limit (eg 0.3).

For example, assuming that the similarity between the channels # 1, channel # 2 and channel #N and the similarity between the channels # 4, channel # 5 and channel # 6 all exceed the upper limit (for example, 0.7), the channel group As illustrated in FIG. 7, the generation unit 130 generates channel # 1, channel # 2, and channel #N as one channel group # 1, and also creates channel # 4, channel # 5, and channel # 6 as another channel. Can be created with channel group # 2.

Also, assuming that the similarity between channel # 3 and channel # 2 is 0.4, the similarity between channel # 3 and channel #N is 0.6, and the similarity between channel # 3 and channel # 4 is 0.5, channel # 3 is channel It belongs to both group # 1 and channel group # 2, and as illustrated in FIG. 8, an upper channel group # 100 connecting channel group # 1 and channel group # 2 may be generated. In this case, a channel group consisting of two or more levels (eg, L1, L2) may be generated, and the number of channel groups generated may be greater than the number N of channels.

Also, similarly, channels having similarities smaller than the lower limit (eg, 0.3) are created as one channel group (eg, channel group # 3), and if not, When dissimilar channels exist, the channels may be independently generated as a single channel group.

The channel group generation unit 130 may generate and manage one or more channel groups generated as shown in FIGS. 6 and 8 in a hierarchical tree form.

Meanwhile, the content belonging degree calculating unit 180 may calculate the belonging degree for each channel of the content broadcasted in each channel using the collected broadcast information. For example, the "movie" content broadcasted on the "channel specialized in movies" may have a high degree of belonging to the channel, and the "drama" content broadcasted on the "specialized channel of drama" may have a high degree of belonging. That is, if the belonging value has a value between 0 and 1, in the above case, the belonging value may be calculated as '1'. On the other hand, the "drama" content broadcasted on the "movie channel" may have a relatively low affiliation value (eg, 0.7) compared to the "movie".

In this way, content broadcasted in a plurality of channels can be extracted, and affiliation values for each channel of each content can be calculated. In this case, specific content may have a value of belonging to more than one channel.

The channel group update unit 190 may update the relationship of each channel group generated by using the content belonging degree value calculated by the content belonging degree calculating unit 180. At this time, the channel group generation unit 190, if the specific content belongs to two or more channels according to the calculated content belonging, and two or more channels belong to different channel groups, the different channels as shown in FIG. You can define the relationship between channel groups by creating an upper channel group for the group.

For example, when content # 1 is broadcast on channels # 1 and # 3, an association relationship between channel group # 1 to which channel # 1 belongs and channel group # 2 to which channel # 3 belongs is generated, which is again channel group # Channel # 2 belonging to 1 and channel # 3 are connected. Through this process, a correlation may also occur between channels having no content broadcast simultaneously to each other, and based on this, content information may be recommended to each other.

9 and 10 are views for explaining viewing history information collected by the content recommendation apparatus according to an embodiment.

The viewing history collection unit 140 collects viewing history information from each user group. For example, as shown in FIGS. 9 and 10, viewing history information 20 for a predetermined period (eg, week, month, quarterly, yearly) from a specific user group (eg, household # 1) Can be collected. At this time, the collected viewing history information 20 may include viewing channels for each user group, viewing time, and content information broadcasted in each viewing channel.

Referring to the viewing history information 20 of FIGS. 9 and 10, furniture # 1 watches content # 1 and content # 2 on channel # 1 in chronological order on Monday, and then views content # 3 on channel # 2 It shows the form of watching content # 4 on channel # 3.

11 and 12 are views for explaining viewing history information for each channel group substituted by the content recommendation apparatus according to an embodiment.

When the viewing history information 20 is collected from each user group, the viewing history substitution unit 150 records the viewing history information for each user group collected by the channel group generation unit 130 by the channel group generation unit 130. It can be replaced with information. That is, as shown in FIG. 11, the viewing history substitution unit 150 uses the channel information in the viewing history information 20 collected from each user group to replace each channel with the channel group to which the channel belongs, and the viewing history. Information 30 may be generated.

For example, if channel # 1 and channel # 3 are the same channel group # 1, and channel # 2 and channel # 4 are separate channel groups # 2 and # 3, respectively, as shown in FIG. 11, channel # The viewing history of 1 and channel # 3 may be replaced with the viewing history of channel group # 1.

