KR101179699B1 - Apparatus and method for contents recommendation using continuous watching feature - Google Patents

Abstract

PURPOSE: A content recommending device using consecutive viewing features and a method thereof are provided to supply a recommendation list using probability of consecutively watching content and probability of watching sub content in case of series content. CONSTITUTION: A first probability calculating unit(110) calculates first probability for each identification information of second content. The first probability is probability of receiving the second content by consecutively receiving first content. A second probability calculating unit(120) calculates second probability in case of series content. The second probability is probability of consecutively receiving second sub content after receiving first sub content. A recommendation unit(140) transmits a recommendation list to a recommendation target device.

Description

Apparatus and method for contents recommendation using continuous watching feature}

The present invention relates to a content recommendation apparatus and method using a continuous viewing feature, and more particularly, a device for recommending content to a recommendation target device among a plurality of content viewing devices based on data collected from the plurality of content viewing devices; It is about a method.

With the advent of the convergence of broadcasting and telecommunications, the number of IPTV subscribers is steadily increasing, while the amount of content provided through IPTV is also increasing. As a result, IPTV users can watch a wider variety of content at any time regardless of time, but have a choice problem that requires a lot of time and effort to find the desired content.

The recommendation technology that selects and provides contents that may be of interest to the user has been actively studied since the late 1990s, and includes representative rule-based recommendation techniques, collaborative recommendation techniques, and content-based recommendation techniques.

The association rule recommendation technique is a method of recommending content highly related to the content used by the user by analyzing the association between the contents used by the users. Association rules generally appear in the form of A-> B, which means that users who use content A are more likely to use content B.

The collaborative recommendation technique first selects users who have a similar preference for content. After that, the predicted preference is calculated by considering similar users' preferences for the content that is not used by the user, and the contents having a large predicted value are recommended.

Content-based recommendation is a technique that recommends items with high similarity by comparing a user model expressing user preference information with the property of an item. The structure of the user model must match the property structure of the content.

However, these conventional recommendation techniques do not effectively reflect the viewing characteristics of the IPTV service.

Most of the content provided by IPTV consists of a series, and users watch the series continuously or periodically watch TV.

Considering this IPTV viewing characteristic, in order to improve the accuracy of recommendation, each sub content (episode) constituting the series should be divided into individual contents. In this case, the number of contents is rapidly increased, and excessive learning time and memory are used. In addition, the association rule recommendation increases the size of the frequent item set, which is the basis of association rule generation, and in the case of collaborative recommendation, the size of the preference matrix for calculating similarity between users increases. On the other hand, if all the sub-contents of the series are divided into one content, it is impossible to distinguish the individual sub-contents, which reduces the accuracy of recommendation and does not effectively reflect the viewing characteristics of the IPTV service called series driving.

The technical problem to be solved by the present invention is to receive viewing data from a plurality of content viewing devices, calculate the continuous viewing probability of the content and the continuous viewing probability of the successive sub-contents, and compare the sub-contents of the recommended target device among the plurality of content viewing devices. A content recommendation apparatus and method using a continuous viewing feature of recommending content to a recommendation target device by calculating a viewing period and a viewing interval are provided.

In order to achieve the above technical problem, the content recommendation apparatus using the continuous viewing feature according to the present invention transmits a first content of the plurality of contents in a plurality of content viewing apparatuses that receive and reproduce a plurality of contents through a communication network. A first probability calculator for calculating a first probability, which is a probability of receiving second content continuously after receiving, for each identification information of the second content; When the first content corresponds to a series of content consisting of a plurality of sub-contents that are temporally contiguous, the first sub-content is continuously received after the first sub-content is received from the plurality of sub-contents. A second probability calculator configured to calculate a second probability that is a probability of receiving a second sub content consecutive to the second probability; A viewing period calculating unit configured to calculate a viewing period, which is a time from when the first sub content is received by the recommendation target device among the plurality of content viewing devices, to the second sub content in succession; And identification information of contents selected from the plurality of contents or sub-content based on the first probability, the second probability, and the viewing period calculated based on the identification information of the content being played in the recommendation target device. And a recommendation unit for generating a transmission to the recommendation target device.

