RATING MEDIA ITEM RECOMMENDATIONS USING RECOMMENDATION PATHS AND/OR MEDIA ITEM USAGE

[0001] The present invention relates to determining a recommendation rating for a recommended media item based on user recommendation paths and/or usage or play of the media item.

[0002] In recent years, there has been an enormous increase in the amount of digital media available online. Services, such as Apple's itunes® for example, enable users to legally purchase and download music. Other services, such as Yahoo!® Music Unlimited and RealNetwork's Rhapsody® for example, provide access to millions of songs for a monthly subscription fee. YouTube® provides users access to video media. As a result, media items have become much more accessible to consumers worldwide. However, the increased accessibility of media has only heightened a long-standing problem for the media industry, which is namely the issue of linking users with media that matches their preferences.

[0003] Many companies, technologies, and approaches have emerged to address this issue of media recommendation. Media item recommendations may be provided to users as suggestions based on information about the user and/or their media likes or dislikes, also called preferences. Media recommendations may be provided either by service provider companies or by the user's online “friends” (typically identified by user id). In the case of company provided recommendations, some companies assign ratings to attributes of an identified media that are assembled to create a holistic classification for the media that is then used by a recommendation engine. Other companies take a communal approach wherein recommendations are based on the collective wisdom of a group of users with similar tastes by profiling the habits of a particular user and then searching similar profiles of other users.

[0004] When a user receives media recommendations, an associated play or popularity rating may follow. The rating is designed to be indicative of the relative play or usage of the media item among users. This allows the user to organize their media item selections for view and/or usage according to popularity. The popularity rating is typically maintained on a remote server or a super peer device accessible to other users' networked devices. The user can select which of their media is played based on the popularity. The play rating may be updated with the user's own usage of given media items which is reported as an input to the popularity rating algorithm for the media item.

[0005] One issue associated with popularity or play ratings is that they are not truly indicative of media item usage as a result of recommendations. The play or usage of a media item may affect its play rating regardless of whether such play or usage was attributable to a recommendation. Thus as an example, a single user playing a given media item one thousand (1000) times may increase the play rating of the media in the same manner as one hundred (100) different users each playing the given media item ten (10) times each. Thus, the play rating can be skewed and even manipulated to artificially increase its popularity by a single or small group of users. One hundred (100) different users playing a media item as a result of recommendations may be more useful information to a user in determining which media item to play. However, it cannot be determined whether a media item's play rating was a result of recommendations, and if so, to what extent.

[0006] The present invention is a media item recommendation rating system and method. A recommendation rating for a media item is established and dynamically updated in response to the media item being recommended to other users. A recommendation server or other device receives a report of a media item recommendation and updates a recommendation rating in response. The recommendation rating may also be updated based on how often media items are used or played in response to a received recommendation. Thus, a media item's recommendation rating is affected by events relating to its recommendation, as opposed to simple play-based ratings that are updated on play actions regardless of whether related to a recommendation or not. Simple play-based ratings do not distinguish between ordinary usages or plays and those resulting from recommendations. Recommendation of a media item to another user is a better indicator of the user's likeability of a media item, since a recommendation is an endorsement of sorts. Many users may be interested or like a media item enough to use or play it, but not enough to recommend it. Uses or plays of media items short of a recommendation do not affect the recommendation rating as provided by the present invention.

[0007] In one embodiment, a recommendation server is provided that is operably connected to a network, such as the Internet for example. One or more user devices are coupled to the network. When a user device makes a recommendation for a media item to another user device, a message is sent to a recommendation server, which includes the originating recommender id, the destination, and the id of the media item being recommended. The recommendation server establishes a recommendation rating for the media item if a record of the media item does not already exist in recommendation tables stored in memory. Each unique media item recommended by a user has its own entry and recommendation rating in the recommendation tables. In this manner, and each time a media item is recommended to another user, the recommendation server dynamically updates the recommendation rating for the media item to affect, and in the typical case increase, its rating. After the recommendation rating is either established or updated, the recommendation server may forward the media item recommendation along with the recommendation rating to the recipient. The recipient client may contain a scoring engine that relies on the recommendation and/or other metrics to calculate a recommendation score for the media item, or to sort the recommendations according to their recommendation rating.

[0008] The recommendation rating may be altered each time a recommendation path for a given media item is made regardless of whether the recommendation path is a duplicate, or altered based only on new recommendation paths for a given media item. The recommendation server updates one or more recommendation tables to store current and previous recommendation paths for media items between users. This allows the recommendation rating to only be altered based on new recommendation paths for a given media item instead of duplicated recommendation paths for the media item.

[0009] The recommendation server may also alter or increase a media item recommendation rating based on the media item being played by another user after receiving the recommendation. In this manner, not only is the recommendation rating affected by number of recommendations made, but the recommendation rating is also based on the effectiveness measured by the recipient's play or usage of the recommended media item. In order to determine if the media item used or played by a user was the result of a previous recommendation, the recommendation tables established and maintained by the recommendation server are consulted. In this manner, it can be determined if a user received the used or played media item as a result of a recommendation. If not, the user is simply playing the media item without the item having been recommended, which is not used to update the recommendation rating. Otherwise, the system would simply just be a play rating system, which is not the intention of the present invention.

[0010] The recommendation server may also be responsible for forwarding the recommended media item to the intended recipient user. Alternatively, the recommendation server may simply be responsible for establishing and maintaining the recommendation rating, wherein the actual recommendation of the media item is forwarded between users in another manner, such as through another server or through a peer-to-peer communication.

