KR20080021069A - Method and apparatus for estimating total interest of a group of users directing to a content - Google Patents

Abstract

The invention which sees about method and the system which presume the whole interest of the user group is the thing about the contents. From here the user group includes the member of at least 2 people, minute description method, about the features which in time available block and the contents the contents are included the phase which acquires information; Hour when differs priority price of the respectively member from within dividing and about the features which are included the phase which acquires the preference of the member; Time available priority price of the respectively member from within dividing and according to preference of the member includes the phase which decides a whole interest about the features which are included. ® KIPO & WIPO 2008

Description

METHOD AND APPARATUS FOR ESTIMATING TOTAL INTEREST OF A GROUP OF USERS DIRECTING TO A CONTENT}

TECHNICAL FIELD The present invention relates to information processing techniques, and in particular, to a method and apparatus for estimating the overall interest of a group of users for content.

Social development and technological advances are rapidly increasing the knowledge people demand. As large amounts of information become available, methods and devices that can automatically process information early on free people from their daily drudgery, enable them to work more efficiently and make their lives more efficient. It became very necessary to be entertained.

Content recommendation is a very buzzing technique in information processing, and its underlying principle is to analyze content based on pre-set decision criteria to find interest of a particular user / user group, and then present the content to the user / user group. Is to decide whether to recommend. This technology has wide application potential in television program recommendation. To this end, several studies are now being conducted on this background.

Those who are recommended for the content may be individual users or user groups consisting of at least two users or members. For example, a user group may consist of family members or roommates living in the same dormitory. Compared to individual users, user groups interact with members in the group; Accordingly, the analysis and evaluation of the interests of the user group is more complicated than that of the individual user.

In US patent application Ser. No. 09 / 819,440, the invention entitled “Method and Apparatus for Generating Recommendations for Multiple Users”, filed March 28, 2001, discloses a method for determining a recommendation result to a user. do. The inventors of the patent application are Lalitha Agnihotri and Srinivas Gutta, and the applicant is Philips Electronics North America Corporation. In the embodiment disclosed in the patent application, the recommendation result is expressed as a recommendation point or a combined recommendation point derived from the calculation of each user of the user group for the features included in the information recommendation. The disclosure of US patent application Ser. No. 09 / 819,440 is incorporated herein by reference.

In a particular embodiment of US patent application Ser. No. 09 / 819,440, the method of determining the recommendation point is to first obtain an electronic program guide (EPG), ie television program forecast information present in electronic form. Thereafter, a profile 300 of the user group member is obtained that includes respective viewer preferences for each feature. Then, if necessary, to convert each viewer's preference for each feature to the same numerical rating. The next processing step is to determine whether the program is worth recommending or the point at which it is recommended for each viewer according to the viewer profile described above. After the recommendation points of all the viewers of the group are obtained, the recommendation points of all the viewers are combined to take the arithmetic mean or weighted average value from the recommendation points of all the viewers to generate the overall recommendation points of the user group. Finally, output the calculated recommended points of the user group.

More factors must be considered in order to better determine the overall user group's interest in the content, that is, the overall interest.

One object of the present invention is to correctly determine the overall interest of a user group for content, and to recommend content to the user group more individually.

For one aspect, the present invention provides a method of estimating the overall interest of a user group for content, wherein the user group comprises at least two members, the method comprising the following steps: Obtaining information about possible time slices and features included in the content; Obtaining a member's preference for the included feature and the priority value of each member within a different time segment; And determining total interest in accordance with the member's preference for the included feature and the priority value of each member within a possible time segment.

In one embodiment of the invention, the priority value of each member is from a plurality of time segments in which the priority value of one member differs on at least two time segments, and the two time segments are divided by day. It is configured. Or the priority value of each member is configured such that the priority value of one member is different on the same time segment on different days of at least two days.

In another embodiment, the determining step includes the following steps: a user for the included feature in the different time segment according to the preference value of each member in the different time segment and the preference for the included feature. Determining an interest value of the group; Determining an interest value of the group according to a possible time partition; And combining the interest values of the group to obtain the overall interest.

In another aspect, the invention provides a method of recommending content to a group of users, wherein the group of users comprises at least two members, the method comprising the following steps: Obtaining information about the included feature; Obtaining a member's preference for the included feature and the priority value of each member within a different time segment; Determining total interest in accordance with the member's preference for the included feature and the priority value of each member within a possible time segment; And recommending to the user group according to the overall interest.