At this time, one content may belong to a plurality of channel groups according to the content belonging to the calculated content, as described above, when viewing the content belonging to a plurality of channel groups in a specific channel as described above, watching for a plurality of channel groups History can be generated.

As illustrated in FIG. 11, the recommendation information generating unit 160 calculates content preferences for each channel group based on the viewing history information 30 for each channel group, and each channel group for each channel group based on the calculated content preferences. Content recommendation information for a user group may be generated.

In this case, the recommendation information generation unit 160 may assume each user group as one user, and generate content recommendation information for each channel group for each user group, not each member of each user group. If there are M channel groups, recommendation information for up to M channel groups for each user group may be generated.

For example, the viewing history information (30a, 30b, 30p) for each channel group generated for each household (# 1, # 2 ~ #P) of FIG. 12 and the generated channel group (# 1, # 2, # Referring to the viewing history information 40a, 40b, and 40c for 3), both furniture # 1 and furniture # 2 watched content # 1, content # 2, and content # 4 in channel group # 1. At this time, if the type of content # 1, content # 2, and content # 4 is an action movie, the preference of the content of the action movie is set high, and among the action movies, the content recommendation is made among the action movies that furniture # 1 and furniture # 2 have not watched. Information can be generated. At this time, additionally, content information such as actors, directors, release dates, and plots appearing in content # 1, content # 2, and content # 4 are further analyzed, and other action movies and directors appearing in the actors appearing in the content The content preferences of other produced action movies and the like can be set high and content recommendation information can be generated using the content preference.

In addition, since furniture # 1 did not watch content # 5, it shows similar viewing history within the same channel group as furniture # 2 showing similar viewing history generates content # 5 viewed as recommended content in furniture # 1. It is also possible to recommend content viewed by other households with high preference.

As described above, according to the present embodiment, the recommendation information generation unit 160 may generate content preferences according to various criteria and generate content recommendation information according to the preferences, but is not limited to the above-described examples.

When the content recommendation information for each user group is generated for each channel group, the recommendation information providing unit 170 provides the content recommendation information to each user group. In this case, the recommendation information providing unit 170 may receive the channel information in real time and provide content recommendation information through the received channel information when the channel currently being viewed by each user group or the channel is moved to watch. That is, the content recommendation information of the channel group to which the user group is currently watching or the channel selected for viewing belongs may be superimposed within a predetermined area of the content currently being viewed, or the recommendation information may be displayed when the user group requests. .

Also, the recommendation information providing unit 170 may analyze content information currently being viewed by the user group and provide recommendation information of content similar to the content being viewed. For example, as described above, each content may belong to different channel groups according to the degree of belonging. That is, since each channel group can be related to each other, if the content currently being viewed on a specific channel has a higher degree of belonging to another channel group, content recommendation information of the channel group can be provided. That is, when a specific user group is watching a movie on an educational channel, the movie may have a higher degree of belonging to the channel group of the movie specialty channel, and thus can provide content recommendation information of the channel group to which the movie specialty channel belongs. .

At this time, the recommendation information providing unit 170, as described above, as shown in FIG. 8, when a channel group of several levels (L1, L2) is generated due to an association between channels, first, the channel of the first level (L1) Content recommendation information for a group may be provided with a high priority, and content recommendation information for a channel group of a second level (L2) may be provided with a lower priority.

13 is a flowchart of a content recommendation method according to an embodiment.

13 may be an embodiment of a content recommendation method performed by the content recommendation apparatus 100 according to the embodiment of FIG. 2.

Referring to FIG. 13, the content recommendation device 310 may collect broadcast information such as an electronic program guide, web information, channel meta information, and content meta information (step 310). For example, the collected broadcast information is broadcast from each channel through channel information such as broadcast content name, broadcast content description, genre, channel name, channel description, and broadcast information published on the web from the electronic program guide and meta information of each channel. Through the meta-information of the content, the content name, actor, director, synopsis, and producer can be included.