[0011] As an alternative to a recommendation server, a super peer device may be provided to perform some or all of the same aforementioned functionalities of the recommendation server. A super peer device may include a database for storing the recommendation tables and recommendation ratings in memory, or a network recommendation database separate from the super peer device may be provided on a network to provide storage for the recommendation tables and ratings. The super peer device may send and receive recommendations to and from other peer user devices, in addition to receiving reports regarding media item recommendations. Alternatively, the other user devices may send the actual media item recommendations directly to other users in a peer-to-peer communication with a separate reporting message only being communicated to the recommendation server or the super peer device to establish and update the recommendation ratings for media items. The peer recipient may then contact the recommendation server or super peer device, as the case may be, to obtain the recommendation rating for the recommended media item, and to perform such actions as calculate a recommendation score at the recipient client for example, if desired.

[0012] The recommendation rating increase may not be the same for each kind of action. It may be desired to weight an increase in the recommendation rating for a media item recommendation more heavily than an increase for play action of recommendations. The same may be true in vice versa. Further, the increase for a recommendation rating could be performed in a non-linear fashion, such as the use of a degrading scale wherein the recommendation rating increase amount becomes less as the number of recommendations and/or play actions for recommended media items increases. The degrading recommendation rating increase method may be performed for play actions for recommended media items, or media item recommendations to other users, or both. The recommendation rating increase could actually convert to a decrease if a recommendation for a media item is never played by a recipient user, or is not played within a certain period of time. Timestamps may be used to track the elapsed time between a recommendation for a media item and its play by a recipient user. A timeout may be used to automatically notify, without requiring a play action trigger, that a recommendation for a media item has not been played with a designated period of time. Alternatively, the elapsed time between plays for recommended media items could be analyzed only after receiving a recommendation or play action for a particular media item. The present invention is not limited to any particular type of increase or decrease system, methodology, or algorithm.

[0013] The recommendation rating system and method of the present invention may also be used in an application to assist a user in choosing media items. The recommendations ratings for media items may be accessed for any variety of applications. For example, the recommendation rating may be displayed to a user in a media application running on a user device. The user may select media items to be displayed according to their recommendation rating, with the highest rating appearing first and the lowest rating appearing last, or vice versa. In this manner, the user can make media item selections based on the recommendation rating of the item. This is one example of the possible applications of the present invention.

[0014] Those skilled in the art will appreciate the scope of the present invention and realize additional aspects thereof after reading the following detailed description of the preferred embodiments in association with the accompanying drawing figures.

[0029] The embodiments set forth below represent the necessary information to enable those skilled in the art to practice the invention and illustrate the best mode of practicing the invention. Upon reading the following description in light of the accompanying drawing figures, those skilled in the art will understand the concepts of the invention and will recognize applications of these concepts not particularly addressed herein. It should be understood that these concepts and applications fall within the scope of the disclosure and the accompanying claims.

[0030] The present invention is a media item recommendation rating system and method. A recommendation rating for a media item is established and dynamically updated in response to the media item being recommended to other users. A recommendation server or other device receives a report of a media item recommendation and updates a recommendation rating in response. The recommendation rating may also be updated based on how often media items are used or played in response to a received recommendation. Thus, a media item's recommendation rating is affected by events relating to its recommendation, as opposed simple play-based ratings that are updated on play actions regardless of whether related to a recommendation or not. Simple play ratings do not distinguish between ordinary usages or plays and those resulting from recommendations. Recommendation of a media item to another user is a better indicator of the user's likeability of a media item, since a recommendation is an endorsement of sorts. Many users may be interested or like a media item enough to use or play it, but not enough to recommend it. Uses or plays of media items short of a recommendation do not affect the recommendation rating as provided by the present invention.

[0031] A recommendation server may be provided that is operably connected to a network, such as the Internet for example. One or more user devices are coupled to the network. When a user device makes a recommendation for a media item to another user device, a message is sent to a recommendation server, which includes the originating recommender id, the destination, and the id of the media item being recommended. The recommendation server establishes a recommendation rating for the media item if a record of the media item does not already exist in recommendation tables stored in memory. Each unique media item recommended by a user has its own entry and recommendation rating in the recommendation tables. In this manner, and each time a media item is recommended to another user, the recommendation server dynamically updates the recommendation rating for the media item to affect, and in the typical case increase, its rating. After the recommendation rating is either established or updated, the recommendation server may forward the media item recommendation along with the recommendation rating to the recipient. The recipient client may contain a scoring engine that relies on the recommendation and/or other metrics to calculate a recommendation score for the media item, or to sort the recommendations according to their recommendation rating.

[0032] In this regard, FIG. 1 illustrates an exemplary media item recommendation system 10 for establishing and maintaining media item recommendation ratings according to the present invention. One or more user devices 12 are illustrated as networked computing devices employed by a user to use and/or play media items. FIG. 1 illustrates a plurality of user devices 12, namely user devices 12A, 12B, 12C, and up through including any number of user devices to 12N. Media items can be any type of media file or information, including but not limited to audio files, video files, songs, music, etc. The user devices 12 are what users or consumers control to provide recommendations of media items to other users as well as using or playing media items received as recommendations from other users. As discussed above and as discussed throughout the remainder of this application, both recommendations to others and usage or play of media items received as recommendations can affect the recommendation rating of a media item.

[0033] The user devices 12 typically contain a web browser and/or client application 14 to provide a graphical user interface to the user. The web browser or client application 14 can establish a connection over a network 16, such as the Internet, a TCP-IP based network, to a recommendation server 18. The recommendation server 18 can be any type of computing device that is coupled to the network 16 and is adapted to provide media related services to user devices 12. Examples of user devices 12 that may be used to access the recommendation server 18 and/or other users over the network 16 include, but are not limited to, personal computers, wireless portable media players (PMPs), personal digital assistants (PDAs), and the like.

