Systems and methods are described herein for providing content item recommendations based on a video. Using feature vectors corresponding to at least one frame of a video (e.g., generated based on texture and shape intensity of a frame), a recommendation system improves content recommendation using analytic and quantitative characteristics derived from a frame of a content item rather than merely manually labeled bibliographic data (e.g., a genre or producer). The recommendation system may generate a feature vector based on a texture, a shape intensity (e.g., generated from a Generalized Hough Transform), and temporal data corresponding to at least one frame of a video. The feature vector is analyzed using a machine learning model (e.g., a neural network) to produce a machine learning model output. The recommendation system causes a recommended content item to be provided based on the machine learning model output.

Methods, apparatus, systems and articles of manufacture to assign viewers to media meter data are disclosed. An apparatus includes processor circuitry to execute computer readable instructions to at least: identify a candidate household from a plurality of second households to associate with a first household based on an analysis of a first duration of time first media was presented by a first media presentation device and a second duration of time second media was presented by second media presentation devices; match different ones of first panelists of the first household with matching ones of second panelists of the candidate household; and impute respective portions of the first duration of time to the different ones of the first panelists based on portions of the second duration of time for which the matching ones of the second panelists of the candidate household were exposed to the second media.

An example system comprising at least one memory, programmable circuitry, and instructions to cause the programmable circuitry to obtain first time information corresponding to one or more first times that a panelist accessed first media via a television, obtain second time information corresponding to one or more second times that a household accessed second media via an over-the-top device, the second time information matching the first time information and associated with corrected demographic information of members of the household, determine that a first genre of the first media matches a second genre of the second media, determine that a member of the household accessed the second media via the over-the-top device based on historical media access events of the panelist, and attribute the access of the second media to the member of the household to create a corrected demographic impression record.

Systems and methods are described herein for providing content item recommendations based on a video. Using feature vectors corresponding to at least one frame of a video (e.g., generated based on texture and shape intensity of a frame), a recommendation system improves content recommendation using analytic and quantitative characteristics derived from a frame of a content item rather than merely manually labeled bibliographic data (e.g., a genre or producer). The recommendation system may generate a feature vector based on a texture, a shape intensity (e.g., generated from a Generalized Hough Transform), and temporal data corresponding to at least one frame of a video. The feature vector is analyzed using a machine learning model (e.g., a neural network) to produce a machine learning model output. The recommendation system causes a recommended content item to be provided based on the machine learning model output.

Systems and methods for providing fast forwarding recommendations based on the user's consumption history are disclosed. The consumption history includes data relating to attributes that were previously rewinded and watched and those that were skipped and forwarded. It also includes scores for attributes that were present and absent in a portion that was previously rewinded or forwarded. A score is assigned to the attributes and used for determining a consumption pattern. If the consumption pattern indicates that the user previously rewinded and watched the attribute, then a recommendation not to skip an upcoming portion that includes the attribute is provided. A graphical timeline that depicts the amount of time saved by skipping the portion of the media asset with the attribute is also provided.

An example to predict demographics for impressions includes a prediction manager to: determine that first demographic data corresponding to a first database proprietor subscriber does not match second demographic data corresponding to a media delivery device, both the first and second demographic data corresponding to an impression; obtain third demographic data corresponding to an Internet protocol address, the third demographic data obtained from a second database proprietor; and generate matched demographic data based on comparing the third demographic data to the first demographic data; and a modeler to generate a prediction model based on the matched demographic data, the prediction model to predict fourth demographic data for the impression.

An intelligent return path data (iRPD) system enables transmission of return path data via a communication network for a television connected to a non-hybrid set top box (STB). The iRPD system is configured to receive the key codes of the keys pressed on a remote control device along with the date time stamps and the location information. The iRPD system analyzes the keypress data along with the date time stamps to recognize the channels accessed in programming operations and the non-programming control operations executed by a viewer operating the remote control device. The viewer's behavior pattern is thus recorded and analyzed to identify the viewer. Upon identifying the viewer, various functions such as collecting the viewership statistics, implementing metered usage billing or ecommerce activities are enabled.

Systems and methods are described herein for curating content that follows a narrative structure. A narrative structure comprises narrative portions that have a defined order. Signature analysis of known content that follows the narrative structure is used to train machine learning models for the narrative structure and the narrative portions that make up the narrative structure. Signature analysis of candidate content segments, along with machine learning models for the narrative portions, are used to identify candidate content segments that match the respective narrative portions. A candidate playlist is generated of the identified candidate content segments in the defined order. In one embodiment, the machine learning model for the narrative structure is used to validate the generated playlist.

A mobile telephone or tablet or the like (40) for viewing commercial television programmes has a display screen (41), a firewall (70), and behind the firewall a first, sealed data storage section (42) which stores personal attributes of a viewer, such as a child, supplied on a token (32) from a trusted third party (10), a second, unsealed data storage section (44) to which the personal attributes are copied, the second section having a flag (48) which is set when the data in the second section has been modified by the viewer, and a third data storage section (46) in which personal preferences can be stored by the viewer. Adverts from advertisers (50) targeted on the data in the second and third sections are sent to the device (40), headers in the adverts are checked against the data in the first section, and adverts are rejected if they are unsuitable for the viewer. Unrejected adverts are shown on the display screen, and the advertiser is informed of their showing, but the identity of the viewer is not disclosed.

Techniques for modifying training data for video response quality optimization are provided. In one technique, training data is identified that is generated based on video presentation data that indicates multiple video items were presented to multiple entities. The training data comprises multiple training instances, each indicating a presentation of at least a portion of a video item to an entity. For each training instance in a subset of the training instances, a quality metric of the presentation of the video item indicated in said each training instance is computed and that training instance is modified based on the quality metric. After modifying one or more of the training instances, the model is trained using one or more machine learning techniques. In response to a content request, the model is used to determine whether to transmit a particular video item over a network to a computing device of a particular entity.