Patent Application
20100205041
Today, digital data is available in numerous forms including multimedia data provided, for example, movies, broadcast television programs, home movies, or user-created video clips. Digital data can be stored in various devices and formats available for consumers. The amount of video media available to the consumers continues to grow at a very high rate. Broadcast, cable, or satellite companies often provide hundreds of different channels for the consumers to choose from. Movie rental companies such as Netflix and Blockbuster have tens, even hundreds, of thousands of titles on DVDs (digital video disc) or video cassettes. More recently, the Internet has facilitated distribution of content media world-wide. Sites such as YouTube have immense video collections, often millions of video clips, contributed by users from all over the world.
Of course, the content of digital data may vary widely. As a result, various rating systems have been developed to help the consumers make informed choices. For example, a star rating system is often used by movie critics and viewers to rate movies or films. For television programs, the Nielsen Ratings is a well-known system that measure audience viewing results. Movie rental companies such as Netflix or Blockbuster and Internet sites such as YouTube or Amazon allow viewers to rate and/or comment on individual movies or video clips manually, such as by selecting a number of stars for each video rated, and these individual ratings are combined or averaged to provide an overall rating for the particular video. These rating systems often reflect the popularity of the videos.
Ratings may be used for different purposes. For example, people may choose to watch a movie based on it star rating. Sponsors, rental companies, or Internet sites may recommend videos based on their ratings. Individuals with similar likes may be offered similar items and/or items selected by one individual can be offered to other individual with similar likes, and so on.
Generally, it is highly useful to know the interest that an individual may have in an item of interest. As such, techniques for determining the interest of individuals are highly useful.
Broadly speaking, the invention relates to determining the interest that an entity may have in one or more items (or items of interest) using computing environments and computing systems.
More particularly, the invention pertains to determining the interest of individual entities based on an interest that may be associated with the entities collectively. It will be appreciated that the invention can, for example, be provided for various computing environments and/or computing systems operable to use digital data stored in various forms.
In accordance with one aspect of the invention, an interest value indicative of the interest of a particular entity in one or more items can be determined based on a general interest value (e.g., a group interest value, a group preference value) associated with a plurality of entities (e.g., person, members of a group, an individual) that include that particular entity. It will be appreciated that the interest value can be determined based on Collaborative Filtering (CF) data and/or individual (or non-collaborative) data. In contrast to the Collaborative Filtering (CF) data which may include data associated with various entities, the individual (or non-collaborative) data typically pertains to one entity, namely, the entity whose interest is to be determined.
In accordance with one embodiment of the invention, an interest value indicative of the interest of a particular entity can be determined based on one or more estimated (or predicted) interest values. The one or more estimated (or predicted) interest values can include a Collaborative Filtering (CF) interest value determined based on Collaborative Filtering (CF) data using a Collaborative Filtering (CF) scheme. In addition or alternatively, one or more individually-based interest values can be determined based on individual (or non-collaborative) data. The individually-based interest values can, for example, be determined based on content-based or non-content-based data including, for example, profile data (e.g., age, gender), usage data (e.g., number of hours spent by a particular individual to watch TV), and various other data (e.g., data pertaining likes and/or dislikes of a particular user with respect to specific types of content, such as, for example, “horror” movies, “rock” music,” and so on)) that may be available. The interest value indicative of the interest of a particular individual can, for example, be determined by effectively comparing the one or more estimated (or predicted) interest values with the general interest value associated with the plurality of entities that include that particular entity. By way of example, the respective difference of each one of the one or more estimated interest values to the general interest value can be determined as one or more prediction errors. One or more prediction errors can be used to determine the interest of a particular individual. As such, given a numerical general interest value of three (3) and corresponding collaboratively-based interest value of “4.8” and individually-based interest value of “5.5”, two differences (or prediction errors) can be determined, namely, a first difference: “1.8” (4.8-3) and a second difference: “2.5” (5.5-3). The first and the second differences can be used to determine an estimated interest value indicative of the interest of the individual based on the general interest value of three (3) as will be known to those skilled in the art. For example, the average of differences can be determined to be “2.15” (2.5+1.8/2) and used to assess the likelihood that the general interest value is reflective of the individual's interest (e.g., 10% probability) and/or estimate the interest of the individual (e.g., determine the interest of the individual to be “4.3”, a value significantly higher that the general interest value).