[0034] The recommendation server 18 may be comprised of several components to facilitate user connections and provide media related services. These components may be comprised of a web component 20, a market component 22, and a server component 24. The server component 24 may also provide an application program interface (API) 26 to provide a downloadable client application to the user devices 12 for interfacing, ordering, and receiving media related services from the recommendation server 18 after an initial session is established between a user device 12 and the server 18 via the web browser 14, for example. A database 28 may be located on location at the recommendation server 18 for storage of data, including but not limited to media items and information relating to recommendation ratings for media items. A network accessible database 28′ may also be utilized to provide storage capabilities remote from the recommendation server 18 and/or as backup data storage, if desired.

[0035] Some media items requested or needing to be accessed may not be stored locally in the recommendation server's database 28 or network database 28′, but rather obtained from subscription services via network 16 access on-demand. In this regard, network accessible subscription services 32 may be provided for user devices 12 to access media items for downloading. The recommendation server 18 and/or the user devices 12 may access the subscription services 32 over the network 16.

[0036] Further, the recommendation server 18 and/or the user devices 12 can communicate over the network 16 with content identification systems 34 to provide identifying and other characteristic information about media items, such as information regarding artists, song titles, genre, etc. in the example of song content in particular. The allows access to user friendly information about media items as part of the services provided by the recommendation server 18 and/or the user devices 12. The identification systems 34 may be coupled with or contain content descriptors 36 that are compared to media items to identify and provide information regarding media items. For an example of songs in particular, the descriptors 36 may include known fingerprints and/or Global Unique IDs (GUIDs) to identify songs and obtain metadata containing information about the songs, such as title, artist, genre, etc. A uniform resource locator (URL) may be included to identify location(s) for the descriptor information.

[0037] Although the recommendation server 18 will be discussed in this application primarily in regard to establishing and updating media item recommendation ratings based on communication with user devices 12, it is noted the recommendation server 18 may also provide other information and services to the user devices 12. These additional services may include, but are not limited to, receiving requests for and establishing new user accounts, managing and storing user account preferences, storing information about the user's media collections, and managing the flow of recommendations for media items to users. In a preferred embodiment, the recommendation server 18 operates in a client-server relationship with users rather than a direct peer-to-peer connection. However, the present invention may be implemented in a peer-to-peer configuration where features of the recommendation server 18 are distributed among one or more peer nodes, super peer nodes, or other devices. Further, note that the recommendation server 18 may be implemented as a number of servers operating in a collaborative fashion.

[0038] The user devices 12 also typically contain an audio/video (A/V) player 38 that allows the user to use or play back any media item desired. Examples of A/V players 38 include but are not limited to Apple® itunes®, Apple® iPOD®, and the like. Media items selected for use and/or play include those stored locally at the user device 12 in a user's A/V collection 40, and/or any media item accessed from the recommendation server 18, the network database 28′, another user device, the subscription service(s) 32, and/or any other system accessible by or coupled to the network 16. Media items can also be recommended to other users in response to a user causing its user device 12 to send a recommendation for a media item over the network 16. The recommendation may go to the recommendation server 18 and then to the destination user, or the recommendation may be sent directly to the destination user without reaching the recommendation server 18. However, the recommendation server 18 will be provided information at a minimum that indicates the recommendation for the media item was made to update the recommendation rating for the media item, as will be discussed herein.

[0039] In regard to the present invention, the user devices 12 contain recommendation engines 42. A recommendation engine 42 is a program, algorithm, or control mechanism that handles sending and/or receiving media item recommendations over the network 16, as will be described in this application. The recommendation engine 42 may also be responsible to communicate with the recommendation server 18 to obtain actual recommendation ratings for received media item recommendations. The receiving user device 12 may use the recommendation rating to sort recommendations as controlled by a user, or to calculate a recommendation score using a scoring engine for the recommended media item. An example of a recommendation scoring system that does not use the recommendation rating as provided by the present invention as an input, but may easily be adapted to do so by one of ordinary skill in the art, is described in co-pending U.S. patent application Ser. No. 11/484,130 entitled “P2P NETWORK FOR PROVIDING REAL TIME MEDIA RECOMMENDATIONS,” which is incorporated herein by reference in its entirety.

[0040] In the system 10, the recommendation server 18 receives information when the user device 12, and in particular the recommendation engine 42, causes a media item recommendation to be sent to another user and/or when the user device 12 uses or plays a media item that has been recommended from another user. In response, the recommendation server 18 updates the recommendation rating for the media item in one or more recommendation tables 43 included in the database 28, so as to make available a dynamically updated media item recommendation rating to any of the user devices 12 and/or any device coupled to the network 16 that can access the recommendation server 18. The recommendation server 18 may also store media item recommendation ratings in one or more recommendation tables 43′ in the network accessible database 28′.

[0041] FIG. 2 illustrates two examples of recommendation paths for media items between different users to illustrate how recommendations for media items may be originated and forwarded. As illustrated, a first or originating user ‘A’, also called ‘node A’, forwards a media item, via a media item id 44, to another user ‘B’ 12B. A node is an originating node for a particular media item if the node makes a recommendation for a media item to another user that was not received as a recommendation from another user. A unique media item id 44 uniquely identifies the media item since typically, it is not desired to upload the actual media item from a user device 12 to the network 16 when a recommendation is made. This avoids the actual time and bandwidth necessary to upload a media item to the network 16 for generating a recommendation as well as any permission rights issues associated with forwarding media items. If the recipient user is authorized to play and/or use the recommended media item, as identified by the media item id 44, the recipient user accesses the media item associated with the media item id 44 via their local A/V collection 40, or through the network 16, such as through the subscription service(s) 32, the recommendation server 18, and/or another server, including by purchase, as examples.

