|
Title:
|
CLICKSTREAM CLUSTERING VIA BEHAVIOURAL SIMILARITY OF WEB USER |
|
Author(s):
|
Syed Tauhid Zuhori and James Miller |
|
ISBN:
|
978-989-8533-82-1 |
|
Editors:
|
Pedro Isaías and Hans Weghorn |
|
Year:
|
2018 |
|
Edition:
|
Single |
|
Keywords:
|
Clickstream Clustering, Similarity Graph, Unique Links of the Website |
|
Type:
|
Full Paper |
|
First Page:
|
109 |
|
Last Page:
|
116 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
In recent times, web personalization has become a strategic tool. To gain a competitive advantage in the marketplace, companies adopt this strategy by promoting their sales using personalized content presentation. This paper focuses on using clickstream data to automatically create (1) a set of clusters modelling the website; (2) the distribution of hyperlink and cluster usage; and finally (3) the distribution of the union of hyperlink and cluster usage. The proposed system estimates user similarity by comparing an individuals historical usage patterns with the historical usage patterns of other known users; and subsequently forms of graph to represent the entire set of user similarity metrics. Iterative Conductive Cut Clustering is used to partition the graph into clusters. The accuracy, precision and recall of the produced clusters are evaluated against clusters produced using a state-of-the-art alternative. |
|
|
|
|
|
|
