Predictively identifying activity subscribers

A computer-readable medium bearing computer-executable instructions which, when executed on a computing system comprising at least a processor, carry out a method, comprising:

• training a machine learning model to predictively identify content sources that are not activity subscribers that could likely become activity subscribers, wherein training includes:
• analyzing a corpus of activity data to identify a first plurality of content sources that are activity subscribers and a second plurality of content sources that are not activity subscribers;
• providing, to the machine learning model as part of the training, first information regarding the first plurality of content sources as positive examples in a training data that is used to train the machine learning model;
• and
• providing, to the machine learning model as part of the training, second information regarding the second plurality of content sources as negative examples in the training data;
• obtaining, from the machine learning model and based at least in part on processing of the first information and the second information by the machine learning model, training outputs indicating predicted likelihood scores;
• tracking an accuracy of the machine learning model by analyzing the training outputs and actual results information;
• and
• modifying, based at least in part on the accuracy and the training outputs, the machine learning model by adjusting at least one of hyperparameters or processing nodes associated with the machine learning model;
• accessing a corpus of user activity records indicating interactions with a plurality of content items;
• identifying, from the corpus of user activity records, a plurality of user activity records indicating activity with content items corresponding to a content source by users other than the content source;
• subsequent to training, providing the plurality of user activity records to the machine learning model as input data to identify the content source as a potential activity subscriber;
• generating activity statistics for the content source based at least in part on the corpus of activity data;
• and
• notifying the content source with the activity statistics and information for becoming an activity subscriber.

A computer system comprising:

• a processor;
• and
• a memory, wherein the processor executes instructions stored in the memory as part of or in conjunction with additional components, the additional components comprising:
• an executable activity subscriber predictor that, in execution on the computer system:
• trains a machine learning model to predictively identify content sources that are not activity subscribers that could likely become activity subscribers, wherein training includes:
• analyzing a corpus of activity data to identify a first plurality of content sources that are activity subscribers and a second plurality of content sources that are not activity subscribers;
• providing, to the machine learning model as part of the training, first information regarding the first plurality of content sources as positive examples in a training data that is used to train the machine learning model;
• and
• providing, to the machine learning model as part of the training, second information regarding the second plurality of content sources as negative examples in the training data;
• obtaining, from the machine learning model and based at least in part on processing of the first information and the second information by the machine learning model, training outputs indicating predicted likelihood scores;
• tracking an accuracy of the machine learning model by analyzing the training outputs and actual results information;
• and
• modifying, based at least in part on the accuracy and the training outputs, the machine learning model by adjusting at least one of hyperparameters or processing nodes associated with the machine learning model;
• accesses a corpus of user activity records indicating interactions with a plurality of content items;
• identifies, from the corpus of user activity records, a plurality of user activity records indicating activity with content items corresponding to a content source by users other than the content source;
• provides the plurality of user activity records to the machine learning model as input data, subsequent to training the machine learning model, to identify the content source as a potential activity subscriber;
• and
• notifies the content source with information for becoming an activity subscriber.