Methods, systems, and apparatuses for determining viewership
Methods, systems, and apparatuses for determining viewership of a content item are described herein. Machine learning techniques may be used to determine which user(s) among a user group at a multi-user location is consuming a content item. A machine learning model may be trained using demographic attributes and content attributes associated with a plurality of single-user locations. A probability engine may train a machine learning model using the demographic attributes and content attributes and one or more machine learning algorithms. The trained machine learning model may be used to determine which user(s) among at least two users is consuming a content item at a multi-user location at which multiple people reside.
1. A method comprising:
receiving, by a computing device, a plurality of demographic attributes associated with a plurality of user locations;
receiving content metadata associated with a plurality of content items presented at the plurality of user locations, wherein the content metadata comprises a plurality of content attributes associated with the plurality of content items;
determining, based on the plurality of demographic attributes associated with the plurality of user locations and the content metadata, a training dataset comprising the plurality of demographic attributes associated with the plurality of user locations and the plurality of content attributes;
generating, based on the training dataset, a distribution of probability scores for one or more demographic attributes of the plurality of demographic attributes associated with the plurality of user locations with respect to one or more values of one or more content attributes of the plurality of content attributes;
training, based on the distribution of probability scores, a machine learning model;
determining, by the trained machine learning model and based on a plurality of demographic attributes associated with at least two users at a user location and content metadata for a content item output at the user location, an initial viewing probability score for each of the at least two users; and
adjusting, based on a plurality of network data associated with the at least two users, the initial viewing probability score for each of the at least two users.
2. The method of claim 1 , wherein each of the plurality of user locations is associated with a single user.
3. The method of claim 1 , wherein the machine learning model is a neural network comprising a plurality of neurons.
4. The method of claim 3 , wherein each neuron is associated with one content attribute of the plurality of content attributes and one demographic attribute of the plurality of demographic attributes associated with the plurality of user locations.
5. The method of claim 1 , wherein the plurality of demographic attributes associated with the plurality of user locations comprises one or more of age or gender, and wherein the plurality of content attributes comprises one or more of a genre, a content type, a production year, a review score, or a duration.
6. The method of claim 1 , further comprising:
adjusting, based the user location, the initial viewing probability score for each of the at least two users.
7. The method of claim 1 , further comprising:
receiving, by the computing device, validation data associated with the at least two users and the user location; and
retraining the trained machine learning model based on the validation data.
8. A method comprising:
receiving, by a computing device, a plurality of demographic attributes associated with at least two users at a user location;
receiving content metadata associated with a content item being presented at the user location, wherein the content metadata comprises a plurality of content attributes;
determining, by a trained machine learning model based on the plurality of demographic attributes and the plurality of content attributes, an initial viewing probability score for each of the at least two users, wherein the trained machine learning model is trained based on a distribution of probability scores associated with the plurality of demographic attributes and the plurality of content attributes;
receiving a plurality of network data associated with the at least two users and the user location; and
adjusting, based on the plurality of network data, the initial viewing probability score for each of the at least two users.
9. The method of claim 8 , further comprising:
retraining the trained machine learning model based on the adjusted initial viewing probability score for each of the at least two users.
10. The method of claim 8 , wherein the plurality of network data comprises one or more of user device wireless network data associated with one of the at least two users, natural language processing data received at a user device associated with one of the at least two users, or account activity data for the user device associated with one of the at least two users.
11. The method of claim 8 , further comprising:
receiving, by the computing device, validation data associated with the at least two users and the user location, wherein the validation data indicates whether each of the at least two users viewed the content item.
12. The method of claim 11 , further comprising:
retraining the trained machine learning model based on the validation data associated with the at least two users and the user location.
13. The method of claim 8 , wherein the trained machine learning model is trained by:
determining, based on a received plurality of demographic attributes associated with a plurality of user locations and received content metadata, comprising another plurality of content attributes associated with a plurality of content items presented at the plurality of user locations, a training dataset comprising a subset of the received plurality of demographic attributes and a subset of the another plurality of content attributes; and
training a machine learning model using the training dataset.
14. The method of claim 13 , wherein the machine learning model is a neural network comprising a plurality of neurons, and wherein training the machine learning model using the training dataset comprises, for each of the plurality of neurons:
generating, based on a demographic attribute of the subset of the received plurality of demographic attributes and a content attribute of the subset of the plurality of content attributes, the distribution of probability scores for the demographic attribute with respect to each value of the content attribute.
15. A method comprising:
determining, by a trained machine learning model, based on:
a plurality of demographic attributes associated with at least two users at a user location, and
content metadata associated with a content item being presented at the user location, wherein the content metadata comprises a plurality of content attributes,
an initial viewing probability score for each of the at least two users, wherein the trained machine learning model is trained based on a distribution of probability scores associated with the plurality of demographic attributes and the plurality of content attributes;
receiving a plurality of network data associated with the at least two users and the user location;
adjusting, based on the plurality of network data, the initial viewing probability score for each of the at least two users;
receiving, by a computing device, validation data associated with the at least two users and the user location; and
retraining the trained machine learning model based on the adjusted initial viewing probability score and the validation data.
16. The method of claim 15 , wherein the plurality of network data comprises one or more of user device wireless network data associated with one of the at least two users, natural language processing data received at a user device associated with one of the at least two users, or account activity data for the user device associated with one of the at least two users.
17. The method of claim 15 , wherein the validation data indicates whether each of the at least two users viewed the content item.
18. The method of claim 15 , wherein the trained machine learning model is trained by:
determining, based on a received plurality of demographic attributes associated with a plurality of user locations and received content metadata, comprising another plurality of content attributes associated with a plurality of content items presented at the plurality of user locations, a training dataset comprising a subset of the received plurality of demographic attributes and a subset of the another plurality of content attributes; and
training a machine learning model using the training dataset.
19. The method of claim 18 , wherein the machine learning model is a neural network comprising a plurality of neurons, wherein each neuron is associated with one content attribute of the subset of the plurality of content attributes and one demographic attribute of the subset of the received plurality of demographic attributes.
20. The method of claim 19 , wherein training the machine learning model using the training dataset comprises, for each of the plurality of neurons:
generating, based on the one demographic attribute and the one content attribute, the distribution of probability scores for the one demographic attribute with respect to each value of the one content attribute.