The invention can be implemented in numerous ways, including, for example, a method, an apparatus, a computer readable (and/or storable) medium, and a computing system (e.g., a computing device). A computer readable medium can, for example, include at least executable computer program code stored in a tangible form. Several embodiments of the invention are discussed below.
Other aspects and advantages of the invention will become apparent from the following detailed description, taken in conjunction with the accompanying drawings, illustrating by way of example the principles of the invention.
The present invention will be readily understood by the following detailed description in conjunction with the accompanying drawings, wherein like reference numerals designate like structural elements, and in which:
As noted in the background section, generally, it is highly useful to know the interest of individuals in various items of interest. As such, techniques for determining the interest of individuals can be highly useful.
In some situations, a general interest associated with a plurality of entities may be known. By way of example, a general interest (or preference) value for a group may be known as a numerical value (e.g., three (3) in a range of one (1) to five (5)). This general interest value may be provided by only one of the members of the group and can reflect the actual and/or expressed interest of the member in one or more items. However, the general interest value may be provided in an anonymous manner without identifying the member of the group which has actually provided the interest value. For example, a general interest value (e.g., an interest rating) can be associated with a family including a father, a mother, and two children of various ages. Any one the members of the family can rate a program viewed on Television and provide the rating for the family without having to identify themselves as the father, mother, or one of the children. In such situations, it would be useful to have the ability to make an assessment regarding the interest of each member of the family based on the general interest.
More generally, individual identity of various entities may not be always known even though a general rating associated with them may be known. Also, it is generally desirable to allow individuals to express their interest anonymously in order to protect their identity. As such, it would be useful to determine the interest of an individual entity based on a general interest that may be loosely associated with that particular entity.
The invention pertains to determining the interest of individual entities based on an interest that may be associated with the entities collectively. It will be appreciated that the invention can, for example, be provided for various computing environments and/or computing systems operable to use digital data stored in various forms.
In accordance with one aspect of the invention, an interest value indicative of the interest of a particular entity in one or more items can be determined based on a general interest value (e.g., a group interest value, group preference value) associated with a plurality of entities (e.g., person, members of a group, an individual) that include that particular entity. It will be appreciated that the interest value can be determined based on Collaborative Filtering (CF) data and/or individual (or non-collaborative) data. In contrast to the Collaborative Filtering (CF) data which may include data associated with various entities, the individual (or non-collaborative) data typically pertains to one entity, namely, the entity whose interest is to be determined.
In accordance with one embodiment of the invention, an interest value indicative of the interest of a particular entity can be determined based on one or more estimated (or predicted) interest values. The one or more estimated (or predicted) interest values can include a Collaborative Filtering (CF) interest value determined based on Collaborative Filtering (CF) data using a Collaborative Filtering (CF) scheme. In addition or alternatively, one or more individually-based interest values can be determined based on individual (or non-collaborative) data. The individually-based interest values can, for example, be determined based on content-based or non-content-based data including, for example, profile data (e.g., age, gender), usage data (e.g., number of hours spent by a particular individual to watch TV), and various other data (e.g., demographic data) that may be available. The interest value indicative of the interest of a particular individual can, for example, be determined by effectively comparing the one or more estimated (or predicted) interest values with the general interest value associated with the plurality of entities that include that particular entity. By way of example, the respective difference of each one of the one or more estimated interest values to the general interest value can be determined as one or more prediction errors. One or more prediction errors can be used to determine the interest of a particular individual. As such, given a numerical general interest value of there (3) and corresponding collaboratively-based interest value of “4.8” and individually-based interest value of “5.5”, two differences (or prediction errors) can be determined, namely, a first difference: “1.8” (4.8-3) and a second difference: “2.5” (5.5-3). The first and the second differences can be used to determine an estimated interest value indicative of the interest of the individual based on the general interest value of there (3) as will be known to those skilled in the art. For example, the average of differences can be determined to be “2.15” (2.5+1.8/2) and used to assess the likelihood that the general interest value is reflective of the individual's interest (e.g., 10% probability) and/or estimate the interest of the individual (e.g., determine the interest of the individual to be “4.3”, a value significantly higher that the general interest value).