[0042] Turning back to example 1 in FIG. 2, after user ‘B’ 12B receives the recommendation for media item id 44, node ‘B’ 12B can in turn use or play the media item associated with the media item 44. Node ‘B’ 12B can also make a recommendation for the media item id 44 to another user, which is node ‘C’ 12C in the illustrated example.

[0043] In example 2 in FIG. 2, node ‘B’ 12B is the originating node for a media item recommendation, identified by media item id 44′. This means that node ‘B’ 12B is sending a recommendation for a media item that it did not previously receive as a recommendation from another user. A node can be an originating node for one particular media item, and then a recipient node for another media item recommendation. Determination of an originating versus a recipient node is made on a per media item basis. In example 2, the media item recommendation, again identified by media item id 44′, is sent to node ‘C’ 12C, from node ‘C’ 12C to node ‘N’ 12N, and then to node ‘X’ 12X.

[0044] When a recommendation for a media item is made, the recommendation, via the media item id 44 in the preferred embodiment, is communicated to the recommendation server 18. The recommendation server 18 maintains the recommendation tables 43, 43′ in databases 28, 28′ for tracking originating nodes or users that made recommendations as well as establishing and maintaining an associated recommendation rating for media items. In this regard, examples of recommendation tables 43, 43′ stored in memory are illustrated in FIGS. 3A and 3B. In FIG. 3A, a “recommendation table” 46 is illustrated. The recommendation server 18 uses the recommendation table 46 to determine if a received recommendation for a media item has already been previously recommended. If so, a table entry for the media item and associated recommendation rating has already been established. If not, the recommendation server 18 can create a new entry for the reported recommended media item in the recommendation table 46 so a recommendation rating is established and can be updated on subsequently reported recommendations and play actions resulting from recommendations for the particular media item.

[0045] For example, the recommendation table 46 in FIG. 3A illustrates that song “867-5309,” represented by its media item id 44, was first recommended by user ‘A’ 12A as an originating user. The song “WHIP IT” is stored in the media item id 44 field of the recommendation table 46 as being first recommended by user ‘B’ 12B as the originating user for this media item. A recommendation id 47 is created within the recommendation table 46 when the recommendation is first made by an originating user. A recommendation rating 48 is established and maintained in the recommendation table 46 for the media item id 44. In the example of FIG. 3A, the recommendation rating 48 is a number rating, wherein a higher value represents a higher rating or higher recommendation value for the media item id 44. A number may be used for the recommendation rating 48 so that the recommendation server 18 can dynamically update the rating 48 by either increasing or decreasing the rating of a particular media item.

[0046] A relative recommendation rating level 50 may also be stored in the recommendation table 46 for a particular media item. For example, a rating level of ‘5’ may be the highest rating, and a rating level of ‘1’ may be the lowest. In this manner, the recommendation ratings 48 for each media item can be ranked relative to each other, rather than deciphering between a raw rating number, like used in the example of the recommendation rating 48 in FIG. 3A. Presenting a relative recommendation rating level 50 may be more easily interpreted by users, such as when recommendation ratings are presented on graphical user interfaces or displayed to users in some other manner, as will be described by example later in this application.

[0047] At a minimum, the purpose of the recommendation table 46 is to track media items that have been recommended by users so that a recommendation rating for the media item can be established and dynamically updated. If a media item is not recommended, a recommendation rating is not established in the recommendation table 46. The recommendation table 46 in the example of FIG. 3A has at a minimum, the originating user for each media item recommended, as well as a recommendation rating 48.

[0048] FIG. 3B illustrates a “received recommendations table” 52 that may also be stored in the recommendation tables 43, 43′ and dynamically updated by the recommendation server 18 for use in maintaining and updating recommendation ratings 48 for media items. The received recommendations table 52 is one form of a recommendation path table. The purpose of a recommendation path table, and the received recommendations table 52 as one form of a recommendation path table, is to track the recommendation paths for media items, such as recommendation paths illustrated in the recommendation path examples of FIG. 2 discussed above. This allows the recommendation server 18 to determine if a particular media item was received by a user as a result of a recommendation from another user. If so, and the recipient uses or plays the media item, the recommendation server 18 updates the recommendation rating 48 for the media item. This is because the media item used or played by the recipient user was received as a recommendation from another user. This is how the recommendation server 18 distinguishes play actions not relating to recommendations, where the recommendation rating 48 should not be updated.

[0049] For example, the received recommendations table 52 in FIG. 3B illustrates that recommendation id (47) #24 (song “867-5309”) was received as a recommendation from user ‘B’ 12B in the “receiver id” field 54 by user ‘A’ 12A in the “recommender id” field 56. Recommendation id #24 (song “867-5309”) was then recommended by user ‘B’ 12B to user ‘C’ 12C, as shown in the received recommendations table 52. This is the recommendation flow path illustrated previously in example 1 of FIG. 2. The recommendation flow path of example 2 of FIG. 2 is also shown in the received recommendations table, where recommendation id #108 was sent as a recommendation by user ‘B’ 12B to user ‘C’ 12C, user ‘C’ 12C to user ‘N’ 12N, and user ‘N’ 12N to user ‘X’ 12X.

