TRANSFORMING DATA DUST TO
DATA GOLD
An inaugural lecture delivered
at the University of Ibadan
on Thursday, 25 August, 2011
By
ADENIKE OYINLOLA OSOFISAN
Professor of Computer Science
Facuity of Science
University of Ibadan
Ibadan, Nigeria.
UNIVERSITY OF IBADAN
Ibadan University Press
Publishing House
University of Ibadan
Ibadan, Nigeria.
© University of Ibadan, 2011
Ibadan, Nigeria
First Published 2011
All Rights Reserved
No part of this publication may be reproduced, stored in a retrieval
system or transmitted in any form or by any means, electronic,
photocopying, recording or otherwise, without permission in
writing from the publisher.
ISBN: 978 - 978 - 8414 - 60 - 5
Printed by: Ibadan University Printery
The Vice-Chancellor, • Deputy Vice-Chancellor (Adminis-
tration), Deputy Vice-Chancellor (Academic), Registrar,
Librarian, Provost of the College of Medicine, Dean of the
Faculty of Science, Dean of the Postgraduate School, Deans
of other Faculties, and of Students, Distinguished Ladies and
Gentlemen.
I feel highly honoured to deliver this inaugural lecture, the
second from the Computer Science Department since its
inception in 1975. The first was given by Prof. Olu Longe ill
the 1983/84 series with the title "If a Divination 2nd Comput£r
Science". The gap in time (1983/84 to 2010/11) is an
indication of the state of affairs of the department over time.
But I will return to this issue later.
Mr. Vice-Chimcellor Sir, today is of historical importance
because this is the very first inaugural lecture to be delivered
in Nigeria by a female professor in the field of Computer
Science. The title of my lecture is "Transforming Data Dust
to Data Gold". I chose this topic advisedly, first, because of
the significance of data to the "Knowledge Economy" of the
present century, especially in our country; then secondly,
because of the many opportunities that we have lost as a
result of not adequately collecting and storing historical data
over time; and thirdly, because of the vast benefits of this
untapped knowledge that we have thereby missed. And
finally, I also chose this topic because it encapsulates much of
the thrust of my research activities in Data Communication
and Knowledge Discovery in Databases (KDD) as well as my
administrative and professional activities within and outside
the university environment. Hopefully therefore, this
inaugural lecture will reveal how data that have hitherto been
gathering dust in our various walks of life may be turned into
valuable powerful decision-making tools that will in turn
create national wealth and add value to our economy and our
lives.
Definitions
First let me begin with definitions.
Dust in the sense that it is used in this lecture refers to the
wasted material, like the particles of dust that settle especially
on household surfaces; something that is worthless. Gold on
the other hand, is something precious, like the soft, yellow
and heavy element that bears the name; something quite rare
but very valuable. The weights of these two definitions rest
on those two adjectives-worthless and valuable.
Data are facts or observations about physical phenomena
or business transactions. Specifically, data are objective
measurements of the attributes of entities such as people,
places, things and events. Data can be numbers, words,
measurements, observations or even just descriptions of
things. Data in Computer Science are often viewed as the
lowest levels of abstraction from which information and then
knowledge are derived. This process of deriving Knowledge
from Data is what has led to a new research area in Data
Bases, referred to as "Knowledge Discovery in Data Bases"
or KDD (Frawley, Piatetsky-Shapiro & Matheus 1992; Date
2001).
The mechanics of effectively storing, organizing, pro-
cessing and communicating data lead to its transformation
from data into information. Information is therefore the data
that have been interpreted and understood by the recipient of
a message (Lucey 1997; Mannino 2001). Consequently, for
data to become information, it has to be communicated, and
both the user and the sender must be involved in such
transformation of data. It is only when these data are com-
municated and understood by the recipient that value is added
to the transformed data, that is, information. And as we know,
information is needed to improve decision making, increase
knowledge, reduce uncertainty and add value.
Now, there is hardly a transaction that does not generate
data somewhere, internally or externally, for a particular
environment. These data accumulate daily in our environment
as hard copy reports, memos, students' results, letters, graphs,
videos, audios, animations, etc. All these data that we care-
lessly allow to gather dust in our various work places and
homes hold valuable information, such as trends and 'patterns,
2
which could be used to improve the business of governments,
schools, or private enterprises, or even home decisions, and
therefore optimize our operational success. This business of
communicating data through computer systems has led to
several areas of research in Data Communications, one of
which is Computer Network in Computer Science. The main
thrust of this evening's lecture, Mr. Vice-Chancellor, will
therefore be on Computer Networks and Knowledge
Discovery in Data Bases which are my own areas of Research
and specialization, and I will be talking about my contribu-
tions to them.
Computer Networks
The origin of computer networks has been ascribed to a series
of memos written by one J.C.R. Licklider of MIT in August
1962 discussing his "Galactic Network" concept. Licklider
envisioned a globally interconnected set of computers
through which everyone could quickly access data and
programs from any site. In spirit, the concept was very much
like the Internet of today (Wikipedia 2011).
As Licklider envisaged it, a computer network would be a
collection of autonomous computers called hosts through
which users can quickly access resources using protocols to
communicate over transmission media. These media of com-
munication may be copper wire, lasers, microwaves, cables
(coaxial, optical) and earth satellites. The distances between
the nodes may range from a few meters to several millions of
kilometers. This computer network would differ from ordi-
nary computer channels in that it would provide:
• A high degree of insulation between processing
nodes;
• Greater variability in communication protocols;
• Less emphasis on very low latency; and
• The possibility of spanning longer distances.
One criterion for Classifying networks is their scale as shown
in table 1.
3
Table 1: Classification ofInterconnected Processors by Scale
Inter processor Processor located Example
distance in same place
0.1 meter Circuit board Data Flow Machine
1.0 meter System Multicomputer
10 meters Room Local Area Network
,100 meters Building (LAN)
1 kilometer Campus
10 kilometers City Metropolitan Area Network
(MAN)
100 kilometers Country Wide Area Network
1,000 kilometers Continent (WAN)
10,000 kilometers Planet The Internet.
Needless to say, Licklider's dream has become a reality
today, Let us proceed to other definitions and clarifications.
The unit of information transferred on a network is called
a message in complete form or in shorter packets. In any
network, there exist collections of hosts intended for running
users' programs. The hosts are connected by a communi-
cation subnet. The work of the subnet is to carry messages
from host to host, and it consists of two basic components:
• Switching elements, and
• Transmission lines
Let me briefly try to explain how the system works. The
switching elements are generally specialized computers often
called:
• Communication computer or Frontend,
• Packet switch,
• Node, and
• Data switching exchange.
Transmission lines are often called channels or circuits.
Broadly, there are two types of designs for the subnet:
• Broadcast channels, and
• Point-to-point channels.
4
In broadcast channel design, there is a single communi-
cation channel shared by all nodes. Packets sent by any node
are received by all other nodes. A portion of control
information in the packet specifies to whom it is intended, so
that if a node receives a packet not intended for it, such a
node just ignores the packet.
Broadcast channei may further be subdivided into static
and dynamic, depending on how the channel is allocated.
Static allocation could divide time up into discrete intervals
and run the channel in a round robin fashion allowing each
node to broadcast only when its time slot comes up. Static
allocation has the problem of wasting channel capacity when
a node has nothing to send during its time allocation.
Dynamic allocation method is either centralized or
decentralized. In centralized channel allocation method, there
is a single entity which determines which node goes next. But
in the decentralized allocation method, there is no central
entity, each node must decide whether to transmit or not.
There are algorithms designed to bring order to any chaos that
might arise.
In point-to-point channels, on the other hand, the network
contains cables or leased telephone lines, each one connecting
a pair of nodes. If two nodes do not share a cable and they
wish to communicate, they must do so indirectly via another
node. When a packet is sent from one node to another via one
or more intermediate nodes, this packet is received at each
intermediate node, stored there until the required output line
is free and then forwarded. A subnet using this principle is
called a point-to-point store and forward subnet. An
important design issue is what the node interconnection
topology should look like.
Topology is the virtual shape or structure of a network.
This virtual design does not correspond to the actual or the
physical shape of the computer network. The transfer of data
in a network is referred to as Logical Topology (the basic
network) while the Physical Topology (the core network)
accounts for the physical structure of the network that carries
5
devices, cable installations and locations (Horak 1999; Bartee
1985; Schwartz 1977; Tanenbaum 1996).
Network topologies are categorized into five basic types
as shown in figures 1-5. They are,
• Star
• Bus/Line
• Tree
• MeshlFull y Connected
• Ring
Let us quickly look at these topologies one by one, to
understand how they work (Wikipedia 2011).
First, a Star Network features a central connection point
called a "hub" that may be a switch or router. Devices
typically connect to the hub with Unshielded Twisted Pair
(UTP) Ethernet. Compared to the bus topology, a star net-
work generally requires more cable, but a failure in any star
network cable will only take down one computer's network
access and not the entire LAN. If the hub fails, however, the
entire network also fails. Many home networks use the star
topology.
Fig. 1. Star Topology
6
Bus Networks use a common backbone to connect all
devices. A single cable that is the backbone functions as a
shared communication medium that devices are attached to or
tapped into with an interface connector. A device wanting to
communicate with another device on the network sends a
broadcast message onto the wire that all other devices see, but
only the intended recipient actually accepts and processes the
message. Ethernet bus topologies are relatively easy to install
and do not require much cabling compared to the alternatives.
lOBase-2 ("ThinNet") and lOBase-5 ("ThickNet") were both
popular Ethernet cabling options many years ago for bus
topologies. However, bus networks work best with a limited
number of devices. If more than a few dozen computers are
added to a network bus, performance problems will likely
result. In addition, if the backbone cable fails, the entire
network effectively becomes unusable.
Fig. 2. Bus Topology
Tree Topologies integrate multiple Star topologies
together onto a Bus. In its simplest form, only hub devices
connect directly to the tree bus and each hub functions as the
"root" of a tree of devices. This bus/star hybrid approach
supports future expandability of the network much better than
7
a bus (limited in the number of devices due to the broadcast
traffic it generates) or a star (limited by the number of hub
connection points) alone.
Fig. 3. Tree Topology
Mesh Topologies involve the concept of routes. Unlike
other topologies, messages sent on a mesh network can take
any of several possible paths from source to destination. A
mesh network in which every device connects to every other
is called a full mesh. Some WANs, most notably the Internet,
employ mesh routing.
Fig. 4. Mesh Topology
In a Ring Network, every device has exactly two neigh-
bours for communication purposes. All messages travel
through a ring in the same direction either "clockwise" or
"counter clockwise". A failure in any cable or device breaks
the loop and can take down the entire network.
Fig. 5. Ring Topology
To implement a ring network, one typically uses Token
Ring, or 'FDDI (Fiber Distributed Data Inter-face), or
SONET (Synchronous Optical Network).