In order to achieve the above technical problem, the content recommendation method using the continuous viewing feature according to the present invention includes (a) a first of the plurality of contents in a plurality of content viewing apparatuses that receive and reproduce a plurality of contents through a communication network; Calculating a first probability for each identification information of the second content, which is a probability of receiving second content continuously after receiving content; (b) when the first content corresponds to a series of content consisting of a plurality of sub-contents that are temporally contiguous, the plurality of content viewing apparatuses receive the first sub-content among the plurality of sub-contents and then continuously perform the first content. Calculating a second probability, which is a probability of receiving second sub content consecutive to one sub content; (c) calculating a viewing period, which is a time from when the first sub-content is received by the recommendation target apparatus among the plurality of content viewing apparatuses, to the second sub-content received continuously; And (d) identification information of a content or sub content selected from the plurality of contents on the basis of the first probability, the second probability, and the viewing period calculated in response to the identification information of the content being played in the recommendation target device. Generating a recommendation list and transmitting the generated recommendation list to the recommendation target device.

According to the content recommending apparatus and method using the continuous viewing feature according to the present invention, it is possible to provide a recommendation list using the continuous viewing probability for the content and the viewing probability of the sub content in the case of the series content, and through this, the viewing feature of the IPTV service. Trends in driving in series can be reflected in recommendations. In addition, by using a viewing interval reflecting a time difference between a viewing period, which is a viewing time interval for viewing consecutive sub-contents, and a viewing sub-content, it is possible to provide a recommendation for a series of contents updated at regular intervals.

1 is a block diagram showing the configuration of a preferred embodiment of a content recommendation apparatus using a continuous viewing feature according to the present invention;
2 is a flowchart illustrating an embodiment in which a recommendation unit of the present invention generates a recommendation list, and
3 is a flowchart illustrating a preferred embodiment of a content recommendation method using a continuous viewing feature according to the present invention.

Hereinafter, exemplary embodiments of a content recommendation apparatus and method using a continuous viewing feature according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram showing the configuration of a preferred embodiment of a content recommendation apparatus using a continuous viewing feature according to the present invention.

1, the content recommendation apparatus according to the present invention includes a first probability calculator 110, a second probability calculator 120, a viewing period calculator 130, and a recommender 140. The apparatus may further include a viewing interval calculator 150 for recommending sub content having a low probability of continuous viewing, and a storage unit 160 storing other user information and content information.

The first probability calculator 110 is a first probability that is a probability of receiving second content continuously after receiving first content among a plurality of contents in a plurality of content viewing apparatuses that receive and reproduce a plurality of contents through a communication network. Is calculated for each piece of identification information of the second content.

The first probability is calculated as in Equation 1 below.

[Equation 1]

는 콘텐츠 c_i를 시청한 후 c_j를 시청할 확률을 의미한다.

는 콘텐츠 c_i를 시청한 후, 콘텐츠 c_j를 시청한 콘텐츠 시청 장치의 수이며,

는 콘텐츠 c_i를 시청한 콘텐츠 시청 장치의 수이다.

Denotes the probability of watching c _j after watching content c _i .

Is the number of content viewing devices that watched content c _j after watching content c _i ,

Is the number of content viewing devices that viewed content c _i .

Among a plurality of content viewing devices that receive and reproduce a plurality of content through a communication network, the number of the plurality of content viewing devices that receive the content c _i and the plurality of content viewing devices that receive the content c _i consecutively and receive c _j The number of devices is stored in storage 160. First probability calculation unit 110 after viewing the content c _i by the expression (1) above using this value can be calculated in the first probability is the probability to watch c _j for each content.

The second probability calculator 120 receives the first sub content of the series content consisting of the plurality of sub content consecutively from the plurality of content viewing apparatuses, and then receives the second sub content consecutively to the first sub content. A second probability that is a probability is calculated. In the series content, the content of the sub content may be continued in succession, and the content may be progressed in separate materials and contents. In the former case, drama is a typical case, and in the latter case, an entertainment program is a typical case. According to the continuity of the content, there is a possibility that the viewing characteristics of the series content are different, and if there is no continuity of the content, the probability of watching the successive sub-contents is low.

If there is sufficient metadata about the content, the continuity of the content may be analyzed and used for recommendation, but currently provided metadata is insufficient to analyze the continuity of the content. Therefore, when watching the sub content in the series content, the second probability, which is the probability of continuously watching the sub content of the same series content, is calculated.

The viewing period calculating unit 130 receives the first sub content of the series content consisting of a plurality of sub content consecutively from the recommendation target device among the plurality of content recommendation apparatuses, and then continues the second sub content consecutively to the first sub content. Calculate the viewing period which is the time until receiving the transmission. In order to calculate this, the storage 160 stores the viewing history of the sub content and the final transmission time of the sub contents transmitted to the recommendation target device for each of the contents.