[0050] FIG. 4 is a flow chart that illustrates an embodiment of how the recommendation server 18 receives a reported recommendation from a user device 12 and how the recommendation tables 43, 43′ and recommendation rating 48 for the recommended media item is established and updated. The process starts (100) when the recommendation server 18 receives a report of a media item recommendation, in the form of a media item id 44, and recommendation path or recipient from a user device 12 (step 102). The recommendation server 18 determines if a record for the particular media item id 44 received exists in the recommendation table 46 for an originating user 12 (decision 104). If not, this means that the particular media item has not been previously recommended, and the user 12 that reported the media item to the recommendation server 18 is an originating user or node. As a result, the recommendation server 18 creates a new record for the originating path for the media item id 44 in the recommendations table 52, as shown in FIG. 3A (step 106). Since the recommended media item is newly recommended and a recommendation rating 48 was not previously established for the media item, an initial recommendation rating 48 is created for the media item id 44 in the recommendations table 46 as well (step 108).

[0051] Next, the recommendation rating 48 for the media item id 44 is increased in the recommendation table 46. This is because the recommendation rules in one embodiment call for increasing the recommendation rating 48 every time a recommendation is sent by a user (step 110). In an alternative embodiment not illustrated, the recommendation server 18 may only increase the recommendation rating 48 for a media item if the recommendation is destined for a new recipient user who has not previously received the recommendation for the media item. The recommendation server 18 could determine this by checking the received recommendations table 52 to determine if the same recommendation path has been previously performed for a given media item.

[0052] The recommendation server 18 then determines if a record for the media item id 44 exists in the received recommendations table 52 for the originating user 12 (decision 112). This decision is made to determine if the path between the originating user 12 and the recipient user for the media item recommendation was previously made. If not, a new record is created in the received recommendations table 52 for media item id 44, since the recommendation is new for the media item id 44 (step 114). Thus, a record must be created in the received recommendations table 52. If a record for the media item id 44 recommended already exists in the recommendation table 52, a new record is not created since one already exists. Thereafter, the receiving or recipient user that is to receive the media item id 44 recommendation is added to the received recommendations table 52 to record the recommendation path (step 116), and the process ends (step 118). The recommendation server 18 needs to know if a particular recommendation path has been previously performed, because only unique recommendation paths trigger an increase in the recommendations rating 48 for play actions in the preferred embodiment, as discussed below for the flow chart in FIG. 5.

[0053] Thus in summary, the flow chart of FIG. 4 illustrates that when a user makes a media item recommendation, the recommendation, via the media item id 44, is reported to the recommendation server 18. Because of the recommendation being made, the recommendation rating 48 for the media item is increased. The recommendation rating 48 may be increased for a given media item id 44 each time it is recommended regardless of the recommendation path, or only for unique recommendation paths. If a record has not been previously established in the recommendation table 46 for the media item id 44 and for the recommendation path for the media item id 44 in the received recommendations table 52, new records are established accordingly. In this manner, the recommendation server 18 can look up information in the recommendation table 46 and the received recommendations table 52. This information is used to dynamically update the recommendation rating 48 for media items based on play actions that result from recommendations as opposed to play actions not related to recommendations. The recommendation server 18 may, in addition to forwarding the media item id 44, also automatically forward the newly established or updated recommendation rating 48 for the media item id 44 to the recipient. Alternatively or in addition, clients or user devices 12 can also, upon their own initiative, send a request over the network 16 to the recommendation server 18 to specifically obtain the recommendation rating 48 for media items. The recommendation rating 48 may be used for any number of reasons, including but not limited to allowing a user to sort through received recommendations by recommendation rating 48, or to calculate a media item recommendation score with the recommendation rating 48 provided as an input to such calculation, for received recommendations as examples.

[0054] FIG. 5 illustrates a flow chart showing the operation of the recommendation server 18 when a play action message is received from a user or user device 12. Because play actions can affect the recommendation rating if the play action is in relation to a recommendation, the recommendation server 18 evaluates play actions for media items. The process starts (step 120), and the recommendation server 18 receives a message for a play or usage action of a media item from a user device 12, with the media item identified by its media item id 44 (step 122). The recommendation server 18 determines if the user has received the media item being used or played as a recommendation by checking the received recommendations table 52 (decision 124). Because the received recommendations table 52 is updated with users that received recommendations in response to a recommendation made by another user, the recommendation server 18 can use information in this table 52 to trace back whether the media item at issue was recommended to the user. If not, the message received by the recommendation server 18 is simply a play action not related to a recommendation, in which case the process ends (step 128). If the recommendation server 18 determines the media item being used or played was recommended previously to the user, the recommendation rating 48 is increased in the recommendation table 46 for the media item (step 126). This is because the play action resulted from a recommendation.

[0055] It should be recognized that although the recommendation rating 48 of media items is discussed above as being increased in response to recommendations for media items (e.g. FIG. 4) and play actions for recommended media items (e.g. FIG. 5), the increase may not be the same for each action. It may be desired to weight an increase in the recommendation rating 48 for a media item recommendation more heavily than an increase in the recommendation rating 48 for play action. The same is true vice versa. Further, the increase for a recommendation rating 48 could be performed in a non-linear fashion, such as by use of a degrading scale, where the recommendation ratings 48 increase amount becomes less as the number of recommendations and/or play actions for recommended media items increases. The degrading recommendation rating 48 increase may be performed for play actions for media items recommended, or media item recommendations to other users, or both. The recommendation rating 48 increase could actually convert to a decrease if a recommendation for a media item is never played by a recipient user, or is not played within a certain period of time. The recommendation tables 43, 43′ could include timestamps to track the elapsed time between a recommendation for a media item and its play by a recipient user. A timeout may be used wherein the recommendation server 18 is automatically notified without a play action that a recommendation for a media item has not been played within a designated period of time, or the recommendation server 18 could analyze elapsed time between plays for recommended media items only after receiving a recommendation or play action for a particular media item. It should be noted that the present invention is not limited to any particular type of increase or decrease system, methodology, or algorithm.