Another aspect of the invention is to provide an apparatus for estimating the overall interest of a user group for content, wherein the user group comprises at least two members, the device comprising: a possible time segment of the content and Obtaining means for obtaining information about a feature included in the content; Receiving means for receiving a priority value of each member within a different time segment and a member's preference for included features; Means for determining overall interest in accordance with the member's preference for the included feature and the priority value of each member within a possible time segment.

In one embodiment, the priority value of each member is configured such that the priority value of one member differs from at least two time segments, and the plurality of time segments are divided into two time segments by day. Or the priority value of each member is configured differently on the same time segment on different days of at least two days of the priority value of one member.

In another embodiment, the means for determining comprises: interest of the user group to the included features within the different time partitions according to the preference values of each member within the different time partitions and preferences for the included features. Generating means for determining a value; Retrieving means for determining an interest value of the group according to a possible time partition; And calculating means for combining the interest values of the group to obtain the overall interest.

Another aspect of the invention provides an apparatus for recommending content to a group of users, wherein the user group comprises at least two members, the apparatus comprising: possible time segments of the content and features included in the content. Acquiring means for obtaining information concerning; Receiving means for receiving a priority value of each member within a different time segment and a member's preference for included features; Determining means for determining total interest in accordance with the member's preference for the included feature and the priority value of each member within a possible time segment; And recommendation means for recommending to the user group according to the overall interest.

Another aspect of the invention is to provide a computer program product for estimating the overall interest of a user group for content, wherein the user group comprises at least two members, the computer program product comprising: possible content Code for obtaining information about a feature included in the time segment and the content; Code for obtaining a priority value of each member within a different time segment and a member's preference for included features; And code for determining overall interest in accordance with the member's preference for the included feature and the priority value of each member within a possible time segment.

The invention also relates to a memory carrier comprising a computer program product.

In the present invention, the priority of each group member is sometimes variable, which is to accommodate various states in which the influence of each group member on the content selection is dynamic, thereby allowing the user to recommend content more appropriately to the user group. do.

In one embodiment, the priority of a member may be set to change from one time segment to another. This implementation has the advantages of simplicity and flexibility and is suitable for television program interest estimation and recommendation. In addition, the user group's interest in each feature within the hourly partition may be pre-calculated based on it, which will be used for the next calculation of the overall interest, which reduces the computational cost and thus simplifies the hardware structure.

These and other objects and advantages of the invention will be apparent, and the invention will be more fully understood from the following description and claims taken in conjunction with the accompanying drawings.

1 is a flow diagram illustrating a preferred embodiment according to the method of the present invention for estimating the overall interest of a group of users for content.

2 is a flow diagram illustrating an example embodiment of calculating a user group's interest in a feature based on a fuzzy logic processing mode.

3A, 3B, and 3C illustrate a membership function used in the fuzzy logic processing mode shown in FIG. 2, corresponding to the user group's preferences, priorities, and interests for a feature, respectively.

4 is a flow diagram illustrating another preferred embodiment according to the method of the present invention for estimating the overall interest of a group of users for content.

5 is a flow diagram illustrating a preferred embodiment of the method of the present invention for recommending content to a group of users.

Figure 6 is a block diagram illustrating an embodiment in accordance with the inventive arrangements for estimating the overall interest of a user group for content.

7A is a block diagram illustrating an embodiment of the device of the present invention for recommending content to a group of users.

7B is a block diagram illustrating another embodiment according to the device of the present invention for recommending content to a group of users.

Like reference numerals in all drawings indicate similar or identical features or functions.

The following is a detailed description of a preferred embodiment of the present invention based on the drawings.

1 is a flow diagram illustrating a preferred embodiment according to the method of the present invention for estimating the overall interest of a user group for content. For example, the content is a television program.

As shown in Fig. 1, in step S110, user data including the preferences of all group members and the priority of members for each feature in selecting a television program is obtained. User data and other user-related data may be stored in a user profile for future use. The user profile is preferably a data file of a given format. They can be corrected by an upgrade by the program provider on a regular basis, or by their correction at the request of group members. Tables 1 and 2 below show data structures of two categories of user data, respectively.

