Method and apparatus for generating television program recommendations based on prior queries

ES 2 346 293 T3

DESCRIPTION

Method and apparatus for generating television program recommendations based on previous inquiries.

Field of the invention

The present invention relates to television program recommenders, and more particularly, to a method and apparatus for generating television program recommendations based on queries that have been made by a user on an electronic program guide.

Background of the invention

As the number of channels available to television viewers has increased, in addition to the diversity of programming content available on such channels, it becomes increasingly difficult for viewers to identify television programs of interest. Historically, viewers identified television programs of interest by analyzing printed guides to television programs. Typically, such printed television program guides contained gridded listings of available television programs with time and date, channel and title. As the number of television programs has increased, it becomes increasingly difficult to effectively identify the programs using such printed guides.

More recently, television program guides have become available in an electronic format, often referred to as electronic program guides (EPGs). Like printed television program guides, EPGs contain grids that list available television programs with time and date, channel and title. Some EPGs, however, allow viewers to sort or search available television programs according to personalized preferences. In addition, EPGs allow an on-screen presentation of available television programs.

Although EPGs allow viewers to identify desired programs more effectively than conventional printed frames, they suffer from a number of limitations, which if exceeded could further improve viewers' chances of identifying desirable programs. For example, many viewers have a particular preference for, or bias against, certain categories of programming, such as action shows or sports programming. Thus, viewers' preferences can be applied to the EPG to obtain a set of recommended programs that may be of interest to a particular viewer.

Therefore, a number of tools have been proposed or suggested to recommend television programming. The Tivo system^TM, for example, commercially available from Tivo, Inc. of Sunnyvale, California, allows viewers to rate the show using "Thumbs Up and Thumbs Down" signals indicative of shows the viewer likes or dislikes, respectively. After which, the TiVo receiver compares the recorded viewer preferences with the received program data, such as an EPG, to make appropriate recommendations to each viewer.

WO 98/37696 describes a broadcast data distribution system that distributes indexed information and directories for selection of display options in broadcast and multicast networks with asymmetric upstream / downstream bandwidths. The system automatically builds both a target profile for each object (program) that is broadcast, as well as a "target profile interest summary" for each subscriber, whose target profile interest summary describes the level of interest of the subscriber in various types. of white objects. The system then evaluates the target profiles based on the interest summaries of the subscriber's target profiles to generate an ordered list of personalized subscriber ranges of objects that are most likely to interest each subscriber, so that the subscriber you can select the ones you want from these potentially interesting targets, which were automatically selected by this system from the plethora of target objects available on the data distribution system.

US 6,005,597 describes a method and apparatus for the selection of television programs that observe the preferences of a viewer to create a dynamic viewer profile that is used for available evaluation programs. Based on the viewer's profile, the available programs are ranked and presented to the viewer in descending order of predicted interest. This allows a viewer to quickly find the program of greatest interest to him without having to tediously search through a large number of available programs.

Such tools for generating television program recommendations provide selections of programs that a viewer may like, based on the viewer's past history. Even with the help of such program advisers, however, it is still difficult for a viewer to identify programs of interest among all the options. Additionally, currently available tools that search the electronic program guide based on a user-defined query require various button activations before the user can review the list of programs that satisfy the query. In addition, there is currently no way to integrate the explicit information determined from the queries made by the user in the electronic program guide with the implicit information found from the user's viewing habits.

ES 2 346 293 T3

There is therefore a need for a method and apparatus for recommending television programs based on the queries that have been executed by a user on the electronic program guide.

Summary of the invention

According to the invention there are provided a method according to independent claim 1 and a system according to independent claim 12. Favorable embodiments are defined in the dependent claims.

Generally, a method and apparatus are described for generating television program recommendations based on queries that have been made by a user on an electronic program guide. The present invention adjusts a standard program recommender score based on previous searches that have been executed by the user. In particular, the conventional program recommender score for a given program is adjusted according to the degree of correlation between the pairs of values ​​of the attributes that define the program and the pairs of values ​​of the attributes that have been previously searched by the program. Username.

A historical search database is maintained to indicate the number of times each attribute-value pair appears in a user search. Each time a manual or automatic search is initiated by the user, the search is decomposed to identify the attribute-value pairs specified by the user. The historical search database captures a user's search activity and provides additional information regarding user preferences. Higher frequency counts for certain attribute-value pairs imply user preference for programs that meet such criteria.

A more complete understanding of the present invention, as well as more features and advantages thereof, will be obtained with reference to the following detailed description and drawings.

Brief description of the drawings

Figure 1 illustrates a television schedule recommender in accordance with the present invention;

Figure 2 is a table of samples from the viewer profile database of Figure 1;

Figure 3 is a table of samples from the program database of Figure 1;

Figure 4 is a table of samples from the historical search database of Figure 1; and Figure 5 is a flow chart depicting an exemplary program recommendation procedure incorporating principles of the present invention.

