Systems and Methods for Predicting Audience Measurements of a Television Program

Abstract

Described herein are apparatuses, systems and methods for predicting audience measurements of a television program. A method comprises inputting a target program for acquisition into a prediction model, wherein the prediction model is based on a plurality of television acquisition performance predictors, and generating a recommendation as to whether the target program should be acquired based on the prediction model and the plurality of television acquisition performance predictors.

1 - 20 . (canceled) 
     
     
         21 . A method, comprising:
 at a predictive modeling server:   identifying each of a plurality of programs;   identifying each of a plurality of networks, each of the networks having aired a subset of the programs;   receiving a network ratings value corresponding to a broadcast of one of the programs on one of the networks;   identifying one of the programs that has not aired on a selected one of the networks; and   determining a predicted network ratings value for the identified program on the selected network by collaboratively filtering the network ratings values.   
     
     
         22 . The method of  claim 21 , wherein the network ratings values are collaboratively filtered using a model-based collaborative filtering employing matrix factorization. 
     
     
         23 . The method of  claim 22 , wherein the model-based collaborative filtering identifies a latent network vector and a latent program vector for the identified program. 
     
     
         24 . The method of  claim 23 , wherein the predicted network ratings value is further determined based on the latent network vector and the latent program vector for the identified program. 
     
     
         25 . The method of  claim 21 , further comprising:
 quantifying a comparison between the selected network and an aired network that aired the identified program based on a network-level variable.   
     
     
         26 . The method of  claim 25 , wherein the network-level variable is an audience size for the selected network and the aired network for one of a given time or a given time period. 
     
     
         27 . The method of  claim 21 , wherein, when the identified program was aired by more than one of the networks, the collaborative filtering uses the networks ratings values for each of the networks that aired the identified program. 
     
     
         28 . The method of  claim 27 , wherein the networks ratings values for each of the networks that aired the identified program includes a variable indicative of a time at which the network aired the identified program. 
     
     
         29 . A non-transitory computer readable storage medium with an executable program stored thereon, wherein the program instructs a processor to perform actions that include:
 identifying each of a plurality of programs;   identifying each of a plurality of networks, each of the networks having aired a subset of the programs;   receiving a network ratings value corresponding to a broadcast of one of the programs on one of the networks;   identifying one of the programs that has not aired on a selected one of the networks; and   determining a predicted network ratings value for the identified program on the selected network by collaboratively filtering the network ratings values.   
     
     
         30 . The non-transitory computer readable storage medium of  claim 29 , wherein the network ratings value is calculated Nielsen ratings. 
     
     
         31 . The non-transitory computer readable storage medium of  claim 29 , wherein the network ratings values are collaboratively filtered using a model-based collaborative filtering employing matrix factorization. 
     
     
         32 . The non-transitory computer readable storage medium of  claim 31 , wherein the model-based collaborative filtering identifies a latent network vector and a latent program vector for the identified program. 
     
     
         33 . The non-transitory computer readable storage medium of  claim 32 , wherein the predicted network ratings value is further determined based on the latent network vector and the latent program vector for the identified program. 
     
     
         34 . The non-transitory computer readable storage medium of  claim 29 , wherein the actions further include:
 quantifying a comparison between the selected network and an aired network that aired the identified program based on a network-level variable.   
     
     
         35 . The non-transitory computer readable storage medium of  claim 34 , wherein the network-level variable is an audience size for the selected network and the aired network for one of a given time or a given time period. 
     
     
         36 . The non-transitory computer readable storage medium of  claim 29 , wherein, when the identified program was aired by more than one of the networks, the collaborative filtering uses the networks ratings values for each of the networks that aired the identified program. 
     
     
         37 . The non-transitory computer readable storage medium of  claim 36 , wherein the networks ratings values for each of the networks that aired the identified program includes a variable indicative of a time at which the network aired the identified program. 
     
     
         38 . A system, comprising:
 a memory storing a plurality of rules; and   a processor coupled to the memory and configured to perform actions that include:
 identifying each of a plurality of programs; 
 identifying each of a plurality of networks, each of the networks having aired a subset of the programs; 
 receiving a network ratings value corresponding to a broadcast of one of the programs on one of the networks; 
 identifying one of the programs that has not aired on a selected one of the networks; and 
 determining a predicted network ratings value for the identified program on the selected network by collaboratively filtering the network ratings values.