The driving characteristics of the series contents come in two forms. In the case of infant contents, the broadcasting time of each sub-content is short, and therefore, the entire series tends to be viewed continuously, whereas in the case of a long broadcasting time, such as drama, it tends to continuously watch one or two episodes a day. In this case, if only the probability model based on the currently viewed content is applied, the continuous sub-content of the series watched yesterday cannot be recommended today. Especially in the case of TV replay, incomplete dramas and entertainment programs are updated weekly, so it is more likely to be watched at regular intervals rather than in succession. In order to solve this problem, the viewing period calculating unit 130 analyzes and stores, for each content viewing device, an average period of viewing the continuous sub content for each series content.

Meanwhile, the content recommendation apparatus according to the present invention may further include a viewing interval calculator 150.

The viewing interval calculating unit 150 calculates a viewing interval that is a time interval between the contents c _i and the contents c _j when the sub contents c _j of the same series content are continuously watched after viewing the sub contents c _i of the series content. to be. In this case, the viewing interval may be represented as the interval between the sub-contents. This is used for recommendation of a series having a low tendency to watch consecutive sub contents, that is, a series having a tendency to select an arbitrary sub content. In addition, it is recommended. In this case, when the average sub content viewing interval is similar to 1, it means that the tendency of immediately watching successive sub contents is strong, and when it is shown as a negative value, it means that the tendency to find and watch the previous sub contents is strong.

The storage unit 160 calculates data from the first probability calculator 110, the second probability calculator 120, the viewing period calculator 130, and the viewing interval calculator 150, and uses the same to calculate the data. The necessary data are stored to enable the recommendation at 140. In addition, content property information such as the content name and whether or not the series is stored is also stored.

The recommender 140 is based on data stored in the first probability calculator 110, the second probability calculator 120, the viewing period calculator 130, the viewing interval calculator 150, and the storage 160. The recommendation list including the identification information of the content or the sub content selected by the recommendation method using the continuous viewing feature is transmitted to the recommendation target device.

2 is a flowchart illustrating an embodiment in which the recommendation unit 140 of the present invention generates a recommendation list.

The recommender 140 first acquires the sub content and the viewing time that the recommendation target device was last transmitted for each series. Regarding the series content whose viewing period has elapsed among the series contents, identification information of the sub content subsequent to the last viewed sub content is added to the recommendation list (S1100). The sub content to be recommended in consideration of the viewing period of the recommendation target device may be limited to the case where the second probability is greater than or equal to the threshold. The second probability threshold may be set by the content provider to an arbitrary value based on information such as the viewer's viewing history or other preferences.

Next, the recommendation unit 140 selects content whose first probability is greater than or equal to a threshold (S1110). The threshold of the first probability may be set by the content provider to an arbitrary value based on information such as a viewer's viewing history or other preferences.

The recommender 140 determines whether the selected content is the same as the content currently being viewed by the recommendation target device because the first probability is greater than or equal to the threshold (S1120). If the selected content is the same as the content currently being watched, the sub content subsequent to the last watched sub content is first added to the recommendation list (S1130). Thereafter, the sub content is added in consideration of the sub content viewing interval corresponding to the content currently being viewed (S1140). For example, if there are 10 pieces of sub content currently being viewed on the recommendation target device and the viewing interval of the sub content is 3, 13 pieces are added to the recommendation list.

On the other hand, if the selected content is not equal to the content currently being viewed because the first probability is greater than or equal to the threshold, the recommender 140 adds the last updated sub-content recently updated to the recommendation list when the selected content is a series content ( S1150). If the selected content is fragment content, the fragment content is added to the recommendation list.

3 is a flowchart illustrating a preferred embodiment of a content recommendation method using a continuous viewing feature according to the present invention.

The first probability calculator 110 is a first probability that is a probability of receiving second content continuously after receiving first content among a plurality of contents in a plurality of content viewing apparatuses that receive and reproduce a plurality of contents through a communication network. Is calculated for each identification information of the second content (S2100).

When the content corresponds to a series of content consisting of a plurality of sub-contents that are temporally continuous, the second probability calculating unit 120 receives the first sub-content among the plurality of sub-contents from the plurality of content-viewing devices and continuously generates the content. A second probability, which is a probability of receiving second sub content consecutive to one sub content, is calculated (S2200).

The viewing period calculating unit 130 receives the first sub content, which is a sub content of the series content, from the recommendation apparatus when the content corresponds to a series content consisting of a plurality of sub contents that are continuous in time, and then continuously performs the first sub content. A viewing period, which is a time until receiving the second sub content subsequent to the content, is calculated (S2300).

The viewing interval calculating unit 150 receives the first sub content, which is a sub content of the series content, from the recommended target device when the content corresponds to the series content consisting of a plurality of sub contents that are continuously temporally contiguous. When the sub content is received, a viewing interval that is a time interval between the first sub content and the received sub content is calculated (S2400).