Detailed description

Figure 1 illustrates a television schedule recommender 100 in accordance with the present invention. As shown in Figure 1, the television schedule recommender 100 evaluates each of the programs in an electronic program guide (EPG) 110 to identify programs of interest to a particular viewer. The set of recommended programs can be presented to the viewer, for example, using a terminal / television set top 180 using well known on-screen display techniques.

In accordance with a feature of the present invention, television schedule recommender 100 generates television program recommendations based on searches that have been performed by a user on electronic program guide 110. As further discussed above, the program recommender score generated in accordance with conventional techniques is adjusted based on previous searches that have been performed by the user. In particular, the conventional program recommender score for a given program is adjusted according to the degree of correlation between the program attributes and the attributes that have been previously searched by the user.

Generally, each time a manual or automatic search is initiated by the user using one or more search commands, the television schedule advisor 100 decomposes the query to identify the attribute-value pairs specified by the user. A historical search database 400, examined below in conjunction with Figure 4, is maintained to indicate the number of times each attribute-value pair has appeared in a user search. Therefore, the value of the corresponding count is incremented in the historical search database 400 each time the attribute-value pair appears in the decomposed search. In this manner, the historical search database 400 captures the user's search activity and thus provides additional information regarding the user's preferences. Higher frequency counts for certain attribute-value pairs imply user preference for programs in accordance with such criteria.

The television program recommender 100 may be incorporated as any computing device, such as a personal computer or workstation, that contains a processor 150, such as a central processing unit (CPU), and a memory 160, such as RAM or ROM. In addition, the television schedule recommender 100 may be incorporated like any available television program recommender, such as

ES 2 346 293 T3 the Tivo system^TM, commercially available from Tivo, Inc., of Sunnyvale, California, or the recommender television programs described in US Patent 6,727,914 B1, entitled "Method and Apparatus for Recommending Television Programming Using Decision Trees", and United States US 7,051,352 B1, entitled "Adaptive TV Program Recommender", or any combination thereof, as modified herein to incorporate the features and functions of the present invention.

As shown in Figure 1, and better discussed below in conjunction with Figures 2 through 5, respectively, the memory 160 of the television schedule recommender 100 includes one or more viewer profiles 200, a database 300 of programs , a historical search database 400, and a program recommendation process 500. Generally, illustrative viewer profile 200 indicates a relative viewer interest level for each program attribute. The program database 300 records information for each program that is available in a given time interval. The historical search database 400 indicates the number of times each attribute-value pair has appeared in a user query. Finally, the program recommendation process 500 generates recommendation scores for each program in a particular time interval, taking into account previous queries that have been made in the electronic program guide (EPG) 110 in accordance with the present invention.

FIG. 2 is a table illustrating an example display profile 200. It is noted that viewer profile 200 may be associated with a particular user or group of individuals, such as a family, as will be apparent to a person of ordinary skill in the art. It should be noted that the viewer profile 200 can be generated explicitly, based on responses to a study, or implicitly, based on the set of samples that were observed (and / or not observed) by the observer during a period of time, or a combination of the above.

As shown in Figure 2, viewer profile 200 contains a plurality of records 205-213 each associated with a different program attribute. In addition, for each attribute set in column 240, viewer profile 200 provides a numerical representation in column 250, indicating the relative level of viewer interest in the corresponding attribute. As discussed below, in the illustrative viewer profile 200 set forth in Figure 2, a numerical scale between 1 ("hate") and 7 ("love") is used. For example, the viewer profile 200 set forth in Figure 2 has numerical representations indicating that the user particularly enjoys programming on the sports channel, as well as programming late in the afternoon.

Although the viewer profile 200 is illustrated using an explicit viewer profile, the viewer profile 200 can also be incorporated using an implicit profile, or a combination of implicit and explicit profiles as will be apparent to one of ordinary skill in the art. For an examination of a recommender television program 100 that employs implicit and explicit profiles to obtain a program recommendation score, see, for example, document, WO-A-02/25938, entitled "Method and Apparatus for Generating Program Scores. Recommendation That They Use Implicit and Explicit Vision Preferences ”.

In an exemplary embodiment, the numerical representation in profile 200 includes an intensity scale such as:

<td>Number</td><td>Description</td>

<td> 1</td><td>Oh, God</td>

<td> 2</td><td>Aversions</td>

<td> 3</td><td>Moderately negative</td>

<td> 4</td><td>Neutral</td>

<td> 5</td><td>Moderately positive</td>

<td> 6</td><td>Similarities</td>

<td>⁷</td><td>Loves</td>

Figure 3 is a sample table from the database 300 of the program of Figure 1 that records information for each program that is available in a given time interval. The data that appears in the base 300 of

ES 2 346 293 T3 program data can be obtained, for example, from electronic program guide 110. As shown in Figure 3, the program database 300 contains a plurality of records, such as records 305-320, each associated with a given program. For each program, the program database 300 indicates the date / time and channel associated with the program in fields 340 and 345, respectively. In addition, the title and genre for each program are identified in fields 350 and 355. Additional well-known attributes (not shown, such as actors, duration, and program description, may also be included in the database 300 of programs.