The Token Ring is a data link technology for LANs. It
operates at layer 2 of the Open System Interconnection
Model. It was developed by IBM during the 1980s as an
alternative to Ethernet. It significantly decreased in popularity
in the 1990s and was gradually phased out of business
networks as Ethernet technology began to dominate LAN
designs.
The Fiber Distributed Data Interface (FDDI), in its own
case, specifies a 100-Mbps token-passing, dual-ring LAN
using fiber-optic cable. FDDI is frequently used as high-
speed backbone technology because of its support for high
bandwidth and greater distances than copper. Furthermore,
FDDI uses dual-ring architecture with traffic on each ring
flowing in opposite directions (called counter-rotating). The
dual-rings consist of a primary and a secondary ring. During
normal operation, the primary ring is used for data
9
transmission, and the secondary ring remains idle. The
primary purpose of the dual rings is to provide superior
reliability and robustness. Figure 6 shows the counter-rotating
primary and secondary FDDI rings.
Fig. 6. FDDI uses counter-rotating primary and secondary rings.
(www.pulsewan.comldataJOl/fddi_basics.htm) (July 24,2011)
SONET on the other hand, is a physical layer network
technology designed to carry large volumes of traffic over
relatively long distances on fiber optic cabling. SONET was
originally designed by the American National Standards
Institute (ANSI) for the USA public telephone network in the
mid-1980s. SONET possesses several characteristics that
make it appealing on the Internet today. For instance,
• It defmes clear interoperability standards between
different vendors' products;
• It carries nearly any higher-level protocol (including
IP); and
• It includes built-in support for ease of management
and maintenance.
One could arrange the Home Network in a circle but it
does not replicate Ring Topology. Graph Theory is used for
studying network topology. Nodes' distance and inter-
connectivity, and the rate of t~ansmission and types of signals
10
of two networks might vary, but their topologies could be
identical. Complex and efficient networks can be built as
hybrids of two or more of the above basic topologies.
It is therefore clear that Network Topology is not some-
thing one can just configure by a rule of thumb or by
favouritism of one node (department or faculty) over the
other. The University of Ibadan has just learnt the bitter
lesson of this. Our UI network topology was not configured
using computer networks engineering and scientific principles
as was advised by my humble self when it was being done.
Now many years afterwards it has taken an expert from the
University of Oregon to tell UI ICT team that there is the
need to address some technical aspects of the Network design
so that when the bandwidth is increased, it will be well
utilized across all nodes. Also, the optical fibre links should
be circular, such that any damage to any route should not cut
off the nodes on that route, that is, there should be alternative
signal routes. Technically, there should be only one Network
Address Translation (NAT) instead of the present two that UI
has, and the issue of authentication at every node should be
limited to a central authentication server.
The simple truth, Mr. Vice-Chancellor Sir, is that node
locations, interconnectivity and rates of transmission cannot
be determined except through hard core data communication
principles. Data dust will be transformed to data gold only
when one uses computer communication theories and not the
rule of thumb, favouritism and/or ignorance.
Knowledge Discovery in Databases (KDD)
Mr. Vice-Chancellor Sir, this is where I come to the subject
of Knowledge Discovery and, specifically, of Data Mining,
which is my own special field. KDD (Knowledge Discovery
in Databases) is a field of computer science that includes the
tools and theories to help us extract useful and previously
unknown information and knowledge from large collections
of digitized data. KDD consists of several steps, and Data
Mining is one of them. Data Mining is the application of a
specific algorithm in order to extract patterns from data. It is a
1 )
field of Computer Science that deals with the mapping of
low-level data into other forms that are more compact,
abstract and useful. KDD achieves this by creating short
reports, modelling the process of generating data and
developing predictive models for future cases.
Due to the exponential growth of data, especially in areas
such as business and technology in our modem times, KDD
has become a very important tool to convert this large wealth
of data into business intelligence. The manual extraction of
patterns has become increasingly outmoded and inadequate in
the past few decades. As a result, KDD is now the preferred
choice for various applications such as social network
analysis, fraud detection, scientific research, investment,
manufacturing, telecommunications, data cleaning, sports,
information retrieval, marketing, and several others.
The KDD process has several steps as shown in figure 7.
Starting with the development of the application domain and
the goal, it then proceeds to the creation of the target dataset.
This is followed by cleaning, preprocessing, reduction and
projection of data. The next step is using Data Mining
algorithms (explained below) to identify patterns. Finally,
discovered knowledge is consolidated by visualizing and/or
interpreting these patterns.
There are two major Data Mining goals as defined by the
specific application, and they are, namely, verification, or
discovery. Verification is checking and confirming the user's
hypothesis about data, while discovery is automatically find-
ing interesting patterns among the data. There are four major
data mining tasks:
• Clustering, which is identifying similar groups from
unstructured data;
• Classification, which is learning rules that can be
applied to new data;
• Regression, which is finding functions with minimal
error to model data; and .
• Association (Summarization) which is looking for
relationships between variables.
12
After this, the specific data mining algorithm needs to be
selected. Depending on the goal, the algorithm could be one
or a combination of the following:
• Market Basket Analysis
• Memory-Based Reasoning
• Automatic Cluster Detection
• Link Analysis
• Decision TreeslRule Induction
• Artificial Neural Networks
• Genetic Algorithm
The first, Market Basket Analysis (MBA), is a form of
clustering used for finding groups of items that tend to occur
together in a transaction. As a clustering technique, MBA is
useful for problems where one wants to know what items
occur together or in a particular sequence. The resulting
information is often used for purposes such as planning store
layouts and bundling of products.
The second, Memory-Based Reasoning (MBR), is a
directed data mining technique that uses known instances in a
model to make predictions about the unknown. It is a
technique directly borrowed from the field of Artificial
Intelligence (AI). MBR looks for the nearest neighbours in
the known instances and combines their values to assign
classifications or prediction values. MBR may be usedon
more complex data types, such as texts and images, when
distance functions are available in these domains. .
The third, Clustering Detection (CD) is the building of
models that find data records that are similar to each other.
There are many methods for finding clusters - geometric
methods, statistical methods and neural methods. CD is a
very good way to start any analysis on data. Self-similar
clusters can provide a starting point for knowing what is in
the data and for figuring how best to make use of them.
The fourth, Link Analysis (LA), follows relationships
between records to develop models based on patterns of
relationships. Link analysis is useful in the law enforcement
13
profession, where clues about crimes are linked together to
help solve them.
The fifth, Decision Trees (DTs), are a powerful model
produced by a class of techniques that include Classification .
and Regression Trees (CART) and Chi-squared Automatic
Induction (CHAID). DTs are used for directed data mining,
particularly classification. The decision tree model is explain-
able because it takes the form of explicit rules. Therefore, the
results can be evaluated and the key attributes in the process
can be identified. The rules can be expressed easily as logic
statements in a language such as SQL, so that they can be
applied directly to new records. Along with Link Analysis,
Decision Trees arose from graph theory and its application to
Data Structures in Computer Science.
The sixth, Artificial Neural Networks (ANN), are simple
models of neural interconnections in brains adapted for use
on digital computers. ANN detects patterns in data in a
manner analogous to human thinking. It can be used for both
directed and undirected Data Mining (DM). In the simplest
form, ANN learns from a training set, and then generalizes
patterns inside it for classification and prediction. It is used
for undirected DM in the form of self-organizing maps and
related structures and time series predictions. Neural net-
works are of particular interest because they offer a means of
efficiently modelling large and complex problems in which
there may be hundreds of predictor variables that have many
interactions. Neural nets may be used in classification prob-
lems (where the output is a categorical variable) or for
regressions (where the output variable is continuous).
Finally, the seventh, Genetic Algorithms (GA) are not
used to find patterns per se, but rather to guide the learning
process of data mining algorithms such as neural nets.
Essentially, genetic algorithms act as a method for perform-
ing a guided search for good models in the solution space.
They are called genetic algorithms because they loosely
follow the pattern of biological evolution in which the
members of one generation (of models) compete to pass on
14
their characteristics to the next generation (of models), until
the best (model) is found. The information to be passed on is
contained in "chromosomes," which contain the parameters
for building the model. For example, in building a neural net,
genetic algorithms can replace back propagation as a way to
adjust the weights. The chromosome in this case would con-
tain the weights. Alternatively, genetic algorithms might be
used to find the best architecture, and the chromosomes
would contain the number of hidden layers and the number of
nodes in each layer. As you can see, Genetic Algorithms and
Neural Network came from attempts to create computing
models using biological processes. Genetic Algorithms are
also used to enhance MBR and neural networks (Fayyad et al.
1996, 2002; Fausset 1994, Berry & Linoff 1997; Golberg
1989; Haykin 1994; Speckt 1991; Witten & Frank 2000;
Ferruci & Lally 2004).
The two terms KDD and Data Mining (DM) are often
used interchangeably, but actually they refer to two related
yet slightly different concepts. While KDD is the overall
process of extracting knowledge from data, DM is a step
inside the KDD process, which deals with identifying patterns
in data. In other words, DM is only the application of a
specific algorithm based on the overall goal of the KDD
process as shown in figure 7.
Mr. Vice-Chancellor Sir, my research work in KDD has
focused mainly on the use of ANN algorithms combined with
Genetic Algorithm and/or Decision trees. With your per-
mission I shall proceed now to provide highlights of my
humble contributions in academics to this process of turning
data dust to data gold.
15
Transformed
- J Data
f Preprocessed t
• Data I
., TDaartgaet •1 I ,.'
! ~
•---•••••••---..- •••••••••••••••••••••••••~•••••• u>:,,,::
Fig. 7. KDD Process (Adapted from r~ ad
My Contributions - . "" ~.•.~.
My contributions have been mainly in Data Comunications,
in Computer Networks in particular, and' Knowledge
Discovery In Data Bases. .?i'
:;~
,.».·t"
NINET
My first contact with research in Data Communications was
during my PhD work at the Obafemi Awolowo University,
Ile-Ife. I had initially wanted to do my research in Data Bases
but there was no one senior enough in the area to supervise
the work. So the then Ag. HOD, Dr. A.D. Akinde (now
Professor Akinde and Bishop of Lagos Mainland Anglican
Communion) asked me if I could switch to Data Communi-
cation, which I gladly did. He agreed to supervise me. Little
did I know then that I would also be making history at the end
of the research work that would lead to the design of an
efficient national computer network, especially for data
handling purposes. We came up with a 3-level hierarchical
(Tree Topology) computer communication network, which
we named the NIgerian NETwork (NINET), as well as a
prototype distributed Operating System (NINOS) (Akinde &
Osofisan 1994). Four design questions were posed and
answered in the research work as follows:
16
1. Topological Design. That is, given the geographical
location of computers and other message sources, as
well as their expected traffic characteristics, where
should network nodes be located? How should they be
connected? What form should the network take? How
many connections between nodes would be needed?