The recommender 140 may include a recommendation list including identification information of a selected content or sub content among a plurality of contents based on a first probability, a second probability, a viewing period, and a viewing interval calculated according to the identification information of the content being played. It generates and transmits to the recommended target device (S2500).

The present invention can also be embodied as computer-readable codes on a computer-readable recording medium. A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored. Examples of the computer-readable recording medium include a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like, and may be implemented in the form of a carrier wave (for example, transmission via the Internet) . The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is clearly understood that the same is by way of illustration and example only and is not to be taken by way of limitation in the embodiment in which said invention is directed. It will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the scope of the appended claims.

110-first probability calculator
120-second probability calculator
130-watch cycle calculator
140-Referral
150-Viewing interval calculator
160-storage

Claims (8)

A plurality of content viewing apparatuses that receive and reproduce a plurality of contents through a communication network, identify a first probability, which is a probability of receiving second contents continuously after receiving first contents among the plurality of contents, and identifying the second contents. A first probability calculating unit for calculating each time;
When the first content corresponds to a series of content consisting of a plurality of sub-contents that are temporally contiguous, the first sub-content is continuously received after the first sub-content is received from the plurality of sub-contents. A second probability calculator configured to calculate a second probability that is a probability of receiving a second sub content consecutive to the second probability;
A viewing period calculating unit configured to calculate a viewing period, which is a time from when the first sub content is received by the recommendation target device among the plurality of content viewing devices, to the second sub content in succession; And
A recommendation list including identification information of a content selected from the plurality of contents or sub-content based on the first probability, the second probability, and the viewing period calculated based on the identification information of the content being played in the recommendation target device; A recommendation unit generating and transmitting to the recommendation target device; And
And a viewing interval calculator configured to calculate a viewing interval, which is a time interval between the sub content of the first content and the first sub content continuously received after receiving the first sub content from the recommendation target device.
The recommendation unit adds identification information of the selected sub content to the recommendation list based on a viewing interval corresponding to the identification information of the content being played when the content being played in the recommendation target device corresponds to the series content. Content recommendation device. delete The method of claim 1,
The apparatus may further include a storage unit configured to store a final transmission time point of the sub contents transmitted to the recommendation target device in response to identification information of each of the contents corresponding to the series contents among the plurality of contents.
The recommendation unit may include identification information of the sub content subsequent to the sub content finally transmitted to the recommendation target device among the sub contents included in the content whose time corresponding to the viewing period has elapsed from the last transmission time, to the recommendation list. Content recommendation device, characterized in that the addition. The method of claim 1,
The recommendation unit adds, to the recommendation list, the identification information of the content having a first probability corresponding to the identification information of the content being played in the recommendation device among the plurality of contents, which is equal to or greater than a preset threshold.
If the content being played in the recommendation target device corresponds to the series content, and the identification information of the content whose first probability is equal to or greater than the threshold is the same as the identification information of the content being played, a plurality of sub-contents constituting the content being played; Add identification information of the sub content continuous to the sub content finally transmitted to the recommendation target device to the recommendation list;
If the content being played in the recommendation target device corresponds to the series content and the second probability corresponding to the content being played is equal to or greater than a preset threshold, the content of the sub content subsequent to the sub content finally transmitted to the recommendation target device is set. And adding identification information to the recommendation list. delete delete (a) The second content is a first probability, which is a probability of receiving second content in succession after receiving first content among the plurality of content in a plurality of content viewing apparatuses receiving and playing back a plurality of contents through a communication network. Calculating for each identification information of the;
(b) when the first content corresponds to a series of content consisting of a plurality of sub-contents that are temporally contiguous, the plurality of content viewing apparatuses receive the first sub-content among the plurality of sub-contents and then continuously perform the first content. Calculating a second probability, which is a probability of receiving second sub content consecutive to one sub content;
(c) calculating a viewing period, which is a time from when the first sub-content is received by the recommendation target apparatus among the plurality of content viewing apparatuses, to the second sub-content received continuously;
calculating a viewing interval which is a time interval between the sub content of the first content and the first sub content continuously received after receiving the first sub content in the recommendation target device; And
(e) identification information of a content or sub-content selected from the plurality of contents based on the first probability, the second probability, and the viewing period calculated in response to the identification information of the content being played in the recommendation target device; Generating a recommendation list and transmitting the recommendation list to the recommendation target device;
In the step (e), if the content being played in the recommendation target device corresponds to the series content, the identification information of the selected sub content is added to the recommendation list based on the viewing interval corresponding to the identification information of the content being played. Content recommendation method characterized in that. delete

Images