determining one or more aspects of a first item based on at least one descriptive text of the first item;

updating a knowledge graph containing multiple nodes that collectively represent multiple items, multiple users, and multiple aspects, wherein updating the knowledge graph comprises linking one or more nodes representing the one or more aspects of the first item to a node representing the first item with one or more first edges, each of the one or more first edges identifying weights associated with (i) user sentiment about the associated aspect of the first item and (ii) an importance of the associated aspect to the first item;

creating a vector space identifying how the nodes in the knowledge graph are connected;

using a machine learning model to generate a mimic network based on the vector space, the mimic network allowing at least one of new users, new items, or new aspects to be inserted into the mimic network without rebuilding the knowledge graph and without requiring specific behaviors of the new users to be revealed, the mimic network comprising a mimicked vector space;

creating training data by representing user-item pairs in the vector space or the mimicked vector space;

training a recommendation model using the training data;

applying the trained recommendation model to at least a portion of the vector space or the mimicked vector space to identify a second item for recommendation to a user;

identifying at least one aspect most in common between the first item and the second item based on values of weights associated with the first item and the second item in the knowledge graph; and

displaying a recommendation of the second item to the user with an explanation, the explanation showing both the first item and the second item, the explanation further showing the at least one aspect most in common between the first item and the second item.