[0056] FIG. 6 illustrates a communication flow diagram of how media item recommendations may be tracked and recommendation ratings established and dynamically updated in accordance with one embodiment of the present invention. FIG. 6 shows the interaction between the user devices 12 and the recommendation server 18. The actions taken internally by the recommendation server 18, including establishing and updating the recommendation rating 48, are shown generally in the flow diagram and relate back to the flow chart operations described above with respect to FIGS. 4 and 5.

[0057] The exemplary flow diagram in FIG. 6 shows media item recommendations being communicated between the exemplary three nodes or user devices 12 illustrated in the recommendation path example 1 of FIG. 2; namely user device ‘A’ 12A, user device ‘B’ 12B, user device ‘C’ 12C. FIG. 6 illustrates the flow communication path between these nodes and the recommendation server 18 to dynamically maintain a recommendation rating 48 for media items.

[0058] Turning to FIG. 6, the user device ‘A’ 12A may choose to use its A/V player 38A to play a particular media item, according to the media item id 44, which was not received as a recommendation from another user (step 130). In this instance, since the media item was not played or received as a result of a recommendation, the recommendation rating for this media item will not be affected. However, the play action by user device ‘A’ 12A initiates a play action message identifying the media item id 44 played to the recommendation server 18 (step 132). This is so the recommendation server 18 can determine if the recommendation rating for the media item id 44 should be updated based on user device ‘A's 12A play. Because the play action was not from a recommendation to user ‘A’ 12A, the recommendation server 18 does not update the recommendation rating for the associated media item id 44. In this manner, as previously described above, simply play actions not relating to a recommendation do not affect a recommendation rating of a media item.

[0059] If user device ‘A’ 12A determines that the media item should be recommended to another user, such as user device ‘B’ 12B for example as illustrated in FIG. 2, the user device ‘A’ 12A communicates the media item id 44 associated with the recommendation to the recommendation server 18 (step 134). The recommendation server 18 establishes and/or stores the recommendation path from user device ‘A’ 12A to user device ‘B’ 12B for the media item id 44 in the received recommendations table 52, as previously provided in FIG. 4 (step 136). The recommendation server 18 then updates the recommendation rating 48 for the media item in the recommendation table 46 based on the media item being recommended (step 138). If the recommendation server 18 is also responsible for forwarding the recommendation for the media item to the destination node, which is user device ‘B’ 12B in this example, the recommendation server 18 will communicate the media item id 44 to the recipient user (step 140).

[0060] In addition to the media item id 44, the recommendation server 18 may also automatically communicate the recommendation rating 48 to the recipient user as previously noted. The recipient user and/or its client application 14 may use the recommendation rating 48 directly to sort recommendations, or the recommendation rating 48 may be used as an input to calculate a recommendation score, which may be used for a variety of applications and purposes. Automatic communication of the recommendation rating 48 when forwarding a recommended media item id 44 also avoids the recipient user from having to send a separate request to the recommendation server 18 for the recommendation rating 48, although the recipient user may do so with such being contemplated by the present invention. Also, note that the recommendation server 18 may only be responsible for creating and maintaining the recommendation rating 48 without also being responsible for forwarding a recommendation to a recipient.

[0061] The recipient, which is user device ‘B’ 12B in the example, may then decide to use or play the received recommended media item (step 142). In this instance, this action is indicative of the popularity of the media item since the recipient user was willing to use or play it based on a recommendation. In response, the user device ‘B’ 12B sends a play action message for the media item id 44 to the recommendation server 18 to report the usage or play (step 144), which is illustrated as being received by the recommendation server 18 in the flowchart of FIG. 5 as an example. The recommendation server 18 looks up the recommendation path for the media item in the received recommendations table 52 to determine if the media item was previously recommended (step 146). If so, the recommendation rating 48 for the media item is increased, as illustrated in the flow chart of FIG. 5 as an example (step 148). If the media item reported as being used or played by user device ‘B’ 12B to the recommendation server 18 was not previously recommended to user ‘B’ 12B, the recommendation rating 48 would not be increased.

[0062] From there, user device ‘B’ 12B may forward a recommendation for the same media item to yet another user, which is user device ‘C’ 12B in the illustrated example (step 150). Again, the recommendation rating 48 for the media item will be increased by the recommendation server 18 (step 152), because of the additional recommendation of the media item to another user. Again, if the recommendation server 18 is responsible for also forwarding the recommendation for the media item to the user device ‘C’ 12C, the recommendation server 18 will perform that action (step 154). The user device ‘C’ 12C may decide to use or play the media item received from the recommendation, just as was the case for user device ‘B’ 12B (step 156). Again, the play action results in a message being communicated to the recommendation server 18 (step 158), to determine if usage or play of the media item was related to a received recommendation (step 160) and to increase the recommendation rating 48 if so (step 162).

[0063] In an alternative embodiment, the recommendation tables 43, 43′ may also include a “sent recommendations table” 170 as illustrated in FIG. 7. The purpose of the sent recommendations table 170 is similar to that of the received recommendations table 52 in that the table 170 is used to determine the recommendation paths for a media item. The sent recommendations table 170 is used to determine if a play or use action by a user is related to a previous recommendation received for the media item. If so, the recommendation rating 48 is increased. The table 170 contains the same information as the received recommendations table 52, including a recommendation id 172 for the recommended media item, but with the remaining fields in reverse order. Namely, the recommender id 174 is provided first, and the receiver id 176 provided second. In this manner, the recommendation server 18 may more efficiently search for recommendation paths when performing recommendation rating related tasks by searching based on the recommender id 174 rather than the receiver id 176, including those described previously.

