Mobile E-Commerce Data Processing Using Relational Memory

Abstract

In this paper, we propose a very simple method for learning relationships between events by accounting for the spatial or temporal sequence of occurrence of the events. The underlying idea behind our proposed method is that for certain data processing application, such as data collected from retail shoppers, relational access to data is more useful and immediately informative than sequential access. We apply the proposed RElational Memory (REM) model on a large retail data consisting of 24,193 shoppers and 915 purchases using a popular mobile retailing iOS app. We illustrate how temporal relativity can play a role in determining the relationships between user actions.