| CPC H04N 21/251 (2013.01) [G06N 3/048 (2023.01); G06N 3/049 (2013.01)] | 20 Claims |
|
1. A method comprising, by one or more computing systems:
accessing content viewing data associated with a first user account, wherein the first user account is associated with one or more client devices, and wherein the content viewing data comprises temporal-based content viewing data;
determining, based on the content viewing data, N arrays of time-series based user viewing data;
generating N number of LSTM layers of a deep-learning model;
processing, by each one of the N LSTM layers, a different one of the N arrays of time-series based user viewing data by executing one or more sequence models of each of the N LSTM layers to determine a set of content viewing features based at least in part on the different one of the N array of time-series based user viewing data;
concatenating the content viewing features determined using the one or more sequence models of each of the N LSTM layers into a single computational array;
providing, using one or more dense layers of the deep-learning model, the single computational array to an output layer of the deep-learning model; and
calculating, based on the output layer, one or more probabilities for one or more labels for the first user account, wherein each label comprises a predicted attribute for the first user account.
|