[0064] FIG. 8 is a flow chart that illustrates another embodiment of how the recommendation server 18 receives a reported recommendation from a user device 12B and how the recommendation tables 43, 43′ and recommendation rating 48 for the recommended media item are established and updated. The flow chart describes a similar, but alternative embodiment to the flow chart of FIG. 4. In FIG. 8, the sent recommendations table 170 illustrated in FIG. 7 is additionally updated and searched to determine if the recommendation rating 48 of a media item is increased. Steps 200-216 of FIG. 8 are essentially the same steps as steps 100-116 in FIG. 4, and thus the description of steps 100-116 for FIG. 4 above are applicable and not repeated here. However, the recommendation server 18, as illustrated in FIG. 8, additionally maintains a sent recommendations table 218 as well.

[0065] As illustrated in FIG. 8, after a media item recommendation is added to the received recommendations table 52 in step 218, the recommendation server 18 determines if a record for the media item also exits in the sent recommendations table 170. If not, a new record is created so that the sent recommendations table 170 has a record of sent recommendations primarily by recommender as opposed to primarily by receiver, as provided in the received recommendations table 52 (step 220). The process then ends 222. Again, the recommendation server 18, by maintaining both a received recommendations table 52 and a received recommendations table 170, may provide for more efficient searching when play or use action messages are received to determine if the play or usage of a media item is from a previous recommendation for updating the recommendation rating 48 of a particular media item.

[0066] FIG. 9 is a flow chart that illustrates another embodiment of how the recommendation server 18 updates a recommendation rating 48 in response to a use or play action of a media item that was previously recommended. The received recommendations table 52 and the sent recommendations table 170 are both checked. The flow chart describes a similar, but alternative embodiment to the flow chart of FIG. 5 that only checks the received recommendations table 52 to determine if the recommendation rating 48 of a media item is increased. Steps 230-236 of FIG. 9 are essentially the same as steps 120-126 of FIG. 5, and thus the description for steps 120-126 is applicable for steps 230-236 in FIG. 9.

[0067] As illustrated in FIG. 9, after the recommendation server 18 has checked the received recommendations table 52 to determine if the play or use action for the media item was previously recommended (decision 234), the recommendation server 18 determines the same thing by checking the sent recommendations table 170 as well or in the alternative (decision 238). If so, the recommendation rating 48 for the media item is increased (step 240), and the process ends (step 240). In this regard, if it is more efficient to check the sent recommendations table 170 as opposed or in addition to the received recommendations table 52, the recommendation server 18 can do so to determine if the recommendation rating 48 for a media item should be increased.

[0068] FIG. 10 illustrates an example of a populated media collection that may be presented by the client application 14 on the user device 12 to a user. FIG. 10 illustrates one example application of how the recommendation rating system 10 of the present invention may be used in an application to assist the user in viewing and/or choosing media items. The illustrated example is for songs, but such is also applicable for any type of media items including but not limited to videos or presentations. As illustrated, a main media collection page or window 300 is provided. The current content being played by the user is shown in a currently playing box 302. User controls 304 are provided to allow the user to control the A/V player 38 on the user device 12, such as reversing, forwarding, and playing selected items. A volume slide control 306 is also provided to allow the user to control the volume of the A/V player 38.

[0069] Media items that are either populated from a server, which may or may not be the recommendation server 18, or from the user's A/V collection 40, or both and sent to the client application 14 are listed in a music item table 308. The music item table 308 consists of numerous columns that provide key information about the music items to the user. A title column 312 provides the title of the music item. A time column 314 provides the length or duration of the music item when played. An artist column 316 provides the artist of the music item. An album column 318 provides the name of the album the song was included on. A score column 320 provides a score in terms of importance to the user according to their user preferences stored as a result of the registration process and/or determined by a server based on the owned media items in the user's A/V collection 40. A genre column 322, a year column 324, a received column 326, and a played column 328 provides the genre type, the year of release, the date and time of receipt, and the date and time of the last play of the music item. A scroll bar 330 is available for the user to scroll up and down through their song items.

[0070] A recommender column 311 contains a user id or name of the person or friend that recommended the content to the user. The user can establish a list of friends that are able to provide recommendations as disclosed in U.S. patent application Ser. No. 11/484,130 entitled “P2P NETWORK FOR PROVIDING REAL TIME MEDIA RECOMMENDATIONS” previously referenced above. The user is able to navigate the media item table 308 by “point and click” to play items, download items, and request content from a server and/or other sources.

[0071] The recommendation rating column 310 is provided to show the user the recommendation rating 48 of the music items in their music item table 308. Alternatively and not shown, a recommendation score may be generated by the client application 14 based on the recommendation rating 48 and displayed in the recommendation rating column 310, and/or used by the client application 14 and/or the media collection window 300. In the example of FIG. 10, the user has selected the music items to be displayed according to their recommendation rating 48, with the highest rating appearing first and the lowest rating appearing last, except for music items not having a recommendation rating 48. This is indicated by a blank field being located beside these media items, which are shown at 332 for the bottom two media items listed in the music item table 308. In this manner, the user can make media item selections, namely music items, based on the recommendation rating 48 of the item. The user device 12 would communicate with the recommendation server 18 over the network 16 to obtain the recommendation ratings 48 for each of the media items accessible by and/or selected by the user. This is one example of the possible applications of the present invention.