The program database 300 may optionally also record an indication of the recommendation score (R) assigned to each program by the television schedule advisor 100 in field 370. In addition, the program database 300 you may also optionally indicate in field 370 the adjusted recommendation score (A) assigned to each program by counselor 100 in accordance with the present invention. In this way, the numerical scores, adjusted by the present invention, can be presented to the user in the electronic program guide with each program directly or represented in a color spectrum or other visual indication that allows the user to quickly locate the programs of interest.

As noted above, the historical search database 400 indicates the number of times each attribute-value pair has appeared in a manual or automated user query. As shown in Figure 4, the historical search database 400 is comprised of a plurality of records, such as records 405 through 415, each associated with a given attribute-value pair. For each attribute-value pair, the historical search database 400 indicates the corresponding number of times the attribute-value pair has appeared in a user query (frequency of use). As previously noted, each time a manual or automatic search is initiated by the user, the television schedule advisor 100 decomposes the query and increments the counter in the historical search database 400 for each attribute-pair. value that appears in the decomposed query.

In addition, to facilitate the calculations performed by the program recommendation procedure 500, discussed below, the historical search database 400 optionally indicates a normalized frequency of use of the term, N, in field 470. For example, the normalized score, N, indicated in field 470 can be obtained by executing a manual representation of the actual frequency of the term of use with a value between zero and one for each of the various associated attribute-value pairs. with an attribute. In an exemplary embodiment, the normalization, in the historical search database 400 includes a usage scale frequency such as:

<td>Real value</td><td>Value Normalized</td>

<td> 0</td><td> 0</td>

<td>0.2 (VALUE MAXIMUM)</td><td> 0,2</td>

<td>0.4 (VALUE MAXIMUM)</td><td> 0,4</td>

<td>0.6 (VALUE MAXIMUM)</td><td> 0, 6</td>

<td>0.8 (VALUE MAXIMUM)</td><td> 0,8</td>

<td>MAXIMUM VALUE</td><td> 1,0</td>

In an alternative implementation, normalization in the historical search database 400 can be obtained by plotting a curve through various count values ​​of the frequency for each of the various attribute-value pairs associated with an attribute, of a known way.

Figure 5 is a flow chart describing an exemplary program recommendation process 500 incorporating principles of the present invention. As shown in Figure 5, procedure 500 of

ES 2 346 293 T3 program recommendation initially obtains the electronic program guide (EPG) 110 during operation 520. After which, the program recommendation procedure 500 calculates the program recommendation score, R, during operation 520 for each program in the time period of interest in a conventional manner (or gets a program recommendation score, R, from a conventional counselor).

After which, the program recommendation process 500 calculates the adjusted program recommendation score, A, during run 530 for each program in the time period of interest, as follows:

k

A = MIN {R + R (£ N¡ · WEIGHT ^ l ^ R} í »l where k is the total number of attribute-value pairs indicated in field 470 of the historical search database 400. Generally, the calculation performed during operation 530 ensures that the fit to the conventional program recommendation score, R, does not exceed an exemplary value of 35%, that is, a maximum of 135% of the recommendation score, R, conventional. In addition, the adjustment in the score of the conventional program recommendation assigned to a given program is obtained by adding the term N of the weighted normalized frequency, for each attribute-value pair associated with the program that has been previously searched by the user.

The WEIGHT contribution of each attribute within a television program can be set by the user, or determined empirically. For example, the date / time attribute can be assigned a weight of 5%, the genre attribute can be assigned a weight of 20%, and the channel attribute can be assigned a weight of 10%. Therefore, if a given program is a comedy, the adjustment to the R score of the conventional program recommendation, attributable to the pair of attribute values ​​"comedy genre" will be .8 (N) multiplied by 20%, which is the weight assigned to the gender attribute.

The program recommendation procedure 500 calculates the combined program recommendation score, C, during run 540 for each program in the time period of interest, as follows:

C = MIN {A, 100}

Thus, in addition to ensuring that the conventional program recommendation R score adjustment does not exceed an exemplary value of 35% (see operation 530 above), the exemplary program recommendation procedure 500 guarantees also during operation 540 that the combined program recommendation score, C, does not exceed 100% (the maximum score).

Finally, the program recommendation procedure 500 provides the combined program recommendation scores (C) for the programs in the time period of interest to the user during operation 550, before the program control ends.

In further variations of the program recommendation procedure 500, the adjusted program recommendation score, A, may be calculated during operation 530 using a bonus scoring system in which a predefined or fixed bonus is determined, for example, based on the number of attribute-value pairs that define the program that has been previously searched by the user. In other words, a voucher can be determined based on the number of attribute-value pairs that match in the current program with those in the historical search database 400. For example, if four attribute-value pairs in the historical search database 400 match attribute-value pairs from the current program, then a bonus of, for example, 10%, can be awarded to increase the score, R, from the recommender of conventional programs, 10%.

It is to be understood that the recommendations and variations shown and described herein are merely illustrative of the principles of this invention and that various modifications may be made by those skilled in the art without departing from the scope of the invention as defined by the claims.