training a wide machine learning model with historical feature vector inputs, using unsupervised learning, to generate a wide ranked element output, wherein the historical feature vector inputs include historical article elements and access records associated with the historical article elements;

training a deep machine learning model with the historical feature vector inputs, using unsupervised learning, to generate a deep ranked element output;

generating a set of inputs specific to an individual entity, the set of inputs derived from element access records associated with the individual entity;

obtaining a set of current article database elements, the current article database elements published according to specified time period criteria;

creating a feature vector input according to the set of inputs and the set of current article database elements;

processing, by the wide machine learning model, the feature vector input to retrieve a wide model subset of the current article database elements having a highest correlation with the element access records of the set of inputs;

processing, by the wide machine learning model, the wide model subset of the current article database elements to generate a wide ranked element list;

processing, by the deep machine learning model, the feature vector input to retrieve a deep model subset of the current article database elements having a highest correlation with the element access records of the set of inputs;

processing, by the deep machine learning model, the deep model subset of the current article database elements to generate a deep ranked element list; and

merging database elements of the wide ranked element list and the deep ranked element list to generate a ranked element recommendation output.