[0072] An iteration of the media collection page 300 is illustrated in FIG. 11, where the relative recommendation rating level 50 is displayed to the user rather than a raw recommendation rating 48. This may prevent constant reordering or churn of the order of the media items displayed in the music item table 300 in the event that the client application 14 updates the recommendation rating for the music items displayed on a continuous or periodic basis. The relative recommendation rating level 50, as provided in the recommendation table 46 illustrated in FIG. 3A and previously described above, may be easier for a user to view and peruse. In addition, since the present invention dynamically updates the recommendation rating 48, the rating numbers 48, as illustrated in FIG. 10, may become too large in digit size to be easily displayed and/or displayed in the allotted real estate in the recommendation rating column 310. Further, recommendation ratings 48 may continuously be in flux or changing; however, the relative recommendation rating level 50 may not be altered as often since a relative recommendation rating level 50 is less susceptible to change in a given period of time. Thus, a relative recommendation rating level 50 may cause less constant reordering of the music items displayed in the media collection page 300.

[0073] In the example of FIG. 11, a star rating system is used to display the recommendation rating in the recommendation rating column 310. In this example, a five-star recommendation rating is the highest, and a one-star rating is the lowest. The recommendation server 18 and/or the client application 14 may convert either the recommendation rating 48 or the relative recommendation rating level 50 into a visually pleasing ranking. For example, if the relative recommendation rating level 50 is a number from one to five, this number can be converted into a scalable visual item same in number, such as but not including a star, circle, diamond, or any other visual indicator desired. A number may also be used for displaying the relative recommendation rating level 50.

[0074] FIG. 12 illustrates a flow chart wherein the recommendation server 18 determines a relative recommendation rating level 50 for media items stored in the recommendation table 42. This process may be used by client applications 14 to display and rank media items by recommendation rating 48 as opposed to raw scores or numbers. Also note that this process may also be implemented on the client application 14 as opposed to the recommendation server 18. The process starts (step 340), and a determination is made as to the number of unique media item ids 44 stored in the recommendation table 46 (step 342). The number of entries is divided by the number of levels desired for the relative recommendation rating level 50 to determine the cut off levels for assigning each media item id 44 to its relative level as compared to the other media item ids 44.

[0075] For example, if there are 10,000 unique media item ids 44 in the recommendation table 46, and there are five relative recommendation levels desired, the cutoffs for each level would be 2,000. The 2,000 media items having a recommendation rating 48 in the lowest of all media item ids 44 would be assigned the lowest relative rating, for instance ‘1’. The 2,000 media items having a recommendation rating 48 in the highest of all media item ids 44 would be assigned the highest relative rating, for instance ‘5’. After the cutoff or levels are determined, the relative recommendation rating levels 50 for the media items ids 44 in the recommendation table 46 are assigned or overwritten with the correct relative rating. Thus, the relative recommendation rating level 50 for a media item can change dynamically and over time depending on the success of other media items having their recommendation ratings 48 increase. What may typically happen is that a new media item release will start with a low recommendation rating 48, move higher and possibly to the highest level over time if recommended heavily, and then over time move back down in levels after a period of time when other media items overtake it, especially if the recommendation rating 48 is decreased or penalized in rating as the media item becomes less recommended over time. Also note that any of the aforementioned features provided for the recommendation rating 48 may also be used to provide an input into a scoring engine for calculating a recommendation score for media items.

[0076] FIG. 13 illustrates an alternative media item recommendation system 10 for establishing and maintaining a media item recommendation rating that does not use a recommendation server 18. The present invention is not limited to a recommendation server 18 establishing and maintaining recommendation ratings for media items. In FIG. 13, a peer system 10′ is illustrated. In this system 10′, a plurality of user devices 12′ act a peer devices communicating with each other in a peer-to-peer manner to forward media item recommendations, similar or the same as the system described in previously referenced co-pending U.S. patent application Ser. No. 11/484,130 entitled “P2P NETWORK FOR PROVIDING REAL TIME MEDIA RECOMMENDATIONS.”. The peer user devices 12′ are coupled to the network 16 to communicate with each other. One user device 12′ (user device ‘B’ in the illustrated example) may be deemed a “super” peer device 12B′, meaning that it has increased responsibilities for establishing and maintaining the recommendation tables 46 and the recommendation ratings 48 for the media items recommended and played as a result of recommendations between peer user devices 12′. The super peer device 12B′ may not contain an A/V player 38B′ since it may be a more dedicated machine not intended to be used for using or playing media items.

[0077] The super peer device 12B′ may include a recommendation engine 42B′ that can perform some or all of the same aforementioned functionalities of recommendation server 18. The super peer device 12B′ may include a database 28″ for storing the recommendation tables 43″, or a network recommendation database 28′″ separate from the super peer device 12B′ may be provided on the network 16 to provide storage for the recommendation tables 43′″. The super peer device 12B′ may send and receive recommendations, and/or their associated recommendation ratings 48 automatically, to and from other peer user devices 12′ in addition to receiving reports regarding media item recommendations. Or, the other user devices 12′ may send the actual media item recommendations directly to other users 12′ in a peer-to-peer communication with a separate reporting message only being communicated to the super peer device 12B′ to establish and update the recommendation ratings 48 for media items. The peer user devices 12′ may access the recommendations tables 43″, 43′″ via communications to the recommendation server 18 or the super peer device 12B′ to access the recommendation ratings 48 for media items for any variety of applications, including those illustrated as examples in the media collection page 300 illustrated in FIGS. 11 and 12.

[0078] Those skilled in the art will recognize improvements and modifications to the preferred embodiments of the present invention. All such improvements and modifications are considered within the scope of the concepts disclosed herein and the claims that follow.