2. Line Capacity Allocation. What size trunks, in units of
data capacity bits/see, should be used i.1n-oughout the
network?
3. Routing Procedures. How did different routing
algorithms compare? Should there be local or centrali-
zed routing control? How often should routing
information and/or routing control. be updated? How
should this be done? What should be the significant
parameters in determining routing?
4. Flow Control Procedures. How should one ensure
smooth traffic flow throughout the network? How are
bottlenecks and deadlocks to be prevented? How
should one stop anyone user, data source, or node
from tying up the network?
In answering these questions the performance criteria we
used included:
(a) Specified response time,
(b) Specified cost, and
(c) Specified reliability.
Load capacity and response time were worked out
between every pair of nodes. The simulation result showed
that the maximum line rate possible then was 56,OOOkpsdue
to the Nigerian Telecommunication environment. Our NINET
was therefore assigned to bring about resource sharing, cost
advantages and performance improvement of computer distri-
buted systems to large communities of computer users in
Nigeria. This, Mr. Vice-Chancellor Sir, was the first major
research work in Nigeria along this line.
17
NINOS
After this, the work then went ahead to propose a prototype
model of a heterogeneous distributed operating system using
clustering, which we called NINOS (Adagunodo, Akinde &
Osofisan 1999).
Mr. Vice-Chancellor, it is very unfortunate that up till
today, Nigerian Universities are still grappling with the issue
of getting connected together after the dismal failure of two
projects called NUMIS and NUNET, which are similar in
nature to NINET. And it is more disheartening to find that the
reason for this is that personal interest has been allowed to
take precedence over collective interest. Needless to say, Mr.
Vice-Chancellor Sir, personal interest over collective interest
will never transform our data dust to data gold.
NUMAPS
Still in the search for Nigeria to use the Internet as an
excellent mechanism for the distribution of educational
resources to improve teaching and learning, in the absence of
effective Internet presence in Nigeria, I then again got
involved with a project funded by UNESCO situated at the
International Institute of Theoretical and Applied Physics
(IITAP) of Iowa State University (ISU), Ames, during the
summers of 1995 and 1996. There we successfully developed
a simulated model of the Internet for use in an educational
environment. This was christened NUMAPS (Network for
Uses in Mathematics and Applied Physical Sciences). Three
inter-woven projects were successfully implemented in
NUMAPS. They were:
• Netshell,
• Netpack, and
• IntraNET (Osofisan & Fils 1996; Osofisan 1996).
These three projects provide educators, teachers and
students the opportunity to acquire knowledge about how the
Internet works, explore the electronic frontier, create World
Wide Web (WWW) pages, and use Uniform Resource Loca-
tors (URLs) to locate resources from other computers even
without full access to the Internet.
. 18
Mr. Vice-Chancellor Sir, I am glad to say that the
University of Ibadan was the first to benefit from this project
in Nigeria, because as soon as I moved from the Polytechnic,
Ibadan, to UI in March 1997, I started supervising final year
projects in the area of Internet Technology. The under-listed
projects were designed and implemented at the University of
Ibadan and our students became the very first to be
introduced to Web Technology in Nigeria. The testimony is
in the following successful projects:
• Development of a Website-Case Study: Faculty of
Science (by Alejo H.S., Durodola P.O. & Ebroo R., in
February 1998).
• Design of a Website for University of Ibadan, College
of Medicine (by Amore B. 0., Ajayi F.T., Olusanya
O.A. & Egwaikhide J.O., in Feb., 1999).
• Design of a Website for the Offices of the Vice-
Chancellor, Registrar and the Registry (by Adeyemo
A., Ebitigha A.K.A. & Ezeh V.N., in Feb. 1999).
These projects were demonstrated to the management of
the University at the time, with the hope that Computer
Science Students would be used to develop the University of
Ibadan Website. The idea was well received by the then Vice-
Chancellor, but surprisingly not by those who were supposed
to make it happen. Hence up till now, since 1998, UI is still
struggling with getting a decent website of world-class
standard.
However, Mr. Vice-Chancellor Sir, I wish to seize this
opportunity to congratulate you on UI's recent outing on the
Webometric ranking as announced last month. As a result of
your fresh initiatives recently, UI is now first in Nigeria. This
is undoubtedly the beginning of good times for us in
Information Technology, although there is still more work to
be done of course. All of us now live in the world of
Knowledge Economy and UI must begin earnestly to trans-
form its data dust to data gold by using sound principles and
theories in Information Technology, instead of the rule of
thumb, favouritism and/or ignorance as has hitherto been the
practice.
19
E-Commerce
In 1998 we extended the idea of NUMAPS to commerce and
came up with a design of an e-commerce site where cus-
tomers and sellers may negotiate on price, participate in an
auction and come to an agreement about prices at which the
merchandise would move. The design can also handle per-
sonalized bundled goods and customized interactions with
negotiations over value rather than prize. Thus, our design
gives a practical example of the ability of e-commerce to
negotiate both price and quality in addition to mass
customization (Osofisan 2000b).
Again, Mr. Vice-Chancellor, this knowledge was trans-
ferred to our final year students, some of whom carried out
their final projects in the area of website design for various
industries in Nigeria. An example of such a project is:
• Designing a website for advertisement of products and
services: Case study of Lever Brothers Nig. PLC,
Dockyard Road Apapa, Lagos (by Oso, 0.0., Orioye,
B.D. & Akinwale, O. N., in May 2000).
I can assure you Sir, that Lever Brothers, eagerly took
advantage of this work. There are many of such projects
produced by students of Computer Science under our super-
vision, which have been of Nigerian industries.
NETBILL
In 2001, Osunade, Adeyemo and Osofisan (2001) came up
with a business model called NETBILL, that is, a set of
protocols and software implementation that allows customers
to pay owners and retailers of information. This is an e-
commerce solution for businesses dealing in the sale of
information such as universities, research institutes, etc.
NETBILL uses a single protocol that supports charging in a
wide range of service interactions. Mr. Vice-Chancellor Sir,
there is no reason why the University of Ibadan cannot
engage in the sale of information and also make money
through foreign exchange. Companies such as Google,
Yahoo, and Amazon, to mention just a few, have made
20
fortunes out of such sales of information. These companies
have transformed data dust to data gold. We have the know-
ledge and the expertise to follow their example.
The UI Network
In 2000, in the bid to get the University of Ibadan network
correctly configured, we embarked on a research on coope-
rative caching in web servers. Historical trace data were
collected from two ISP's in Nigeria. With our traces, we
evaluated quantitatively the performance improvement poten-
tial in inter-proxy cooperation between 500 to 1000 clients of
one ISP operating within a small geographical location, and
over 5000 clients nationwide. We used existing models to
demonstrate that cooperating caching has performance bene-
fits within a particular population boundary to confirm earlier
findings. The results are displayed in figures 8 and 9.
8~r-----------------------------------,
t; '.4C
CJ
:::J
o
~ 3C
2C
1C
I I
2000 40CO 6COG
C lENT PO?ULATIO~
Fig. 8. Cacheable curves for ISP A and B.
21
sooe -
400e
.§.
>- 300e -
U
.ZW..
:.5..
z 200C -
w
::;;
'U0"
0
loaD -
C-,
zccr sex
c:...~::.r-.:-:- POPULATION
Fig. 9. Mean latency as a function of client population.
The cacheable curves of the two ISPs almost overlie each
other. In both cases the hit ratio is above 50% as shown in
figure 8. This is not surprising because more than 70% of the
requests go to cacheable documents, given that most of the
Web users visited almost the same sites (Music, CNN and
some few entertainment sites). This made the population look
homogeneous in its web behaviour. The two graphs in figure
8 have a sharp knee at about 650 and 800 clients. The
increase in the hit rate below the knee implies that there is
potential benefit from cooperating caching for multiple
proxies with a small population. However, the result of about
650-800 clients falls short of the 2500 that was the case in a
previous study reported in Osofisan (2001). The low result is
due to the low bandwidth available then in Nigeria.
Sensitivity analysis was therefore carried out on popula-
tions that are strictly heterogeneous in nature, such as a
university environment. The graph of mean latency as a func-
tion of client population is shown in figure 9. This figure
shows that caches have little impact on the mean latency
22
beyond a particular population. Therefore, unless the band-
width offered by ISPs in Nigeria increases, no ISP will be
able to support the needs of the university, because the
university audience cannot exhibit the homogeneous
behaviour that the population we studied exhibited. We
immediately shared this information at that time so that some
solution would be proffered and we would get value out of
our network. We were eager to transform our data dust to data
gold.
To our surprise however, rather than praise, what we got
was a severe rebuke for our efforts. Well almost immediately
afterwards, we were proven right, because when more people
in UI came on board the network, the hit ratio of web access
fell drastically and the university Internet users became
disenchanted. Hardly could any user get a mail across from
the Faculties except from the ICT unit. Going to the ICT unit
to send and receive mails and download files became the
norm. As we all know, University of Ibadan users are com-
pletel y heterogeneous in nature; even a single department in
the university cannot possibly exhibit homogeneity in its web
behaviour because there are various research interests and
pastimes within it. That is why we have been encountering
the kind of problems we are having and will continue to have,
until the situation is rectified. Fortunately as we indicated
earlier, the new Administration is fully aware of the problem
and is taking steps already to solve it. For this, I thank you
Mr. Vice-Chancellor Sir, and also pledge that we on our part
will not relent in our efforts to make the University of Ibadan
network functional as in the other world universities. May the
Almighty God help us all. Amen.
Content Distribution Networks (CDNs)
The Internet has experienced a phenomenal growth world-
wide as a result of increasing demands for contents, content
distributions and other services. Present development trends
in content networking facilities or resources provisioning has
stirred up interest in interconnecting CDNs. In order to
achieve a cost effective content delivery and better overall
23
service, distinct CDN providers seek ways to cooperate and
coordinate their services. Customers' interests or preferences
are forming a very important part in the provisioning of CDN
services while taking into account some specific QoS require-
ments. Analytical modelling is a very good and effective tool
that can be used to solve the resource sharing and manage-
ment problems among autonomous CDNs in order to justify
the overall system goals.
CDNs have since evolved as cooperative and collabo-
rative groups of networks over the Internet where contents are
replicated over the surrogate servers for efficient delivery
performance to the clients and improved service cost to the
CDN providers (Osofisan & Idowu 2009a). However, a CDN
is limited in terms of Point of Presence (PoP) and capability.
Therefore, in 2011 we became concerned about content
object replication among peering CDNs and then explored the
issue of Web content delivery among them (peering CDNs)
(Osofisan & Idowu 2009b, 2011). We precisely considered
content object replication. We then provided an analytical
model for the replication problem in terms of a constrained
optimization problem subject to a mix of QoS requirements
(bandwidth and delay).