A B C D E F G H I J K Child 1 1.0 0 0.6 0 0.4 0.8 0 0 -0.6 0 0.9 Child2 0.9 0 0 0.6 0 -0.4 0.1 0.5 0 0.8 0 Mother 0.6 0 0 0 0.6 0.7 -0.5 0.2 0.9 0 0 father 0.7 0 0 0 0 0.3 0.8 0.9 0 -0.8 0.9

Time block Child 1 Child2 Mother father 18: 30-19: 30 One One 0.1 0.6 19: 30-20: 30 One One 0.4 0.4 20: 30-21: 30 0.4 0.5 0.7 0.3 21: 30- 0.2 0.2 0.4 0.8

User data is obtained in various ways. Using a set-top box to perform a function of total interest calculation, the set-top box can obtain user data by downloading user data from a program provider's remote server via a cable television network or LAN. The user data provided may be statistical data obtained from a sample of multiple families; Family members personally set user data via an input device and store them in a device such as a set-top box for calculation of overall interest; The information of family members of television viewing is automatically gathered through monitor devices (e.g., video cameras and remote controls) installed around the television set, so that the information is sent to the set-top box or transmitted to the remote server via the network, It is formed through statistics and analysis from user data.

Table 1 contains data of the preferences of all group members for each feature. The user group consists of father, mother, child 1 and child 2. The television program watched includes a total of eleven features A-K, where A shows the program being broadcast in English (or mainly English); B, C and D show computer, math and chemistry related programs, respectively; E shows a program of romantic movies; F and G each show a cartoon or television series program; H shows a news program; I show a commercial program; K shows a sports program. The numerical values in the table show the members' preferences for a particular feature, i.e., in this embodiment, the numeric values are used to describe or indicate the preferences of all group members for the features included in the content. For example, child 1 likes to watch cartoon movies; His preference for an animated film is 0.8. Thus he is not interested in television series and news; The corresponding value is zero. Furthermore, child 1 hates seeing commercials; His preference for this feature is -0.6. Other family members may also be given corresponding values according to their respective preferences.

Table 2 contains priority data for all group members. As the description of the interest calculation mode shows, the priority here indicates the influence of that member on the overall interest of the user group. Higher priority means that member's relatively prominent influence on overall interest.

As Table 2 shows, in this example, the television program viewing time is divided into four time segments, and the priority of each member within the different time segments is not fixed. For example, if you take a time segment between 18:30 and 19:30, since the time is still there, parents generally allow their children to watch television, so that child1 and child2 have both priority values of 1. . As late night approaches, the likelihood that children will sit in front of a television set decreases, indicating that their priorities are gradually decreasing. Thus, the slight change in priority over time reflects the change in viewers of television programs. As another example, the daily mother's priority is lower than the other members at 18: 30-19: 30 (takes a value of 0.1), but is higher than the other members at the time segment of 20: 30-21: 30 (0.7 Takes the value of). The low priority in the early hours is because they are busy doing household chores. The late hour compartment is the mother's break. In order to satisfy her urgent demand for watching television, the priority can be set to a very high value, which ensures that she has a greater breath of television program selection.

In this embodiment, the priority is a different value within different time periods of the day. This mode is relatively appropriate for the estimation and recommendation of television program interest. In addition, different values are taken for changing the priority of the time segment of weekdays and holidays to reflect the effect of the holiday element on the viewing habits of family members. However, it is worth pointing out other modes over which time priority change can also be used. For example, it may be taken in the form of an arithmetic function that takes a different priority value for each year, month or day to reflect holiday and periodic factors or the priority changes over time.

In conclusion, when determining the total interest by appropriately setting a combination of preference and priority values within different times, the time varying element of the object of the received content is taken into account to facilitate understanding and give accurate recommendation results.

Referring back to FIG. 1, after performing step S110 of acquiring user data, the process enters step S120 where the user group's interest in each feature within each time segment (also a feature interest in the present invention). Is known as In the next section, the specific calculation mode of feature interest will be described in detail.

Entering step S130, a user profile is created to store the interest of the user group for each feature within each time segment calculated in step S120. Table 3 shows an exemplary structural form of the profile. In Table 3, A through K have their respective calculated values in four time partitions, which represent the user group's interest in one of the features in each time partition. Just for the purpose of representing the profile structure, some spaces do not represent specific values. Only the calculated values of A, E and G in the time segment 20:30 to 21:30 are shown.

Time block A B C D E F G H I J K 18: 30-19: 30 19: 30-20: 30 20: 30-21: 30 0.37 0.145 -0.015 21: 30-

Accordingly, a profile of the user group is created. It is worth noting that the overall user's interest in the program depends on the user's interest in each of the features included in the program; Therefore, in the next step of calculating the total interest, it is sufficient to take the corresponding value of the interest in the features included in the program from the profile of the user group, without having to repeat the steps S110 to S130.

