System and method for dynamically displaying images on electronic displays according to machine-learned patterns

A method of recommending at least one item for display on a user device, the method comprising the steps of:

• receiving N item descriptions, each item description in the N item descriptions including an image;
• extracting, from each image in the N item descriptions, a plurality of features including machine-understandable visual parameters;
• arranging a subset of the N item descriptions in an arrangement on an electronic display device according to the machine-understandable visual parameters;
• sending, to the user device, the arrangement of the subset of the N item descriptions;
• receiving, from the user device, signals indicating an interaction with or relative to the electronic display device;
• in response to the signals, replacing an item from the subset, removing an item from the subset, or adding a new item to the subset to produce a modified subset of the N item descriptions based on a machine learning model;
• rearranging and displaying the modified subset of the N item descriptions on the electronic display device;
• wherein the machine learning model is modified according to at least how a user of the user device interacts with the subset of the N item descriptions displayed on the electronic display during an interaction session, wherein the interaction session includes (a) which of the subset of the N item descriptions is selected as the interaction, (b) how much time is spent before a selection is made of at least one of the subset of the N item descriptions, or (c) a scrolling interaction relative to the electronic display device to identify those of the subset of the N item descriptions that are removed from a visible area of the electronic display in response to the scrolling interaction.