Embodiments of these aspects of the invention are discussed below with reference to
Referring to
An item of interest Ii can, for example, be one or more media items, one or more applications, one or more audio files, one or more video files, one or more songs, one or more movies, and so on. Typically, the general interest value 102 can be loosely associated with one or more of the entities (E1-En) and it represents the actual and/or expressed interest of one or more of the entities (E1-En). However, the one or more entities can be anonymous with respect to their actual and/or expressed interest. By way of example, the general interest value 102 can be a general interest value associated with first, second and third entities (E1, E2 and E3) represented by a first group (G1). Furthermore, the general interest value 102 can, for example, be indicative of the expressed interest of a particular one of the entities in one or more items of interest. By way of example, the general interest value 102 can be indicative of the actual and/or expressed interest of a second entity (E2) in a third item of interest (I3). However, the general interest value 102 can be received anonymously as a general (or ambiguous) interest value associated with the first group (G1) including the first, second and third entities (E1, E2 and E3).
It will be appreciated that the computing system 100 can determine an interest value 104 for an entity E1 based on the general interest value 102. The interest value 104 indicative of the interest of a particular entity Ei can, for example, be expressed as the likelihood and/or probability that the general interest value 102 is indicative of the particular entities' (Ei's interest). By way of example, when the general interest value 102 is obtained as a general interest value associated with a group of entities including the first, second and third entities (E1, E2 and E3), the interest value 104 for the first, second and third entities (E1, E2 and E3) can, for example, be expressed as a probability (e.g., 10% probability that the general interest value 102 is indicative of the interest of the first entity E1, 75% probability that the general interest value 102 is indicative of the interest of the second entity E2 and 15% probability that the general interest value 102 is indicative of the interest of the third entity E3). It should be noted that the interest value 104 determined for a particular entity 104 can be expressed in many different forms including, for example, an estimated interest value. As such, if the general interest value 102 is expressed as a numerical (e.g., “4” in a range of “0-6”) the interest value 104 can be expressed as a numerical for each one of the entities associated with the general interest value 102.
As noted above, the computing system 100 can determine the interest value 104 as a particular or specific interest value for a particular entity Ei based on a general interest value 102 associated with a plurality of entities including the entity Ei. It should be noted that the computing system 100 can also obtain a Collaborative Filtering (CF) interest value 106 and individual data 108 for an entity Ei. Moreover, the computing system 100 can determine the interest value 104 for a particular entity Ei based on the CF interest value 106 and/or individual data 108 for the entity Ei. It will be appreciated that CF interest value 106 can represent a Collaborative Filtering (CF) interest value determined based on a Collaborative Filtering (CF) technique as a collaboratively estimated interest of a particular entity Ei in one or more items of interest (I1-In), as will be known to those skilled in the art. The CF interest value 106 can be determined based on CF data 110.
It should be noted that the computing system 100 can be operable to determine the CF interest value 106 based on the CF data 110 and/or obtain the CF interest value 106 as input. In addition to the CF interest value 106, the computing system 100 can also use the individual data 108 pertaining to a particular entity Ei in order to determine the interest value for that particular entity as an interest value 104. In contrast to the CF interest value 106 determined based on the CF data 110, an estimated individual interest value 112 can be determined based on the individual data 108 representing non-collaborative data or data specifically pertaining to a particular entity Ei. In other words, the computing system 100 can be operable to utilize both collaborative data and non-collaborative (or individualized data) in order to determine the interest value 104 based on a general interest value 102.
Generally, the individual data 108 pertaining to an individual entity Ei can include one or more known properties, characteristic and/or attributes of the entity. The individual data 108 can, for example, include content-based data and non-content based data. The content-based data can be indicative of one or more attributes and/or factors that can be considered in view of content of one or more items of interest I1-In in order to make an assessment regarding the interest of a particular individual Ii with respect to the one or more items of interest given the general interest value 102. In contrast, non-content based data can be indicative of one or more attributes and/or factors that can be considered regardless of content of the one or more items of interest (I1-In) in order to make an assessment regarding the interest of a particular entity Ii with respect to the one or more items of interest (I1-In).
As noted above, the individual data 108 can be used to effectively determine an estimated individual interest value 112 for a particular entity Ei. The computing system 100 can be operable to effectively consider an individual interest value 112 and/or a CF interest value 106 in order to determine the interest value 104 given a general interest value 102. As a result given a general interest value 102, CF data 110 and/or individual data 108 can be effectively used by the computing system 100 in order to determine an individual interest value 104 for a particular individual Ei expressed in various forms.
The computing system 100 can, for example, be operable to determine the interest value 104 for a particular entity Ei based on the difference of the CF interest value 106 and the general interest value 102 and/or the difference between the general interest value 102 and an estimated individual interest value 112 determined based on the individual data 108.