The profile of a user group may not change for a period of time (eg, a week). During that time, each time a program recommendation is performed, the profile of the user group can be directly retrieved and used without having to repeat steps S110 to S130. For retrieval and ease of use, the profile of the user group may be recorded in a data file for later use in the step of calculating total interest. Once the overall interest is calculated in the set-top box, the user group's profile can be retrieved in a manner that is downloaded from a remote server over the network or obtained from memory contained in the set-top box. After updating the user data regularly or intermittently to reflect the change to the user data, it should be pointed out that the profile of the user group will be recalculated using steps S110 to S130.

Referring to Fig. 1, the process proceeds to step S140, where data relating to features and broadcast time included in the television program is obtained. The data is included in the EPG stored in the supplier database or sent to the user group in a given data structure, and used as a means of performing a function of obtaining a full interest. If the means for providing is physically independent of the means for acquiring the full interest, it is desirable to enter all these data into the means for acquiring the total interest for only one time. Assume that the television program for which full interest has been calculated is the original English television series that shows the emotional love story, and therefore the features it includes are A, E, and G. Further, suppose that the broadcast time of a program is in a time division of 20:40 to 21:20 of 20:30 to 21:30.

Entering step S150, the preference for each feature of the television series is retrieved from the profile of the user group created in step S130. Here, since the broadcast time belongs exactly to the time division of 20:30 to 21:30, the corresponding preferred data can be immediately seen from the table shown in FIG. Assume that the interest data for features A, E, and G included in the television program are 0.37, 0.145, and -0.015, respectively. The broadcast time of a program frequently occurs over two time divisions, that is, as shown in FIG. 2, the broadcast times of 20:25 to 21:15 are respectively 19:30 to 20:30 and 20:30 to It belongs to two time divisions of 21:30. This can be handled in the following way.

The first scheme proceeds in this way. The time division to which the program belongs is determined by the time at which the program starts broadcasting. Under such a hypothetical situation, the program is considered to be in the 19:30 to 20:30 time zone. The second approach is exactly the opposite, where the time segment is determined by the time the program ends. Thus, the program is considered to be in the 20:30 to 21:30 time division. Other schemes exist, where the interest in each feature is calculated based on the individual time segments according to the duration of the program within two different time intervals. That is, the weighted average value of interest for a feature is calculated as the final value of interest of the user group for the feature, as the broadcast duration as a weight.

The process then enters step S160 to combine the interest of the user group for each feature in the corresponding broadcast time segment obtained in step S150 with the overall interest in the program. For example, an arithmetic mean or weighted average of the user's group's interest for each feature included in the program calculated according to the following equation may be used as the overall interest (P):

는 시간 구획 i에서의 특징 j에 대한 사용자 그룹의 관심이고,

는 특징 j의 가중치이며, m은 텔레비전 프로그램의 특징의 개수이다. 단계(S140)에 관련하여 취해진 예에 대하여, 산술적 평균값이 취해질 때, 전체 관심은 특징 A, E 및 G에 대한 3개의 관심의 산술적 평균값과 동일하며, 즉,

이다.here

Is the user group's interest in feature j in time partition i,

Is the weight of feature j and m is the number of features of the television program. For the example taken in relation to step S140, when an arithmetic mean value is taken, the overall interest is equal to the arithmetic mean value of three interests for features A, E, and G, i.e.

to be.

As another example, a scoring method, which is often used for all kinds of competition or competition, with very few, i.e., ten features, may also be used, where the average value indicates that the feature has the maximum and minimum values of interest. Calculated without consideration. In any case, various combinations can be used as long as they reflect how much each feature contributes to the overall interest.

Thereafter, step S170 is made to determine whether there are any other television programs that require calculation of overall interest in them. If so, the process returns to step S140 or the entire process ends.

The following is a detailed description of the method of calculating the feature interest in step S120. Obviously, the user group's interest in the feature depends on the member's preference for the included feature and the priority value of each member within different time segments; Thus, there are several ways to calculate preferences or combine preferences with priorities to obtain such interest data. Here are two examples along the line.

The first example is based on the concept of weighted average. Specifically, the priority of each member within each time partition is treated non-uniformly. Thus, the priority is converted to the weighted value of the member within each time segment. Table 4 shows the uniformly processed weight values of each group member within different time segments. As shown in Table 4, the user group consists of three members: father, mother and child. In each time segment, the sum of the weighted values of each member is always equal to one. Table 5 shows the preference of each group member for features A, B and C. Each time in the table means the same as in Table 2 and is therefore not described herein any further.