Our objective was to minimize the service cost which
consists of both the storage and consistency management
costs. In order to ensure the consistency of content objects in
replica placement, different values of reading and writing
rates were considered, assuming a flat update delivery. We
developed a greedy algorithm to obtain a near-optimal solu-
tion using AMPLICLEX. The computational results obtained
are shown in tables 2 and 3.
24
Table 2: Content Objects Replication Costs
NoofCDN X>=O X=O or 1 X>=O X=O or 1 X>=O X=O or 1
2 CDN-l I CDN-2 I. CDN-3 CDN-4 CDN-5 CDN-6 CDN-7
Objl 0.741 I 0 10
Obj2 0 0302 0.603
16.25 29.27
Obi 1 0.819 0 0 I .r , 0,273 . 0 .- . ,..J~ ~'" '" "~"I~ -,
Obj2 0 0.333 0.667 ,.'. . .I,', '.{).'< .',' ,0.222'· ' ,yo~. t/*'.. ~ ~1-'..?.~.~'.~" r ,i,,.! , .,..~
~S,35, . 9
I '" 1,,0;;1;42 I 0 ..• ,': '''4 ...
4..:!t ,\; 'fo: ' .•.. ''}
Obil 0.969 0 0 0.162 0
Obj2 0 0395 0.789 ~(}' -, 0.263 - ~. 0 0,056
~ 53.02 193.74
~<
3 Objl 1.403
,
0 0 ,,'~~(~~'~..':'1;l f, .'"~,"(,~.-~-t~'\:.., .• r : .•...r,.
Obj2 0 0.571 0 i":'''',;,''-~- e , ~"t~
Obi3 0 0 0.281 , .~ .~~~:::. ~., -. ."
19.99' 55.05' " ''''5" ," "'" ~ "'-\t ' ~, j~~~ -.': ~~'''-~~ ..•
Objl 0,771 0 0 0.4214 1"0 ~t-.;.~":J ~~~-J
Obj2 0.229 0.543 1.086 0"'; , 0362 l'#?~"'..-.~ ~ ,,:;,'
Obj3 0 I 0 0 ;.0 0 .\"..,,: t~~'1"W '.
.~.~" '~
I Obil I 0.551 I 0 10 I 0.414 0 ".J:.... 0.276 0
Obj2 I 0 I 0.673 I 0 I 10 0 ~•.. 1.: · . 0.276 0
Obj3 0 0 0,905 10 0.453 0 0.071
72.63 157.44
4 Obil 1.900 0 0
Obi2 0.099 0..873 0
Obi3 0 0 0.430
Obj4 0 0 0
30.57 55.05
25
0 0 0 0 00
~~ ~...,~N ~ ~~~...,
00
~N ~
00 00 00 00
<::> g'0(:g-; ;00z;"":;;: g;
00 -0 00 0
~ '0
0
00 ~
00 00 00 00
-N..I -~N'D
!=,O 00 0s0 000V~I
00~ 00 00 0a,0
~ 0...,
....a
~
•..
e~
00 00~
N
'D
O!=--, 00
••..CO..
U•...
N
~
~
Table 3: Content Object Replication Cost Summary Results
No of objects No of peering CDN Total replication cost (trc)
X>=O X=O or 1
2 3 16.35 29.27
2 5 35.35 94.11
2 7 53.02 193.74
3 3 19.99 55.05
3 5 57.88 119.89
3 7 72.63 157.44
4 3 30.57 55.05
4 5 77.00 128.03
4 7 101.51 266.35
This is still an ongoing research. We believe that further work
on modelling in Content Distribution Internetworking (CDI)
in particular with respect to clients' requests, re-direction and
content replications can still be explored. This can be done by
focusing on two main aspects;
1. New model realization to include new situations or
parameters to help facilitate the solution of problems
being represented.
2. Algorithm development for more complex situations
with more QoS requirements.
These two issues are being addressed in the doctoral work of
S.A. Idowu, whom I am supervising.
Cyber Crime
The use of the Internet in Africa has grown so rapidly with
the explosion in the number of Internet Service Providers
(ISPs), Internet cyber cafes and access points within the last
decade. This has had several positive impacts on the social,
economic and educational sectors on the continent. Unfortu-
nately, the image of nations such as Nigeria, Ghana and
Cameroon has also suffered as a result of the nefarious
activities of some users who, instead of exploring the Internet
for constructive purposes, turn it into a ready channel for the
perpetration of criminal activities. We are all familiar, for
27
instance, with the 'Advance Fee Fraud' (AFF) , known more
familiarl y as "419".
Nigeria, in particular, has gained recognition as a source
of fraudulent Spam mails characterized by bogus business
proposals and fraudulent joint ventures. If these mails do not
entirely emanate from the suspected nations, then spammers
have scored another point by allowing efforts to be directed
and focused solely on a few nations in Africa that only
contribute a small percentage to the advance fee fraud prob-
lem. Since most of these nations do not even have reliable
databases for apprehending cyber criminals, these criminals
can simply change location, move to other nations even in the
Western world and continue to perpetuate their heinous
crimes.
The increase in the volume of cyber criminal activities
noticed in the United States, Canada and parts of Europe may
be a consequence of this scenario. Are spammers migrating to
the Western world from resource-poor regions of the world;
from sub-Saharan Africa in particular? The implications for
cyber security and information systems research is that mea-
sures must be adopted that correctly identify causation and
locate criminals in order to direct preventive efforts in the
right direction and provide appropriate solutions.
Thus the correct identification of the origins of advance
fee fraud mail is a major concern. Although previous research
(Cuckier et al. 2007; Gbenga 2007; Igwe 2007; Profgame
2(07) suggest that these mails originate mainly from Nigeria
and other West African Countries, we decided in 2010 to
carry out research using available tracking tools to validate
previously held notions about the issue of advance fee e-fraud
mails (Longe & Osofisan 2011).
We harvested in real-time aggregated advance fraud e-
mails over a two-year period using the sinkhole aggregation
methodology as proposed by Abhinav et al. (2008). Using
freeware e-mail and Internet protocol address tracers, we
obtained results (table 4) that deviate from the generally held
belief about the origins of advance fee fraud e-mails. In table
4, the amount lost by selected fraud type is reported as
Monetary Loss.
28
Table 4: Amount Lost by Selected Fraud Type for Reported
Monetary Loss
% of Reported Average (median)
Complaint Type total dollar $ loss per
loss complaint
Nigerian letter fraud 1.7% $5,100.00
Check fraud 11.1% $3,744.00
Investment fraud 4.0% $2,694.99
Confidence fraud 4.5% $2400.00
Auction fraud 33.0% $602.50
Non-delivery 28.1% $585.00
Credit/debit card fraud 3.6% $427.50
Source: Internet Crime Complaint Centre Report (2006-2008),
http://www.ic3.gov/media/annualreports.aspx
As shown in table 5, the ICC report for the period 2006-2008
reflects that it is the United States that tops the list of nations
that perpetuate cyber crime (62% - 66.1% in three years).
This is followed by the United Kingdom where cyber crime
activities by percentage dropped from 15.9% to 10.5%
between 2006 and 2008. Nigeria is only third on the list, with
a marginal increase of 1.6% between 2006 and 2008. Two
other African countries, Ghana and South Africa are listed in
the midst of other European and Asian nations.
29
Table 5: Top Ten Countries-Perpetrators of Cyber Crime
Year 2006 Year 2007 Year 2008
United States 60.9% United States 63.2% United States 66.1%
United Kingdom 15.9% United Kingdom 15.3% United Kingdom 10.5%
Nigeria 5.9% Nigeria 5.7% Nigeria 7.5%
Canada 5.6% Canada 5.6% Canada 3.1%
Romania 1.6% Romania 1.5% China 1.6%
Italy 1.2% Italy 1.3% South Africa 0.7%
Netherlands 1.2% Spain 0.9% Ghana 0.6%
Russia 1.1% South Africa 0.9% Spain 0.6%
Germany 0.7% Russia 0.8% Italy 0.5%
South Africa 0.6% Ghana 0.7% Romania 0.5%
Source: Internet Crime Complaint Centre Report (2006-2008) - http://www.ic3.gov/media/annualreports.aspx
30
What is not reflected in table 5 however, is the typology
of these cyber crimes. For instance, even though the United
States is consistently ahead of other nations in terms of cyber
criminal activities as reflected in the table, the diversity of
cyber criminal activities responsible for the figure includes
advertisement spam, cyber pornography, and product market-
ing, bulk mails, and so on. However, a focus on cyber fraud
in table 4 shows that monetary loss due to advance fee fraud
tops the list of reported dollar loss to cyber criminal activities.
This unfortunate scenario begs the question, Are these losses
traceable to mails that emanate from Nigeria and other West
African Countries?
We answered this question in Longe and Osofisan (2011)
by using a modified version of the sinkhole methodology in
real-time as proposed by Abhinav et al. (2008). Our e-mail
accounts were activated and used from locations outside
Nigeria, specifically in the United States of America, United
Kingdom, and Canada to ascertain the source of these mails.
Subscription to all forms of online promotion and marketing
were avoided in order to avoid other forms of spam. We
received over 36,000 fraud spam mails over a two-year
period, and out of these we made a random selection of 400
for analysis. To find the actual locations from which the e-
mails originated, we picked the "Received From" IP that is at
the bottom of the list on the header view. We ran the e-mails
through specific open-source e-mail and IP address tracers
such as IPSLocator, E-Mail Trace, and IPGeolocator. The
.results we obtained are shown in tables 6 and 7 as well as
figures 10-13.
31
Table 6: Different Forms of E-mail Harvested from Our Experimental E-mail Accounts '"
Type Total no of mails(%)
Bayesian poisoning 35
4' Dating spam
"
32
Table 7: Advance Fee Fraud E-mail Distribution by Continents
Country Percentage of 400 mails
·AMERICA AND CANApA 28.5
EUROPE 23.2
AFRICA 20.4
SOUTH AMERICA - ' , 15.2
ASIA '« 9.5
AUSTRALIA 4.2
'. ft;~~ table 6, one very disturbing trend in the spam war is
, the. ~n1¥g,et:lce of individuall y targeted e-mails .and academic
" conference' (invitation) spam e-mails. Quite a number of
Individuals have fallen victims to this genre of spam. The'
fraudsters use key-loggers and access to privileged inform-
ation on individuals such as travel itineraries, to target
unsuspecting victims.