Then, attribute information for each channel may be generated based on the collected broadcast information (step 320). At this time, the attribute may be classified into "specialized channel for movies", "specialized channel for dramas", "children's channels", "education channels", "economic broadcasting channels", "news channels", and the like.

Then, when attribute information for each channel is generated, similarity between each channel may be calculated using the attribute information (step 330). In this case, the similarity may be calculated by additionally using the duplicated information of the content broadcasted in each channel, for example, whether it is duplicated or the number of duplicated contents, in addition to the attribute information of each channel. Therefore, even if the channel properties are different, the similarity calculated according to the overlapping degree of broadcast content may increase relatively, and even if the channel properties are the same, the overlapping degree of content is low between channels having a higher content duplication level. Similarity can be calculated higher than between channels.

Then, when the similarity between each channel is calculated, one or more channel groups may be generated using the calculated similarity between the channels (step 340). At this time, by comparing the preset similarity threshold with the calculated similarity, channels having similarities exceeding the similarity threshold are generated as one channel group, and the generated channel groups may be correlated with each other. On the other hand, channels having similarities that do not exceed the similarity threshold may be generated as one channel group, and this channel group may not be associated with other channel groups.

In addition, the similarity threshold may be preset to a range having an upper limit and a lower limit, and channels having similarities exceeding the upper limit are created as one channel group, and in the case of channels having similarities between the upper and lower limits, channels associated with each other It is possible to belong to a group, and also, create an upper channel group for a channel group associated with each other.

In this case, additionally, the degree of belonging of a channel of each content may be calculated, and channels belonging to different channel groups may have an association relationship according to the calculated degree of belonging of the channel. In this case, it is also possible to create upper channel groups for different channel groups associated with each other. For example, the content belonging degree may belong to each channel when one content is broadcast on different channels. At this time, if the type of content is a movie and one of the channels to be broadcast is a "film specialty channel", and the other channel is a "drama specialty channel," the content is calculated with a high degree of affiliation with the "movie specialty channel." The degree of affiliation with " channel " can be calculated relatively low.

Subsequently, content recommendation information for each user group may be generated for each created channel group (step 350). At this time, the viewing history information may be collected from each user group, and the collected viewing history information may be replaced with viewing history information for each channel group. Content preferences for each channel group may be calculated for each user group by using the replaced viewing history information for each channel group, and content recommendation information may be generated using the calculated preferences.

For example, the content preference is based on the type of content for each channel group, for example, if the content is a movie, the content of a movie of a type similar to the type of the content mainly viewed in the channel group is higher based on the movie type such as action, melo, SF, etc. Can be calculated with preference. As described above, content preferences may be calculated according to various other criteria.

Then, the generated content recommendation information may be provided to a corresponding user group (step 360). At this time, when a channel currently being viewed by each user group or a channel is moved to watch, the channel information may be received in real time and content recommendation information may be provided through the received channel information.

In addition, the user group may analyze content information currently being viewed and provide recommendation information of a channel group broadcasting content similar to the content being viewed. In addition, when channel groups of several levels (L1, L2) are generated due to association between channels, first, content recommendation information for the channel group of the first level (L1) is provided with high priority, and then the second level ( Content recommendation information for the channel group of L2) can be provided with a lower priority.

Embodiments of the present invention are implemented in the form of program instructions that can be executed through various computer means and can be recorded on computer readable media. Such computer-readable media may include program instructions, data files, data structures, or the like alone or in combination. The program instructions recorded on the computer-readable medium may be specially designed and configured for the present invention or may be known and usable by those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs, DVDs, and magnetic media such as floptical disks. Includes all types of hardware devices specifically configured to store and execute program instructions such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of program instructions may include high-level language code that can be executed by a computer using an interpreter, etc., as well as machine language codes made by a compiler. Such a hardware device may be configured to operate as one or more software modules to perform the operation of the present invention, and vice versa.

As described above, the present invention is not limited to the configuration and method of the embodiments described above, but may be configured by selectively combining all or part of each embodiment so that various modifications can be made. .

According to the present invention, an apparatus for recommending appropriate content in a device such as a TV used by a plurality of users is disclosed. According to embodiments of the present invention, it is necessary to collect separate profile information and the like in order to recommend customized content information to multiple users by recommending content for each channel group provided on the TV, not recommending content for each user. There is not that utilization will not be small.