Time block Child 1 Mother father 18: 30-19: 30 0.8 0 0.2 19: 30-20: 30 0.4 0.3 0.3 20: 30-22: 30 0.3 0.6 0.1 22: 30-24: 00 0 0.7 0.3 24: 00- 0 0 0

A B C child One 0.3 0.2 Mother 0.5 0.6 One father 0.3 0.9 0.7

Assuming that the television program includes feature A and its broadcast time is 18: 30-19: 30, the user group's interest in that feature is calculated according to the following equation:

는 시간 구획 i에서의 특징 j에 대한 사용자 그룹의 관심이고,

는 특징 j에 대한 사용자 k의 선호이며,

는 시간 구획 i 내에서의 사용자 k의 가중이며, n은 그룹 멤버들의 수이고,

와

의 곱은 시간 구획 i에서의 특징 j에 대한 사용자 k의 관심을 나타낸다.here

Is the user group's interest in feature j in time partition i,

Is the preference of user k for feature j,

Is the weight of user k in time partition i, n is the number of group members,

Wow

The product of denotes user k's interest in feature j in time partition i.

As far as considering this example, by querying Tables 4 and 5, the weights of father, mother and child are 0.2, 0 and 0.8 respectively within time segments 18: 30-19: 30 and their preference for feature A is respectively Since 0.3, 0.5 and 1, therefore, their respective interest in feature A is 0.06, 0 and 0.8, and the user group's interest value is 0.86.

Another example is that the function of step S120 is performed based on fuzzy logic processing.

FIG. 2 is a flow diagram illustrating an example embodiment of calculating a user group's interest in a feature based on a fuzzy logic processing mode, and the environments shown in Tables 1 and 2 are exemplary.

In order to calculate the interest of the user group in the time segment (e.g., 18: 30-19: 30) for the feature (e.g., A), as shown in FIG. The interest of the member for example (eg, father) and its priority within the corresponding time segment is entered. According to Tables 2 and 3, the values are 0.7 and 0.6, respectively.

Thereafter, step S220 is used to project two determinants of preference and priority entered in step S210 to the membership of the fuzzy value using the selected membership function. The type of membership function depends on the specific environment of the application. For example, as an example embodiment herein, the membership function shown in FIGS. 3A and 3B can be used, where FIG. 3A represents a membership function of preference, the X axis e ₁ represents a preference, and the vertical axis represents a membership μ. It is possible to use a membership function to project a preference with a membership of three fuzzy values of "dislike", "medium" and "like". 3B shows the priority membership function, with the X-axis e ₂ representing the priority and the vertical axis representing the membership μ. The membership function can be used to project priorities with membership of three fuzzy values of "not important", "medium" and "important".

Thereafter, step S230 is entered to deduct using a predetermined fuzzy logic rule to obtain a fuzzy output. This is a membership group of different fuzzy values that reflects the interest of the member (here father) on feature A. 3C shows the membership function of the fuzzy output. The X-axis α shows interest in feature attributes and the vertical axis shows membership μ. As shown in FIG. 2, a member's interest in feature attribute A is represented by the membership of several fuzzy values of "very dislike", "dislike", "medium", "like", and "very like". The specific rules of the deduction method depend on the characteristics of the application environment. For example, in the present example embodiment, the following rules may be used:

I. when e ₁ is "dislike" and e ₂ is "not important", α is "medium";

II. when e ₁ is "dislike" and e ₂ is "medium", α is "dislike";

III. when e ₁ is "dislike" and e ₂ is "important", α is "very dislike";

Ⅳ. if e ₁ is "medium" and e ₂ is "not important", α is "medium";

Ⅴ. when e ₁ is "medium" and e ₂ is "medium", α is "medium";

Ⅵ. if e ₁ is "medium" and e ₂ is "important", α is "medium";

Iii. if e ₁ is "likes" and e ₂ is "not important", α is "medium";

Iii. when e ₁ is "likes" and e ₂ is "medium", α is "likes";

Iii. when e ₁ is "favorite" and e ₂ is "important", α is "very much like";

로 변환한다(즉, 소위 역퍼지화 처리(defuzzification processing)). 여기서 i는 멤버의 수이다. 통상적으로 사용되는 역퍼지화 처리 방법은 "무게 중심", "최대 중심(center-of-maximum)" 및 "최대 평균법"을 포함한다. 처리에 대해서 "무게 중심"이 적용되면, 계산에 다음의 수학식이 사용될 수 있다.After that, the process proceeds to step S240 where the interest of the member for the feature attribute A obtained in step S230 is determined.

(Ie, so-called defuzzification processing). Where i is the number of members. Commonly used reverse purge treatment methods include "center of weight", "center-of-maximum" and "maximum averaging". If the "weight center" is applied to the process, the following equation can be used for the calculation.

