Global vector recommendations based on implicit interaction and profile data

摘要

A digital media environment is described that simplifies recommendations based on vectors generated by feature word embedding. A recommender system receives data describing at least one attribute for a user profile, at least one element, and an interaction between the user profile and the at least one element. The recommender system associates each user profile attribute, element and interaction between a user profile and an element as a natural language processing word and combines these words into sentences. The sentences are entered into a word embed model to determine feature vector representations describing relationships between the user profile attributes, elements, and explicit and implicit interactions. From the feature vector representations, the recommender system determines a similarity between different features. Thus, the recommender system may provide customized recommendations based on implicit interactions even for a user profile that is not associated with historical interaction data.