Findings from our experiment as shown in table 7, also
showed that advance fee fraud e-mails do not only emanate
from. Nigeria and some West African nations alone, but also
from the Western world (America and Europe), the Middle
East, and Asia. This result is very significant for us because it
reflects the fact that the usual focus on Africa, and West
Africa in particular, as the major source of fraudulent mailing
and cyber activities is misleading. It remains to be investi-
gated however if the increase noticeable in the volume of
fraudulent spamming activities emanating from these other
parts of the world correlates with the number or volume of
Africans, Asians and other immigrants' moving into the
Western nations, as some have alleged.
\ ' ,
33
lit uIf Uk,OIOIIW, "M! t:lWllldowlllvt 'U.hlOlo.cilforpl9Cn ItGIMSc
,n IINIt ~ [Ad4l '11II
lnWMt 14•••• fl/It Proflllll (IMPI US
11O*I~'OO12 101'
IkoItI., ioltntdo
OIU
1...'
'" --
AlU Lonery 11I11 AvardZOOt
Mo,um., ~lft"l19 •••• t,
~.~~.~~! •.~~.~~~.~seOOO,
YOUR laNAI!. AllOWSlION TIllS yw ,uIAHMILLIOLNOrrlRY I
lir/Illdal.
W. !li.h ~o acnormlm you over your ••• U IUCctU ill M c.uttr blUotillO beld on .tb DtctBer 200t.
TIIu 11 • IUU.Ml111Sounutlc _ttr OUt ln lIhich ••• 11 addre" •• litre und, It i. I pfOllCtionll
prooUI aUld It IftCOllUOll1m0lfllit ulm, therefore YO\dl o not nted to _bIIyuctn to mer for it.
YOIbIave bllllapproved for thl aur prill of Um17ISU,OO (Sllllllndrld AndSUty Sevu TIIoII,IJId,FlvtIIwIdHd
Andromy hve United Imu Dollm onlYI,To ohia yoIIr !liMillOprill yo\I art to acntaCt tbe appoiDttd elm
aOlnt •• 1001III JIO'l1blt for the ilMdlm rel •••• of your 111nnlno'with the toHovino illfOrlAtiOll1belov:
YO\lf Pull KIII't t I • I •• I I , , I I , I I I I I I I t I I I I I I • I
Fig. 10. Screenshot of e-mail purportedly from Kuala Lumpur.
34
~,,.-".-.---jet ~~~""""""""""N* -'lnl .•1I11rJ(C 11.,( e·.nl.)iI ••ndr,.iJ~kl. :v:
.•••••• h,,, •• 1?:::11'~"" eHP e~w...",. eMIniIU,..IlIll""e>MPlW •••••• e> •••••••~W~UVW fg,VttittKll<lllnO-CMlfItfPtip,
I;jemllil TrIIctn~ IlUc~tun
f,,... .••. '1Jiijiill'ililiiilii ., "''''til:ii=ii''~:l:fM:~~· __ U_···''··=:7.2••••. ~ "; ". . ••••••••
L .~~~'1:.t ~ .%•1•&../,...... ""._._•.._.
=" ':,H...•ICo_ j
"._ ••••_"",", ,"'
'If :
~", '
I
~""""'1-7.J-I~, .".'.
"f .- !;
~ "_-
.-
_~~~"I~>
:
1I"4"! !
_'I1'I_WSf2Jt _OO3._·A4II>V."'~ 1
Applicant Track'nA Softwa •••· I
~anl Tracking SOftware Into. Ac;c;efjs 10 :
. s..-c:h EPgNIS AC~.. i
••••• ,lnto.GDft\ !
'nstam I!ny!JI ,.,...,. j
t) TVfHJIn All En'I3iIAddfotllO 2) FitJd OW".,.. i
~&lnto :
am.itF ••.•••• CXWft • i
Eny ProdUC1tT!)IICUbI.1tJ i
j ~_ .-.. ~ Tho IoI1d1nQfood tnleeotHly eylllWm from froid 10 !
! ,~,s..itw~ i
:~_. www~ •
I _ _.. '-_'1'em»QIU""g 1
l.............................................................•...~.".:!.~.....•.•..:.......................
fa .IUIU ""'ill"".;)
~ ~ WhAt it my'" tkM" tree ~ ,..", c;.twMt ~. m 1M wng It OptMJnflP' M"te 1!!..fAQ reLe=Mn QWtaMM
Fig. 11. E-Mail trace result for the email showing it originates from Brazil.
35
@@nIUllp@nd@tt
'rom l"t.r.wlUh"'lt.l'I~t.Jmltu.A!U' 3.i @ft Ii UltlJllll i€Ji')§
.~A~p.r"fltIY~TDI'f.",801l.l",,88@\Vith88IE8ffi \+11167;1§S.l5.2i.OI Sit. 1; e».t II'.'0t O~I01t4\J·t
X-Y'_h"«.~i",U.,if''"''~,,,·,lPh'.'ktlIU ,·•• """.li@",I€@I:@ffillleIlH8fl¥ol£Hlflj;c&m;.,1.6;CMtUtt." ,tv] , ":X';V"'IIJJJ~.I.>"'lvOViWI.:OJjI(§j(J:lffl~IilWf3FY§§§A"A7¥tr&~\;I"t't\3J.t;~lIUt ;U_&hlt ;:Nfi l V'O~ ••....
tv) x.gritfl.".tJ"V~JPI ,,,'U60"i,O,fJ41
AIJ•••• ntic:.\~,;ft~t"t-.c:•.•wlIft.~.,~~:·'·Im,,..'i"II,atr •• millil,f@lf,¥ilh88.E8ffl tf8mii I a6miIiU~I,. •• ria;vtt,1 (flU AI,) I hOfH;; I €1~,~'tCM)4 ((;J:lb@JlijijHIi"mall;ulll:;.Stj) (i6Cdl.O.14) b, 'I\tn!'l/J.t,i.
".1ftf!ti"iJl"",~••'•'i'·f,. 'tOm U~l"8§1:§~'U4: l§; itl~,8jl'Ir.RU.,ji6f@I&EiiJl,/1iit (It, I t\;t H".I i\;?H (I'c30CHUOUiift7 ,R§H~7§l'tHtl@iHi§B.jillitiilf.mjlf.titl<l,'il"4J~
Ifro •.•'•·. "'"htf§I'#tiANji@flil blffllS.8~A§ ~ .<f 8RhA4i.4if"'tn """,II.It>t.14",tt h tHHI)I+."
• ~Htt MUH 'f4t, "&lif Ai''''' ~iluj ~.fi 13•• 11S&;jpil"d.ti0..... .." 17 ~ ~ O§I§7'i4 ;(jj(JljMI"C...y~. 1.0Fig. 12. Letter purportedly from Interswitch Nigeria Limited.
36
1111 Country I(Full) Flag ISP I
160.36.0.84 . US UNITED STATES Ii-TENNESSEE KNOXVIUi UNIVERSITY OF TEN~ESSEE
160.36.0.84 US UNITED STATES TENNESSEE KNOXVIUi UNIVERSITY OF TENNESSEE'
160.36.0.84 US UNiTED STATES iii TENNESSEE KNOXVlUi UNIVERSITY OF TENNESSEE
Fig. 13. Result from IP2Location on the validity of the IP address.
What is certain anyway is that as the Internet expands,
opportunities for unethical use will continue to increase if
nothing pragmatic is done in terms of policies, technology
and legislation to protect users against online criminals. This
is reflected in the routine activity theory which posits that
crime can be motivated by opportunities provided in routine
activities. Incidentally, the use of the web falls perfectly into
the domain of routine activities, howbeit on a global scale
(Cohen et al. 1979). But correlating poverty with cyber crime
and using this as a major factor to point to Africa and some
resource-poor environments as the sources of advance fee
fraud mails does not capture the entire picture and could be
totally misleading. Future research will explore the contribu-
tions of other subtle but extraneous factors, such as the
economic meltdown and immigration policies, to the increase
in fraudulent cyber activities now noticeable in the Western
world.
Data Mining in the Service of VI
In 2002, we obtained a 3-month fellowship from the UNDP
to Swinburne University of Technology in Australia, and it
was while there that we researched into a project in KDD
using the University of Ibadan as case study. We in particular,
set out to find a solution to the Vision goals that the
University Governing Council set in 2001. The Vision Com-
mittee had set out to find answers to the following questions:
37
• Where were we before?
• Where are we now and how did we get there?
• Where do we want to be and how do we get there?
Interestingly enough, the University over the years had
accumulated valuable data in various formats (text, video,
audio, etc) which could be used to answer their questions. But
unfortunately, the data were mostly gathering dust in several
shelves, files, and storerooms, scattered and unharnessed
together, all over the place. Within these data lies valuable
information from which knowledge can be derived to find
solutions to the three questions outlined by the Vision
Committee. We came up with a KDD framework that we
successfully demonstrated in a seminar presented at the
Swinburne University of Technology for solving the three
questions (Osofisan 2002).
Unfortunately, Mr. Vice-Chancellor Sir, the difficulties
encountered in obtaining relevant data from appropriate units
of the university has stalled the conclusion of this work. The
framework for the Decision Support Tool is ready. It is a
matter of urgency that UI must develop an enterprise Data
Base for the University, and all I can say is that given the
right cooperation, the Department of Computer Science will
be willing to assist in getting it done. As a matter of fact, we
had already proposed, as far back as ten years ago, a work-
able Data Warehouse Architecture for transactions in an
academic environment (Osofisan 2000a). The Architecture is
a hybrid of data marts that feed information into the central
data warehouse (Osofisan 2000b). These data marts were
designed with regard to cross-functional information require-
ments. The architecture is appropriate for any environment,
such as ours, that consists of an inconsistent legacy system of
islands of data. The proposed architecture combines the
advantages of locality of reference, lower entry cost, faster
implementation, and centralized system management.
38
Students' Records
In 2009, in a bid to see how we can get around the problem of
frequently missing or confused data, Olamiti A.D. and
Osofisan A.O. embarked on another research of how to make
sense out of some students' historical data available in the
department. We investigated two treatment methods for deal-
ing with missing values in students' data in the department,
and found out that the mean error rate of Embedded-C4.5
method was 3.2% less than that of Pre-processing-CD method
(Olamiti & Osofisan 2009a). The statistical t-test indicated
that there is no significant difference between the two
methods at 5% level of test. However, the two methods
showed that there were some misclassifications in the final
grades awarded to some students. Some fields that were
expected not to have null values, such as the SSCE results in
English, were found to be null. These are interesting facts that
need to be further investigated. We would like to state that the
historical data mostly fell within the period we classified as
the period of decline in our previous study (Osofisan 2002).
Further work was carried out on the same set of data (Olamiti
& Osofisan 2009b), and the following results were derived:
1. Only students with grade AI, A2, or A3 in SSCE
Mathematics graduated with First Class honours.
2. Students with grades of AI, A2, or A3 in SSCE
Mathematics but who spent extra years before-com-
pleting the programme graduated with at least Second
Class Lower only if they took further Mathematics at
SSCE.
