Discovering Frequent Navigation Patterns For Constructing User Profile: The Case Of Ebiz Online Solutions Plc Official Website
No Thumbnail Available
Date
2015-09-04
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
World Wide Web is a huge repository of web pages and links. It provides abundance of
information for the Internet users. To reduce users browsing time a lot of research has taken
place. Ebiz (e-business) being a dynamic and fast growing online service giving organization,
it should continuously assess and monitor customers’ usage behavior and restructure the
website as well as Personalize the pages accordingly.
Web Usage Mining is a type of web mining which applies mining techniques in log data to
extract the behavior of users which is used in various applications like personalized services,
adaptive web sites, customer profiling, prefetching and creating attractive web sites. In this
study, an attempt is made to discover useful patterns from the server log files of Ebiz Official
website used as input for user profile.
In this research, web usage mining process model suggested by Lalithadevl et al. is used. This
model has five distinct phases; Data gathering, Data preprocessing, Navigation pattern
discovery, Pattern analysis and visualization, and Pattern applications. Web Server log is used
for statistical and pattern discovery. Moreover, TENDER content database is used to create
user profile. Next the log files have been preprocessed. That is, data cleaning, data integration,
feature selection, data grouping into different web content categories, transaction identification
and transformation of the data to Weka understandable format was performed using WUMprep
tool, Python programming, MS Access and MS Excel. Finally a total of 42,154 transactions
have been prepared for the experiment.
The experiment have been conducted using FP-Growth and K-means algorithms in order to
discover interesting patterns of the different web content categories. WebLog Expert is used to
yield different useful statistical reports. MS Access 2013 and W3Perl are used to create user
profile. The statistical analysis shows that 83.86 % of the user of Ebiz do not browse further
than four pages into the site. According to the evaluation result obtained on the user profile,
87.5% of the experts believe that the user profile is useful in all aspect of the questions.
The finding of the study indicates that the Ebiz (e-business) official website should be
restructured and the page needs to be personalized using the user profile as a base.
Description
Keywords
Web Usage Mining, Pattern Discovery, Log file, User Profile