10-06-2017, 06:51 AM
A recent trend in society is social networking, a powerful tool for people to meet and
interact based on common interests. Data mining is another powerful tool used to
understand the vast amounts of data that are produced by social interactions, in order to
enhance the services being provided, and for marketing. University alumni systems exist
to promote active and ongoing relationships between graduates and their alma mater.
This research proposes to incorporate selected features of social networking and data
mining into alumni systems. Such an alumni system is termed a smart alumni system
(SAS).
There are two major contributions of this research, a framework for smart alumni
systems, and a proof-of-concept prototype implementation of an SAS subset. The SAS
framework expands stakeholder roles beyond alumni to include current students, faculty,
staff, and guests. The framework supports social networking style interactions within and
across stakeholder types, for activities such as mentoring, fund-raising, curriculum
development, etc. In the SAS framework, the primary purpose of data mining is to
provide recommendations for establishing associations between stakeholders; a
secondary purpose is analyzing results from university and departmental surveys.
The proof-of-concept smart alumni system prototype has been implemented as a webbased
web system. The prototype implements stakeholder roles for students, faculty and
alumni, and supports social networking features of friends, groups and messaging. Basic
data mining algorithms are used to provide a ranked list of recommendations for
stakeholder relationships for friends or groups.
1
1. INTRODUCTION
Alumni are one of the most important assets to any university. They are the people
who represent the university in the real world. Many alumni networks were initially
started from regional groups of alumni brought together for university fundraising
activities. Later, these networks slowly gained added importance in the development of
the universities because of their enormous outreach potential that benefits the university
and helps current students in their career paths. The alumni groups have been in existence
for decades and they are constantly changing with time. There have been very big
changes in the recent years with the development of the internet and social networking
that forces the alumni system to undergo huge changes. Therefore, it is really important
for universities to focus on the alumni networks and find ways to enhance their growth
and development.
A new intelligent system called Smart Alumni System (SAS) proposed in this
thesis offers an inter-mix of traditional alumni systems and social networking sites. A
brief study on both systems is undertaken and the important features of these systems are
incorporated into the new proposed system. Then the concept of data mining is used to
pre-process large and complex datasets in extracting non-obvious patterns of correlations
by removing unrelated data (noise) to discover hidden non-obvious patterns that represent
valuable knowledge discoveries. It is a process of observing patterns in the data and
summarizing the findings in terms of usable information. When carrying out the data
mining operations on the system, some essential associations are revealed which should
be studied in the development of more useful university alumni systems. This system
2
informs the university of the main areas to seek donations, improve interactions and
determine from which parts of the world people are more interested to join the university.
This thesis describes SAS and explains the important features and the results of data
mining in detail.