It should be noted that the interest value 104 can be provided as CF data 110, thereby effectively providing feedback based on an estimated interest value. It should also be noted that the interest value 104 can be effectively marked in the CF data 110 in order to identify it as an interest value that has been provided as an estimation of interest rather than an interest value that may reflect a real and/or expressed interest or determined based on other data.
Furthermore, it will be appreciated that the interest value 104 can be used for various applications including making recommendations with respect to one or more items of interest.
Referring to
It should be noted that the individual interest values 204 can be provided as effective feedback to the CF based system 212 and/or individually-based system 214 and effectively marked accordingly.
Referring to
However, if it is determined (260) to use individual data, it is determined (264) whether to use content-based data and a second interest value can be determined based on the content of the more items of interest in order to determine a second difference value on the difference of the second interest value and the group interest value.
Referring now to
It should be noted that if it is determined (254) not to use Collaborative Filtering (CF) data, the method 250 proceeds to use at least content-based data, non-content based data (262) or other individual data pertaining to the first member in order to determine (262) the interest of the member.
Generally, the interest of a member can be determined based on one or more of the difference values noted above. Content-based data includes one or more of the following: a profile of said person, a profile of said person that includes his or her age, occupation, state and/or country of residence, address, and a phone number. A recommendation regarding one or more items can be made to an entity based on the determined interest of that entity.
As noted above, the techniques described above can be effectively combined by various other techniques related to rating content.
For example, the techniques of the invention can be effectively combined the techniques described in U.S. patent application Ser. No. 12/120,217 entitled: “SYSTEM AND METHOD FOR AUTOMATICALLY RATING VIDEO CONTENT” (SISAP022) filed on May 13, 2008 which is hereby incorporated by reference herein in its entirety and for all purposes.
More particularly, a rating system can automatically determines the content ratings of the videos that have been operated on a media device in the sense that users of the media device are not required to take any specific actions to rate these videos, such as in the case of, for example, Netflix, YouTube, or Amazon, where users need to manually select a rating level for each video. Instead, with the rating system described above, the device usage actions are monitored and used to adjust the video content ratings in reference to the rating rules without requiring user actions with respect to the rating process. The device usage actions are actions associated with performing video operations on the media device, not rating the video content. Rating rules may be added, modified, or deleted as needed. Different rules may be designed for different types of media devices according to the types of video operations they support.
An automatic rating system may, for example, be used to help provide personalized video content to users of various types of consumer devices, such as televisions, video recorders, cable or set top boxes, and portable media players. System and methods for providing personalized video content are described in more detail in U.S. patent application Ser. No. 12/120,203 (Attorney Docket No. SISAP021/CSL07-NW14-A), entitled “PERSONALIZED VIDEO SYSTEM” by Gibbs et al., filed on May 13, 2008 which is hereby incorporated by reference herein in its entirety and for all purposes. Video content ratings obtained from a specific consumer electronic device can, for example, be used by a video content provider server to rank the pieces of video content selected for the user(s) of that consumer electronic device so that the selected pieces of video content are presented to the user(s) based on their ranks.
Techniques for ranking pieces of video content for individual users is described in more detail in U.S. patent application Ser. No. 12/120,211 (Attorney Docket No. SISAP035/CSL07-NW16), filed on May 13, 2008, entitled “COMBINATION OF COLLABORATIVE FILTERING AND CLIPRANK FOR PERSONALIZED MEDIA CONTENT RECOMMENDATION” by Nemeth et al. and co-pending U.S. patent application Ser. No. 12/120,209 (Attorney Docket No. SISAP036/CSL07-NW17), filed on May 13, 2008, entitled “CLIPRANK: RANKING MEDIA CONTENT USING THEIR RELATIONSHIPS WITH END USERS” by Rathod et al., both of which are hereby incorporated by reference in their entireties and for all intents and purposes. According to some embodiments, the video content ratings obtained from a specific consumer electronic device are used obtain personalized ClipRank on a set of video content for the user(s) of that consumer electronic device.
The various aspects, features, embodiments or implementations of the invention described above can be used alone or in various combinations. The many features and advantages of the present invention are apparent from the written description and, thus, it is intended by the appended claims to cover all such features and advantages of the invention. Further, since numerous modifications and changes will readily occur to those skilled in the art, the invention should not be limited to the exact construction and operation as illustrated and described. Hence, all suitable modifications and equivalents may be resorted to as falling within the scope of the invention.