Where μ [1] is the height of the output area that satisfies Deduction Rule No. 1, Y _i is the horizontal coordinate of the center of gravity of the output area that satisfies Deduction Rule No. 1, and m is the output area that satisfies this rule. Is the number of.

Thereafter, step S250 is reached in which it is determined whether all members S _i have been calculated. If this condition is not satisfied, the process returns to step S210 or enters to step S260.

In step S260, S _i of all group members (eg, father, mother, child, as in this example embodiment) is calculated to obtain the group's interest in a feature within a given time segment. Or, use the average value of S _i of all group members as the user group's interest in the feature within a given time segment.

4 illustrates another exemplary embodiment according to the method of the present invention for estimating the overall interest of a group of users for content.

As shown in FIG. 4, in step S410, user data may also be obtained in various ways. The structure of the data may be as shown in Tables 1 and 2 or Tables 4 and 5, but it may also be in other forms.

Thereafter, step S420 is entered to obtain data regarding broadcast time and characteristics of the television program. Assuming that the program information taken from the program source database is a television program containing features B and C, it is broadcasted at 20:30 to 21:00.

Thereafter, step S430 is entered in which a corresponding priority is determined according to the obtained broadcast time. For example, the priority variable of that time segment is obtained by checking the table.

Then, enter the step S440, the interest of the user group for each feature in the broadcast time segment according to the preferences of all members obtained in step S410 and the priority obtained in step S430. Calculate The manner of calculating the user group's interest in the feature has been described in detail above and will not be described herein any further.

Thereafter, step S450 is entered to convert the user group's interest in each feature in the broadcast time segment obtained in step S440 into total interest in the program. See the related description above for a specific calculation method.

Thereafter, the flow advances to step S460 to determine whether there are other programs required to calculate the overall interest. If so, return to step S420, or end the entire process.

If the broadcast time of the television program spans a plurality of time segments in Tables 2 and 4, the above-described scheme can be used. For example, assuming that the first and second schemes are used, find a time segment that is the time at which the program starts or ends in step S430, and accordingly each feature in the time segment in step S440. Calculate the user group's interest in. Assuming that the third scheme is used, find all time segments in which the program broadcast duration spans in step S430, and as the broadcast time in each time segment determined in step S430 as weighting in step S440. Compute the weighted average of interest for the features within the different time intervals and use them as the values of interest of the user group for the features.

As an alternative approach, it is worth noting that in step S440, it is possible to calculate the interest of each group member for a program rather than the interest of the user group for each program. Thus, in step S450, each group member's interest in the program is combined with the overall interest in the program. As a comparison between the two exemplary embodiments described by FIGS. 1 and 4, they first define a mode in which the user group's interest in each feature substantially changes from one time segment to another. Create a profile of the user group (steps S120 and S130), and use the profile of the user group to determine, for each particular program, the user group's interest in the features contained in the program within the broadcast time segment. (Step S150) is necessary, but the latter omits the step of creating a profile of the user group and instead the priority of the group members of the program in the broadcast time segment for each particular program is prioritized respectively. User group's interest in the feature that is calculated (step S430) and then included in the program within the broadcast time segment. It is calculated (step (S450)) is different from the point.

Using this method to gain full attention, it is possible to recommend content to a group of users more fully and appropriately. In the following, a television is taken as an example according to FIG. 5, and a preferred embodiment of a method for recommending content to a group of users is described.

As shown in FIG. 5, in step S510, the overall interest of the user group for one or more television programs is obtained using the method for acquiring the overall interest in FIGS. 1 and 4.

Thereafter, step S520 is entered, which determines the television program to be recommended to the user group according to the overall interest obtained in step S530. The decision can be made in the following manner: The overall interest for each program is compared with a predetermined threshold. If it is greater than the threshold, the program is put on the recommendation list provided to the user group, otherwise, the overall interest for the next program is compared and processed. In addition, the overall interest for all programs is prioritized first, and then the first nth programs in the ordered sequence are put in the recommendation list. Where n is a predetermined positive number.

Finally, in step S530, the obtained recommendation list is proposed to the user group. If the content recommendation is completed on the program provider side, the recommendation list is provided with the EPG by the program provider over a network to a user group device such as a set-top box or a PC. If it is completed on the user side, it may be obtained by the user device's processing of program information (eg, features and broadcast times included in the program) provided by the program provider.

The following description focuses on the development of equipment and apparatus and content recommendation methods that perform the above method of gaining full attention.