3. All students who graduated with First Class honours
had Al in SSCE Chemistry.
Knowledge I is to be interpreted as meaning that if a
student graduated with First Class, then the Student's grade in
SSCE Mathematics is either AI, A2, or A3. The converse of
this knowledge is not true, that is, all students with AI, A2, or
A3 in SSCE Mathematics graduated with First Class.
39
Knowledge 2 supports this converse interpretation, but
also gives another interesting knowledge-that any student
who took Further Mathematics in SSCE would graduate with
at least Second Class Lower degree in Computer Science.
Knowledge 1 and 2 therefore underscore the importance of
mathematics to Computer Science.
Knowledge 3 is very interesting too. It shows that all
students who graduated with First Class degree had Al in
SSCE Chemistry. Now combining knowledge 1 arid 3 means
that we can make the following proposition: "Any student
who comes into the Computer Science department with at
least A3 in SSCE Mathematics and Ai in SSCE Chemistry
is a potential First Class Student".
We will continue to investigate this proposition for
validity with historical data from other universities in Nigeria.
But if it proves to be true, then it may be necessary to find out
the reasons behind the proposition for decision making in
respect of science education and the production, of highly
skilled software engineers. Students who score any grade of
A in SSCE Mathematics and Al in SSCE Chemistry may
therefore be targeted for special training in Software
Engineering. This crop of Software Engineers may bring
fortune to Nigeria in terms of production of world-class
marketable software as India is currently doing. This might
lead to wealth creation for Nigeria. .
Contributions to the Outside Community
Mr. Vice-Chancellor Sir, my contributions have not been
limited to the university alone. In all modesty, I have also
tried to use my knowledge in Computer Science to analyse
existing problems in the outside society and create solutions
to them. For the rest of this lecture, I will try to give a few
instances of these.
The School System
The first concerns schools and the educational system. In
2003, we decided to focus our research effort in knowledge
discovery on this area by using historical data of JSS
examination results in one state to determine the success or
40
failure of the 6-3-3-4 Educational system, most specially, the
3-3 segment. We created a data warehouse consisting of
student enrollment, performance, local government, school
ownership, core subjects, pre-vocational and elective (non-
vocational) subjects as specified in the National Policy on
Education Revised (1982). We found that overall per-
formance P can be any of the following eight values:
• All Rounder (good in core subjects, pre-vocational
and elective) - A
• Science/Humanities - SIH
• Science/Vocational - S/V
• Humanities/Vocational - HlV
• Science only - S
• Humanities only - H
• Vocational only - V
• No classification - N
The Data warehouse relation, School (S), has three compo-
nents:
• Facts
• Dimension
• Attributes
Facts attributes (LGA-ID, Sch-ID, Std-ID, Time-ID,
Name, Sex, Age, Results in courses registered for, P)
Dimensions (School, LGA, and Time)
Attributes (Ownership, Location, School Type)
We then used Classification and Clustering technique to
find data groupings, data dependencies, relationships, as well
as classifications.
The result was that we found a cluster representing about
25.9% that could not proceed to Senior Secondary School,
and another cluster of about 3% that did not fit into any
academic classification even though they passed six subjects.
These two groups should be candidates for trade centers as
proposed by the National Policy on Education. This means
that the state government should make provision for 28.9% of
JSS students to enter various trade centers in the state.
41
However, the statistics of the number of trade centers in the
state showed that only 14.5% of these students could be
absorbed by the available trade centers. This means that, for
government to make a success of the 3-3 segment of the
educational system in the state, trade centers must be
increased by about 620% (Osofisan et al. 2003). The state
therefore needed to set up many more trade centres for
various artisans in areas like plumbing, carpentry, electrical
fittings, and so on, that is, artisans in the building industry,
who are almost going into extinction nowadays.
Banks
In 2004, because of the ailing nature of the Bank Industry in
Nigeria, we investigated the Portfolio Management of one of
the Banks, also using the KDD procedure (Osofisan, Inanga,
& Adeyemo 2004). We found out that the Central Bank of
Nigeria Policy as well as the internal management of the bank
affected the bank's portfolio management. Precise! y, we
found out that no bank can adhere strictly to all the CBN
guidelines, because some of the CBN policy guidelines are
conflicting. For example the credit expansion goal and capital
adequacy goals should complement each other, but analyses
show that they conflict. The analyses also show that bank
managements tend to set unrealistic standards. The basic
result of goal achievement is shown in table 8 while the
results for variants 1 to 5 of the basic model are illustrated in
tables 9-11. PK=Oimplies that the goal at the k priority level
is achieved, while Pk=1 means that the goal at the k priority
level is not achieved. None of the variant models gave the
same level of goal achievement as the basic model solution,
and none of them was identical to another (see tables 9-11).
The results of these various combinations explain the effect
of changes in the internal policy on the performance of the
bank. Also in all the variants as well as in the basic model, at
least one of CBN statutory requirements was routinely
violated.
42
Table 8: Degree of Goal Attainment of the Basic Model
Priority level Constraints Goal achievement
TI T2 T3 T4 T5
0 Artificial Y Y y y y
I Fund allocation Y Y Y Y Y
2 Sector credit allocation Y Y Y Y Y
3 Other statutory goals K N N N ~
4 Customer deposit mix/profit y y y y y
5 Principal ratios N N N N N
6 Profit/deposit growth Y Y Y Y Iy
7 Lowe bound constraints Y Y N Y N
8 Balance sheet Y Y y y y
43
Table 9: Sensitivity Analysis Results for Time Periods for Variants 1 & 2
SIN Constraints Revised model combination
Variant 1 Variant 2
I 2 3 4 5 I 2 3 4 5
I fund availability P,=O P,=O P1=0 P1=0 1'1=0 P,=O P,=O P,=O P,=O P1=0
2 Sector credit allocation 1'3=1 P4=1 P5=1 P2=0 P2=0 P3=1 P4=1 1'5=1 P2=0 P2=0
3 Other statutory allocation P2=0 P]=1 P3=0 P3=1 P3=1 P2=0 P3=1 P3=1 1'3=1 1'3=1
4 Customer deposit mix/profit P5=1 P5=1 1\=1 P6=1 P6=0 P5=1 P5=0 1'6=1 P6=1 P6=1
5 Principal ratios goals P4=1 P2=1 P4=1 P4=0 P5=0 1'4=1 P2=1 1\=1 1'4=1 1'5=1
6 Profit and deposit growth goals 1\=1 1'6=1 1'7=0 1'7=0 P7=0 P6=1 P6=0 P7=1 P7=0 P7=0
7 Lower bound constraints P7=1 P7=1 1'2=0 1\=1 1'4=1 P7=1 1'7=1 P2=0 1'5=1 P4=1
8 Balance sheet identity 1'8=0 1'8=0 P8=0 P8=0 P8=0 P8=0 P8=0 P8=0 P8=0 P8=0
44
Table 10: Sensitivity Analysis Results for Time Periods for Variants 3 & 4
SIN Constraints Revised model combination
Variant 1 Variant 2
I 2 3 4 5 I 2 3 4 5
I Fund availability P1=O P1=O P1=O Pj=O P1=O P1=O P1=O PI=O P1=O PI=O
2 Sector credit allocation P3=1 P4=1 1\=1 P2=O P2=O P3=i P4=1 Ps=l P2=O P2=O
3 Other statutory allocation P2=O P3=i P3=O P3=1 P3=1 P2=O P3=i P3=1 P3=1 P3=1
4 Customer deposit mix/profit Ps=1 Ps=1 P6=1 P6=O P6=O Ps=1 Ps=O P6=1 P6=i P6=1
5 Principal ratios goals P4=i P2=O P4=i P4=O Ps=O P4=1 P2=1 P4=i P4=1 Pj=I
6 Profit and deposit growth goals P6=1 P6=1 P7=O P.,=O P7=O P6=i P6=O P7=1 P7=O P7=i
7 Lower bound constraints P7=1 P,,=I P2=O Ps=l P4=1 P7=i Pv=I P2=O Ps=l P4=1
8 Balance sheet identity P8=() Pg=O Pg=O Pg=O Pg=O Pg=O Pg=O P8=O Pg=O Pg=O
45
Table 11: Sensitivity Analysis Results for Time Periods for Variant 5
SIN Constraints Revised model combination
Variant 5
1 2 3 4 5
1 Fund availabilitv P]=O p]=O p]=O p]=O p]=O
2 Sector credit allocation P3=1 P4=1 Ps=l P2=O P2=O
3 Other statutory allocation P2=O P3=1 P3=O P3=1 P3=1
4 Customer deposit mix/profit Ps=1 Ps=l P6=1 P6=O P6=O
5 Principal ratios goals P4=1 P2=1 P4=1 P4=O Ps=O
6 Profit and deposit growth P6=1 P6=1 P7=O P7=O P7=!
goals
7 Lower bound constraints P7=1 P7=1 P2=O Ps=! P4=1
8 Balance sheet identity Ps=O Ps=O Ps=O Ps=O Ps=O
All these show therefore that, for the banking system to
function properly, the CBN Monetary Policy and bank
managements policies must be scientifically determined,
based on historical facts.
The Lagos Ibadan Express Way
In 2005, we turned our attention to the issue of accidents on
the Lagos- Ibadan Express Road. We wanted to look for
interesting patterns in the nature of the accidents that had
occurred in the past. Historical data about road accidents were
therefore collected from the Road Safety Commission for the
first 40 kilometers from Ibadan to Lagos. We employed
Neural Network using Multilayer Perception as well as
Multidimensional Data Analysis method and discovered that
the dark spot on the first 40 kilometers of the Ibadan-Lagos
road is between kilometers 10-20 where 60.2% of all
accidents took place. We found three interesting clusters at 0-
9 kilometers, 10-20 kilometers, and 21-40 kilometers. All the
accidents due to wrong overtaking took place in cluster 2,
while no accident caused by robbery took place in cluster 1.
One third of the accidents that occurred in cluster 1 were
caused by loss of control, followed by over-speeding and
dangerous/careless driving. Therefore, the accidents that
occurred in cluster 1 can be attributed to the drivers, and may
therefore be drastically reduced if drivers have the right
attitude to driving.
46
Cluster 2 has a fair share of all causes of accidents, with
tyre burst (71.3%) being the highest, followed by over-
speeding and loss of control. Going by the fact that in cluster
1, only 2.1% of the accidents occurred as a result of tyre
burst, with 71.3% in cluster 2, and 26.6% in cluster 3, then
tyre burst in cluster 2 may be a result of other extraneous
causes. It may therefore be necessary to further investigate
cluster 2 in order to identify these extraneous causes in order
to reduce accidents in this cluster.
In all the clusters, most of the accidents occurred in the
morning (41.4%), and in the afternoon (42.1 %). Only 16.5%
of the accidents occurred at night. Furthermore, there were
more accidents in the dry season than in the wet season.
