Clickstream Analytics: An Experimental Analysis of the Amazon Users' Simulated Monthly Traffic
Konstantinos F. Xylogiannopoulos, Panagiotis Karampelas, Reda Alhajj · 2018
Online shopping in recent years demonstrate a constant increase and as a result the study of user behavior through clickstream has attracted again the interest of the research community. This increase though requires novel approaches to clickstream analytics since the volume of the products available online and the corresponding transactions is huge. In this paper, a sequential frequent itemsets detection methodology (SAFID) is adopted to solve a clickstream analytics problem by analyzing a composite dataset which simulates monthly traffic of Amazon U.S. online retail shop. It is shown that the methodology can perform the analysis very efficiently in a simple desktop and detect all the frequently bought together products which can provide valuable knowledge to marketers of online retail stores. The methodology used can further be improved to handle larger datasets by considering a cloud computing environment.