6 is a block diagram illustrating an embodiment of an apparatus for performing the method for obtaining full attention. Apparatus 600 includes an acquisition unit 610, a receiving unit 620, and a determining unit 630.

Acquisition unit 610 is for acquiring user data including a time varying mode of priority of all group members and their preferences for each of the features shown in Tables 1 and 2 and Tables 4 and 5.

For other ways of acquiring user data, the acquiring unit 610 may have a different implementation. For example, if the user data is obtained through estimation, the acquisition unit or device includes an automatic monitoring unit installed at the place; If user data is generated by the user himself, a device such as a keyboard, mouse, remote control, voice-input device can be used as the user data acquisition unit; Calculation of the total interest data is done in a device of the user group (e.g., television set, set-top box or PC), the user data is provided by the remote terminal computer via the network, and the communication function in the device of the user group The module for performing may be considered as a data acquisition unit.

The receiving unit 620 is used to receive program-related information (including the features included in the program and the broadcast time of the program). Of course, in certain embodiments, the receiving unit 620 may be integrated in hardware with the acquiring unit 610.

The determination unit 630 is used to obtain the overall interest of the user group for the content, the determination unit 630 including a profile generation unit 631, a retrieval unit 632, and a calculation unit 633 of the user group. .

The device 600 also includes a first memory unit 640, a second memory unit 650, and a third memory unit 660, where the first memory unit 640 is an acquisition unit 610 and a user group. Connected to a generating unit 631 to store user data; The second memory unit 650 is connected to the profile generation unit 631 of the user group to store profile data of the user group to be used by the retrieval unit 632. After the user data is updated regularly or intermittently, the profile generation unit 631 of the user group must recalculate the profile of the user group and update the second memory unit 650 with the new calculation result. The third memory unit 660 is connected to the calculation unit 633 and stores the storage result of the calculation unit 633.

The first, second and third memory units 640, 650, 660 may be volatile memory or nonvolatile memory, and may be realized as one memory if they are substantially different storage areas of the memory.

The generating unit 631 takes the user data from the first memory unit 640 and in each time partition according to the priority of each group member within each time partition and its preference for each feature. Determine the interest of the user group for each feature. These data of interest are stored in the format given as user profile data in the second memory unit 650. See the following detailed description for specific modes of calculation of relevant interest data.

The retrieval unit 632 is connected to the receiving unit 620 and corresponds in the second memory unit 650 according to the information on the features included in the program provided by the receiving unit 620 and the broadcast time of the program. Search for interest data and obtain interest of the user group for each feature included in the program within a time segment corresponding to the broadcast time of the program. In an environment where the broadcast duration of a television program spans a plurality of time divisions, the retrieving unit 632 performs a retrieval operation according to different processing modes.

The calculation unit 633 is connected to the search unit 632 to evaluate a weighted average or arithmetic mean value of the user group's interest in the features included in the television program obtained from the search. The arithmetic mean or weighted mean value is stored in the third memory unit 660 as the overall interest of the user group for the program.

7A is a block diagram illustrating an example of an apparatus 700a in accordance with the present invention for recommending content to a group of users.

The recommendation unit 710a includes a threshold comparison unit 711 and a recommendation list memory unit 712.

The threshold comparison unit 711 is connected to the third memory unit 660 in the unit 600 to compare the total interest stored one by one with a predetermined threshold. If the former is larger than the latter, the id of the corresponding program is output to the recommendation list memory unit 712. The recommendation list may be further displayed to the user.

7B is a block diagram illustrating another embodiment of an apparatus 700b in accordance with the present invention for recommending content to a group of users. The apparatus 700b shown in FIG. 7B differs from that of FIG. 7A in that the recommendation unit 710b and the ordering unit 713 replace the threshold comparison unit 711.

Ordering unit 713 is connected to third memory unit 660 of apparatus 600. It sorts the stored total interest in order and outputs the ids of the first one or several programs with the highest total interest to the recommendation list memory unit 712.

It should be understood that all of the above units and some and all of the parts included therein may also be realized in software.

The invention can also be realized as an application by a computer. The code contained in the program included in the computer is provided to a processor to form a device, and the code executed in the processor may perform the following steps: Priority value and feature of each group member in different time segments. Obtaining his preference for; Obtaining information about the content including the provision time of the content and the features included in the content; And determining total interest in accordance with the priority of each group member in the presentation time section and its interest in the features included in the content. The computer program product may be stored in a memory carrier.