Then, no accident due to robbery occurred at night, and big
vehicles were more involved in accidents than small vehicles
in the ratio of 59.9:40.1. This is useful information for the
authorities concerned. The results show the measures to be
taken to reduce accidents and make the roads safe (Osofisan,
Komolafe, & Akinola 2005).
University Administration
Mr. Vice-Chancellor Sir, I wish to turn now very briefly to
our modest contributions in the areas of administration. When
in September 1999, I became the Acting Head of Computer
Science Department, there were only 5 lecturers, and I was
the only one among them with a PhD degree. Hard decisions
had to be taken therefore on how to move the department
forward. The first decision was to start a postgraduate
programme, but how? After a lot of reflection and consulta-
tion, we decided that our best option was to start with
MPhillPhD conversion for all staff in the department with an
M.Sc. degree. The next question to be answered then was:
With me as the only supervisor, would all of them now be
constrained to do their research in my area and so end up with
making the department a monolithic research department?
The answer of course had to be in the negative, since that
would not be good for the system and I was determined to
build a robust department. Then God, Who sees the heart in
His infinite mercy, sent help to us.
47
This help came through the following: Prof. A. David of
the University of Nancy. 2 in France along with his
colleagues, Prof. Odile Thiery, and Dr. G Duffing; Dr. V. H.
Dang of IIST, United Nations University, Macau; Prof. A. B.
Sofoluwe (Current Vice-Chancellor of Unilag), Prof. O.
Abass; Prof. J. O. A. Ayeni and Prof. C. O. Uwadia, all of
University of Lagos; Drs. E. R. Adagunodo and G.A.
Aderounmu, now both professors of Obafemi Awolowo
University, Ile- Ife. With the assistance of this selfless lady
and the gentlemen, we were able to diversify our research
base in the department. It also meant my shifting slightly
away from my original research areas in Data
Communication and Data Base to other areas of research in
Computing since I had to co-supervise PhD work in these
other areas with colleagues within and outside Nigeria.
I take pride in the fact that, within the time of my tenure,
seven of the members of staff listed below have graduated
with PhD degrees within the department.
• OLUWADE, Bamidele Ayodeji in Code Theory
("Design and Analysis of Computer Coded Character
set", 2004)
• AKINKUNMI, Babatunde Opeoluwa in Knowledge
Representation ("Temporal Properties and Applica-
tion of Recurrent Entities", 2005)
• OSUNADE Oluwaseyitanfunmi in Data Commu-
nication ("Data Migration Patterns for JAVA-Based
Mobile Agent Systems", 2007)
• OKIKE, Ezekiel Uzor in Software Engineering
("Measuring Class Cohesion in Object-Oriented
Systems Using Chidamber and Kemerer Metrics:
Java as Case Study", 2007)
• AKINOLA, Solomon Olalekan in Software
Engineering ("A Comparative Study of the
Effectiveness of Checklist Based and Ad-Hoc
Software Inspection Reading Technique in Paper and
Tool-Based Environments", 2010)
48
• ONIF ADE, Olufade Williams in Intelligent Systems
("A Model for Information Risk Management in
Economic Intelligence Systems", 2010)
• OLADEJO, Fausat Bolanle in Knowledge Capitali-
zation ("User-Centred Capitalization of Knowledge in
the Context of Economic Intelligence System", 2010)
I am also happy to say that, apart from the 6th and 7th
carjdidates who were co-supervised by me with Professors A.
.'~Davi.~ 'CUldOdile Thiery respectively, all the remaining five
PhD candidates were solely supervised by my humble self.
Furthermore, the first PhD student, B. A. Oluwade has
since joined the services of Salem University in Kogi State as
a Professor. He is here with us in the audience today. Three
othet. members of staff also got their PhD degrees as listed
b~lQ.Jr: "
,~.~, ~QB~RTS, Charles Abiodun in Systems Modelling
'~. ,: eVannotation Pour La Recherche D'information
~' ~t~ '" . P.8¥§ .,' ,~e Con~ext D'intelligence Economigne'.
":'~~;, ·:tJnlverslty of Loria 2 France, 2007).
" •.• !Ji". \ \~EYEMO, Barnabas Adesesan in Data Mining
'!'<,., ,i-, ,:('IDe.xelopment of Data MiningSystem for Oil Well
-:,,~ Lithology and Fluid Contents Analysis". Federal
',' University of Technology, Akure, 2008).
• LONGE Olumide in Cyber Crime ("SPAMAng - A
Domain Specific Outbound Antispam system for
. Filtering Fraudulent Electronic Mail". University Of
"." Benin, 2010).
The result, Mr. Vice-Chancellor Sir, is that today, our
department can boast of a staff body of 21, 9 of whom have
PhDs. Sir, there is no other Computer Science Department in
Nigeria to my knowledge that can boast of a similar strength.
Conclusion .
Mr. Vice-Chancellor Sir, I wish to emphasize that, talking
about these projects and challen~~s, my aim is to showcase
49
the almost limitless possibilities that exist within the field of
Computer Science, and especially of KDD, to grapple with,
and offer solutions to the many problems that beset us in this
nation. Sir, the creation of Enterprise Data Bases and of Data
Warehouse is a must in every sector of the Nigerian
enterprise in this era of knowledge economy.
As many nations have discovered, the organization of life
and progress in the modem world is almost impossible with-
out the systematic gathering and organization of information.
That is what is meant when people say we live in the
information age. Unfortunately, we in this part of the world
tend to be careless about the collection of data, and extremely
nonchalant towards the management of information. This is
principally why we remain backward in the 21st century and
why we continue to suffer the terrible nightmare of under-
development. But we have the means of changing our lives
for the better and joining the globalization train not just as
passengers, but as drivers and pilots. One of the weapons in
our hands, Mr. Vice-Chancellor, is Information Technology
(IT), and by IT, I mean the Convergence of the Tele-
communication and Computing Industries. I hope, in the brief
scope of this lecture, I have been able to demonstrate
sufficiently how IT can surely transform our Data Dust to
Data Gold!
Acknowledgements
I now wish to make a few acknowledgements. First and
foremost, I acknowledge the Almighty God, my Father and
my Maker Who has moved in many mysterious ways in my
life for His wonders to perform.
I thank my father, Late Hon. Chief Josiah Orisabinu
Adedipe, the Elemo of Akure and Mayegun of Osogbo, (aka
Yagbo Yaju, Osun Osogbo) who gave me not only sound
education, but also moral uprightness. Papa, you taught your
children to respect all and fear only God. You taught us
dignity of labour, and the pursuit of excellence in all that we
do. Above all you gave equal opportunities to all your child-
ren whatever their gender, and as a result Papa, you have
50
three female professors (daughters) to your credit. May I here
humbly mention and thank my two sisters as follows:
• Prof. Adefunke Oyemade (Medicine). First Female
Medical Doctor in the old Ondo Province of the
lit.defunct Western State;
• Prof. Adeola Abaelu (Bio-Chemistry), one of the first
two students to be awarded the PhD degree of the
University of Ife, now the Obafemi Awolowo
University, Ile-Ife; and
• I am, of course, the third professor after them, and the
very First Female Professor of Computer Science, in
all of sub-Saharan Africa. '
Papa, no legacy is as good and precious as a successful
child. We can never thank you enough.
As for you, my mother, Late Chief Mrs. Felicia Fehintola
Adedipe (Nee Ogundare, and the former Akuwajo of St.
Paul's Anglican Church, Emure Ekiti), words alone cannot
express my gratitude to you. You were selfless and big-
hearted. You took care of all our children while they were
growing up. You were the "Ark of God" in our household.
Your Royal Highness, Oba Adebiyi Adesida, the Deji of
Akureland, I feel highly honoured by your presence. I remain
truly your obedient subject. To all the members of the
Adedipe Dynasty of Akure land here present, led by our
Patriarch, High Chief Hon. Bola Adedipe, the Elemo of
Akure land and former Minister of the Federal Republic of
Nigeria, I say a big thank you, to you all. Our family is a
special one in Akure, whose enduring hallmark is love for
one another. We are always there for each other through thick
and thin. I thank God for creating me to be an Adedipe, and
now an Osofisan.
I thank my brother in-law, Chief Adefemi Adekanye and
my sister, Chief Mrs. Adefunmilayo Adekanye, the Erelu of
Akure land for their love and particularly for looking after
Wale and Yemi while I was away studying at the Georgia
Institute of Technology, Atlanta, Georgia.
51
I thank all my relations from Emure-Ekiti here tonight. I
remember with nostalgia, my maternal grandmother, Late
Mrs. Alice Siyanbola Ogundare, who taught me lots of words
of wisdom. I can never forget her telling me that 'Inu mimo
ja ju ogun 10'. This statement has been proven right in my
life many times over.
I sincerely appreciate the entire Osofisan family, and its
extensions, especially those here present. You accepted me
the way I am as a daughter and not as a wife. These include
Prof. Philip Babatunde Osofisan, my husband's older brother,
and his other siblings-Mr. Olusola Osofisan, now in the
USA, and Mrs. Dupe Ilori, a retired civil servant. They also
include their spouses and numerous relations all over the
world. I have indeed blossomed among you. I also appreciate
my in-laws - the Akinrindes; the Nwabuezes; and the
Usidames. Thank you for giving me three more lovely
daughters-Jumoke, Uzo and Ehi. I appreciate my brother
Engr. Leke Adedipe, retired staff of Shell Petroleum
Development Company, who accompanied me to my new
marital home in 1974; my cousin, Hon. Mrs. Bunmi
Adelugba, a Commissioner in Ekiti State, and my brother,
Hon. Adegboyega Adedipe, the present Ondo State Secretary
of the Action Congress of Nigeria-who were the first "two
children" I raised. I am proud of you. Words are not enough
to express my deep love for late Funmi Jeyifous aka "Funmi
sharp" my late adopted daughter, may your soul rest in
perfect peace. I appreciate the dedication of my wards, Elijah
and Anu as well as my nephew Akinyemi who are currently
staying with me.
I thank all the teachers who have taught me right from
primary school to the university. I wish to appreciate most
especially, Miss Rosaline Jane Pelly, my principal at
Fiwasaiye Girls' Grammar School Akure, for teaching her
students time management. I thank Dr. Olumide Kuti, the
then Comprehensive High School Aiyetoro Guidance
Counsellor, for asking me to change my A' Level subjects
from Physics, Chemistry and Biology to Physics, Chemistry
52
and Mathematics so that I could study Computer Science,
even when Computer Science was not studied in Nigeria then
in 1968. I appreciate Dr. Isaac Odeyemi for making my
studying Computer Science possible. I remember vividly that
day at the University of Ife Computing Centre in 1972, when
I visited you and you requested me to change my course from
Physics honours to Computer SciencelEconomics that was
just starting then in September 1972. I remember the argu-
ment that took place between you and Prof. Abiodun Oluwole
(NNOM) who would rather have me study Computer
Science/Mathematics. My innate spirit chose Economics even
though I knew nothing about the subject.. That was how I
started a journey into the world of two unknowns-one that
controlled world economy then and the other that would
move the world economy into that of knowledge. It is
amazing!