1: Content recommendation system 100: Content recommendation device
110: broadcast information collection unit 120: channel attribute generation unit
130: channel group generation unit 140: viewing history collection unit
150: viewing history replacement unit 160: recommendation information generation unit
170: recommendation information providing unit 180: content belonging calculation unit
190: channel group update unit

Claims (20)

A channel group generating unit that calculates similarity between channels based on attribute information of the channel and generates one or more channel groups using the calculated similarity between channels;
A viewing history replacement unit that replaces the content viewing history information for each user group with the viewing history information for each channel group; And
And a recommendation information generation unit that generates content recommendation information for each user group for each channel group based on the viewing history information for each channel group.
The recommendation information generating unit,
A content recommendation device that calculates content preferences for each channel group based on viewing history information for each channel group, and generates content recommendation information for each channel group based on the calculated content preferences. According to claim 1,
Collects broadcast information including one or more of electronic program guide information for each channel, meta information for each channel, meta information of content broadcasted on each channel, and web information A content recommendation device further comprising a broadcast information collection unit. According to claim 2,
And a channel attribute generator configured to generate attribute information of each channel using the collected broadcast information. According to claim 3,
The channel group generator
A content recommendation device that checks content duplication information between the channels by using the collected broadcast information, and calculates similarity between the channels based on the content duplication information. The method of claim 4,
The content overlap information
A content recommendation apparatus including one or more of overlapping content and the number of overlapping content between the channels. According to claim 1,
Further comprising a content affiliation calculation unit for calculating the affiliation for each channel of the content broadcast on each channel,
The channel group generator
A content recommendation device that updates a relationship between the created channel groups based on the belonging degree of the calculated content. The method of claim 6,
The content affiliation calculation unit
A content recommendation device that calculates a degree of belonging for each channel of the content based on the association between the attribute of each channel and the type of content broadcast on each channel. According to claim 1,
The channel group generator
A content recommendation apparatus that compares the similarity between the calculated channels and a preset threshold, and generates channels as one or more channel groups based on the result. The method of claim 8,
The channel group generator
As a result of the creation of the channel group, if one channel belongs to more than one channel group, a content recommendation device that creates an upper channel group for the two or more channel groups. delete According to claim 1,
A content recommendation device further comprising a recommendation information providing unit that provides content recommendation information for each channel group generated for each user group to each user group. The method of claim 11,
The recommended information providing unit
A content recommendation device that provides content recommendation information of a channel group to which each channel is currently selected by each user group, to each user group. The method of claim 11,
The recommended information providing unit
A content recommendation device that provides content recommendation information of a channel group that broadcasts content similar to the content currently being viewed by each user group to each user group based on the degree of belonging to each channel of the content. Calculating similarity between channels based on attribute information of the channel;
Generating one or more channel groups using the calculated similarity between channels;
Replacing content viewing history information for each user group with viewing history information for each channel group; And
And generating content recommendation information for each user group for each channel group based on the viewing history information for each channel group.
Generating the content recommendation information,
A content recommendation method for calculating content preferences for each channel group based on viewing history information for each channel group, and generating content recommendation information for each channel group based on the calculated content preferences. The method of claim 14,
Collects broadcast information including one or more of electronic program guide information for each channel, meta information for each channel, meta information of content broadcasted on each channel, and web information To do; And
And generating attribute information of each channel using the collected broadcast information. The method of claim 15,
The step of creating the channel group
And checking content duplication information between the channels using the collected broadcast information.
A content recommendation method for calculating similarity between the channels based on the identified content duplication information. The method of claim 14,
Calculating a degree of belonging for each channel of content broadcasted in each channel; And
And updating the relationship between the created channel groups based on the belonging degree of the calculated content. The method of claim 14,
The step of creating the channel group
Comparing the similarity between the calculated channels and a preset threshold;
Generating channels into one or more channel groups based on the comparison result; And
And generating an upper channel group for the two or more channel groups when one of the channels belongs to two or more channel groups. The method of claim 14,
And providing content recommendation information for each user group generated for each user group to each user group. A computer-readable recording medium on which a program for executing any one of the content recommendation methods of claims 14 to 19 is recorded.

Images