In the present invention, a feature means a feature included in content that affects a user's interest; As such, there may be unique feature combinations for different user groups and content. Taking television programs as an example, they can be represented by features such as broadcast channels, titles, program casts, and program genres (e.g., drama, comic play, love stories, action movies, sports events), It affects what you miss. It is worth noting that how to customize feature combinations for user groups and content is not intended to be solved by the present invention and does not have a causal relationship with the technical effects achieved by the present invention; Accordingly, the knowledge of the foregoing aspects does not impose any limitation on the claims of the present invention.

Although the above description takes a television program as an example, the content referred to in the present invention should be understood in a broad sense. In fact, it includes all information that can be perceived by human organs such as visual, auditory, tactile and taste information. Its physical form includes, but is not limited to, various forms of optical, electrical and acoustic signals. Here is another example of content: Travel promotions sent to user groups. This material has attributes such as destination, cost and selection conditions that affect the recipient's interest as its content feature. In addition, the priority of group members may vary over time. For example, when Children's Day in June approaches, children in the family have more breathing than usual when deciding where to travel. Similarly, on Mother's Day, the mother's priority is high and the other members are low.

Those skilled in the art will also understand that all kinds of variations, modifications and variations based on the foregoing can be made. As such variations, modifications, and variations fall within the spirit and scope of the appended claims, they are encompassed within the invention.

Claims (15)

A method of estimating the overall interest of a user group for content, the user group comprising at least two members, wherein the method includes: a. Obtaining information about a possible time segment of the content and features included in the content; b. Obtaining a priority value of each member within a different time segment and the member's preference for the included features; c. Determining the total interest in accordance with the priority value of each member within the possible time partition and the member's preference for the included features. The method of claim 1, And the content is a television program. The method of claim 1, The priority value of each member is of overall interest, wherein the priority value of one member is different on at least two time segments, and wherein the two time segments are from a plurality of time segments divided by day. Estimation method. The method of claim 1, Wherein the priority value of each member is configured such that the priority value of one member is different on the same time segments on different days of at least two days. The method of claim 1, Step c, c1. Determining an interest value of the user group for the included features in a different time segment according to a priority value of each member in a different time segment and the preference for the included feature; c2. Determining an interest value of the group according to the possible time partition; And c3. Combining the interest values of the group to obtain the overall interest. The method of claim 5, The user group's interest value for the features combined in one time partition is the sum of the product of each member's preference for the included features and the priority value of each member within the time partition. Same as, overall interest estimation method. The method of claim 5, Step c1, Determining the interest value of the group of users in a fuzzy logic manner according to the priority value of each member within the different time partition and the preference for the included features. A method of recommending content to a user group, the user group comprising at least two members, the method comprising: a. Obtaining information about a possible time segment of the content and features included in the content; b. Obtaining a priority value of each member within a different time segment and the member's preference for the included features; c. Determining the overall interest in accordance with the priority value of each member in the possible time partition and the member's preference for the included features; And d. Recommending to the user group according to the overall interest. A device for estimating the overall interest of a user group for content, the user group comprising at least two members, the device comprising: Obtaining means for obtaining information about a possible time segment of the content and features included in the content; Receiving means for receiving a priority value of each member within a different time segment and the member's preference for the included features; And Determining means for determining the overall interest in accordance with the priority value of each member within the possible time partition and the member's preference for the included features. The method of claim 9, Wherein the priority value of each member is configured such that the priority value of one member differs on at least two time segments, and wherein the two time segments are from a plurality of time segments divided by day. The method of claim 9, Wherein the priority value of each member is configured such that the priority value of one member is different on the same time segment on different days of at least two days. The method of claim 9, The determining means, Generating means for determining an interest value of the user group for the included features in a different time segment according to a priority value of each member in a different time segment and the preference for the included features. ; Retrieving means for determining an interest value of the group according to the possible time partition; And Computing means for combining the interest values of the group to obtain the overall interest. A device for recommending content to a user group, wherein the user group includes at least two members, and the device includes: Obtaining means for obtaining information about a possible time segment of the content and features included in the content; Receiving means for receiving a priority value of each member within a different time segment and the member's preference for the included features; Determining means for determining a total interest according to the priority value of each member within the possible time partition and the member's preference for the included features; And And recommending means for recommending to the user group according to the overall interest. A computer program product for estimating the overall interest of a user group for content, the user group comprising at least two members, the computer program product comprising: Code for obtaining information about a possible time segment of the content and features included in the content; Code for obtaining a priority value of each member within a different time segment and the member's preference for the included features; And code for determining overall interest in accordance with the priority value of each member within the possible time partition and the member's preference for the included features. A memory carrier comprising a computer program product according to claim 14.

Images