My gratitude goes to my PhD Supervisor, Prof. Bishop
Adebayo Dada Akinde, for facilitating my goal of having a
PhD degree in Computer Science. As I said earlier, little did
we know that history was in the making that day in your
office in 1984, that I would become the first female to have a
PhD degree in the Faculty of Technology, Obafemi Awolowo
University Ile- Ife, as well as the first female to have a PhD
degree in Computer Science in Nigeria. God's hands were
obviously working behind the scene.
I thank those who in one way or the order contributed to
the success of my PhD work, among whom I must mention
Prof. Biodun Jeyifo, my husband's bosom friend, and my
cousin, Mr. Tubosun Adedipe living in USA, especially for
paying for my subscriptions to various international journals,
conference proceedings and books. I also thank the Joint
Komputer Kompany (JKK), DEBIS Nig Ltd., and the then
DATAMATICS owned by Chief Dapo Sarumi, former
Minister of the Federal Republic of Nigeria, for sponsoring
my attendance at the io' World Computer Congress in
Dublin in 1986; and the International Council for Computer
Communications (ICCC) for sponsoring my attendance at all
ICCC conferences between 1986 and 1989.
53
I thank all my classmates at the University of Ife, in
particular, the foundation set of Computer Science Education
in Nigeria - Yemisi Osiname (nee Odedeyi), Saka Fawole,
Femi Aladesulu, Funso Akinniyi, Gbadebo Adeleke, Charles
Ikemere, Kola Ogunlana, Gbenga Okusaga, and Taiwo
Yakubu, for your comradeship. I thank every member of
Fiwasaiye Old Girls' Association (FOGA), most especially,
Wura Ajibola for always being there for me. .
My special appreciation goes to Prof. Akinbo Adesomoju
for being instrumental to my moving from the Polytechnic
Ibadan to the University of Ibadan in 1997, and for believing
in my ability. Prof. Funso Olorunsogo, my technical class-
mate at All Saints' Primary School, Osogbo, thank you for
encouraging me to stay on in UI when the road was very
rough and the environment extremely hostile. Your godly
statement-"Y ou are not going back to Egypt"-struck a
chord within me and I resolved to cross the Red Sea, fight all
the 'Canaan's kings' and pull down the walls of Jericho with
God by my side and He, in His infinite mercy, saw me
through. May God continue to use you both as vehicles of
grace. To Professors O. Osonubi, L. A Hussain, AT. Hassan,
R.A Oderinde and T. Odiaka, I say thank you all for the roles
you have played in my life. And to all the 'Canaan's kings', I
also say thank you.
To every member of staff and student of Computer
Science Department, past and present since I joined the
services of UI, I say God bless you all, especially my
postgraduate students and my colleagues collaborating with
me in research, who have been sources of inspiration to me.
You will all reach the top of your profession in Jesus' name
(Amen).
Many thanks to my professional colleagues in the Com-
puter Professional Registration Council of Nigeria (CPN), the
Nigeria Computer Society"(NCS) and the Nigerian Institute of
Management (NIM) (chartered). I thank most especially,
Prof. Senator Iya Abubakar and Senator Joy Emordi for the
role they played while I was the President and Chairman of
CPN. During my tenure, the Council was transformed from a
54
one-office affair in Lagos, with just five members of staff, to
a nationwide organization with nine zonal offices in Lagos,
Abuja, Owerri, Jos, Ibadan, Port Harcourt, Gombe, Kano ar-d
Yola, and a total staff membership of 78! The Council was
also able to acquire 5.404.03sq.mtr of land in Abuja,at least
one vehicle for each zonal office, and a modem e-library, all
within the 4 years of my stay in office. To God be the Glory.
Great things He has done.
My appreciation goes to my brothers and sisters in
Christ, at The Chapel of the Resurrection, University of
Ibadan, most especially members of the Resurrection Lilies;
to the entire congregations of the Church of Transfiguration,
Anglican Communion at Ikolaba Ibadan and St. David's
Cathedral Akure, as well as every member of the Bible Study
Fellowship, Ibadan EW and EM.
To my Children Wale ("Oshofs 1"), Yerni
("Shoffybrown"), Akin ("Oshofs3") and Tomi ("SisOshofs"),
MamOshofs says you are precious and dearly loved and my
little Oshofs -Toni, Mayo, Rayoke, Todun and Lase-you are
bundles of joy. My darling husband, Ferni "the Elereko", you
are the love of my life, my pillar of strength and the husband
of my youth. You have supported me all through my entire
academic and professional career. So I say to you:
Like an apple tree among the trees of the forest
is my lover among the young men.
I delight to sit in his shade,
And his fruit is sweet to my taste.
He has taken me to the banquet hall,
And his banner over me is love
-Song of Songs 2:3-4
Now to God Immortal, Invisible, the only Wise, be all
honour and glory now and for evermore.iMr. Vice-Chancellor
Sir, ladies and gentlemen, I thank you for listening.
55
References
Adagunodo, E.R. Akinde, A.D. & Osofisan, A.O. (1999) Protocols
specification & verification: A graph - Theoretic model
approach. The Journal of Computer Science and its Applica-
tion. 7(1) June, pp. 55 - 64. , '
Akinde, A.D. & Osofisan, A.O. (1994) A conceptual design and
specification of a data handling computer network. lfe Journal
of Technology 4(1) pp. 49 - 53.
Bartee, T.C. (1985) Data communications, networks and systems .'
New York: Howard W. Sams & Co.
Berry, J.A. & Linoff, G. (1997) Data mining techniques. New
York: John Wiley & Sons,Inc.
Cohen, Lawrence, and Marcus Felson (1979) Social change and
crime rate trends. American Sociological Review 44: ·S88.
doi:1O.2307/2094589. ~
Cukier Wendy, Eva J. Nesselroth, Susan Cody (2007) "~eRre,
Narrative and the Nigerian Letter in Electronic 'Mail ",
Proceedings of the 40th Annual Hawaii International Con-
ference on System Sciences HICSS 2007: 70.
Date, C.J. (2001) An introduction to database systems. Seventh
edition. New Delhi: Addison-Wesley.
Fausset, L. (1994) Fundamentals of neural networks. Upper Saddle
River, NJ: Prentice Hall.
Fayyad, U., Piatetsky-Shappiro, G. and Smyth, P. (1996) From data
mining to knowledge discovery in databases AAAI 1 pp. 37-54.
Fayyad, U. Wiesse, A. and Grinstein, G.G. (2002) Knowledge
discovery and information systems. San Fransico: Morgan
Kaufmann publishers.
Federal Govt, Nigeria (1982) National Policy on Education
Revised.
Ferrucci, D. & Lally, A. (2004) "UIMA: An architectural approach
to unstructured information processing in the corporate
research environment" Natural Language Engineering 10(3/4),
pp 327-348. .
Frawley, W., Piatetsky-Shapiro, G., & Matheus, C. (1992)
Knowledge discovery in databases: An overview. AI Magazine,
Fall 1992, pgs 213-228.
Gbenga, S. (2007) "The growing Menace of Cyber Crime in
Nigeria-Causes, Effects & Solutions"
http://events.tigweb.org/I2657
·56
Goldberg, D.E. (1989) Genetic Algorithms. Reading M.A.:
Addison Wesley.
Haykin, S. (1994) Neural networks: A comprehensive foundation.
New York: Macmillan.
Hofstetter, F.T. (2005) Internet technologies at work. Boston:
McGraw Hill Technology Education.
Horak, R. (1999) Communications systems and networks, 2nd ed.
Chicago: M&T Books.
Igwe, e. (2007) ''Taking Back Nigeria from 4 I9: What to do about
the Worldwide E-mail Scam - Advance Fee Fraud".
Bloomington: iUniverse.
Longe, O. & Osofisan, A. (2011) On the origins of advance fee
fraud electronic mails: A technical investigation using internet.
Information Systems 3(1), pp 17-26.
Lucey, T. (1997) Management information systems. London: DP
Publications.
Mannino, M.V. (2001) Database application development &
design. Boston: McGraw Hill.
Nisbert, R., Elder, J. & Miner, G. (2009) Handbook of statistical
analysis & data mining applications. Elsevie Inc. Amsterdam.
O'Brien, J.A. (2003) Introduction to information systems. Irwin:
Boston: McGraw Hill.
Olamiti, A.O. & Osofisan, A.O. (2009a) "Academic background of
students and performance in a computer science programme in
a Nigerian university". European Journal of Social Science
9(4),2009.
_____ (2009b) Experimental comparison of missing value
treatment methods in students' enrolment data. European
Journal of Scientific Research 33(4).
Osofisan, A.O. (1996) Designing and writing Web documents: A
case study of NUMAPS. INFOTECH (no. 2) pp 184 - 198.
(2000a) A virtual data warehouse architecture for
academic environment. Journal of Science Research. 6(2) pp
122 - 125.
____ (2000b) "Creating Synergy on the Net" In Computer
Association of Nigeria Conference Proceedings on Deployment
of Telematics Systems: Trends, Techniques and Tools, Vol. 11,
Eds. e.0. Uwadia, H.O.D. Longe, and A.D. Akinde, pp. 255 -
261.
57
Osofisan, A.O. (2001) Performance trace analysis of abstract co-
operative caching in web servers. The Journal of Computer
Science and its Application. 8 (1) June pp. 13 - 16.
____ (2002) "Application of Data Warehousing Technology
to Economic Intelligence: A case Study of University of
Ibadan" In Information and Communication Technology
Applied to Economic Intelligence ICTEI'2002 Institut National
De Recherche en Informatique et en Autornatique, INRIA,
Lorraine. pp 21- 3 1.
Osofisan, A.O. & Fils, D. (1996) "Network for Use in Mathematics
and Applied Physical Science". Computer Association of
Nigeria Conference Proceedings on Trends in Computing:
Hardware, Software and Networking. Vol. 7. Eds. E.R.
Adagunodo, A.D. Akinde, L.O. Kehinde and T. Onabanjo, pp.
85 - 115.
Osofisan, A.O. & Idowu, S.A. (2009a) "Content and Service
Delivery Techniques on the Internet: Trend of Development
and Characterization" IJAS Conference for Academic
Disciplines held on the Suffolk University campus in down-
town Boston (June 22-25,2009).
(2009b) QoS-based request-routing among peering
content distribution networks (CDNs) in a Virtual Organization
(VO)-Based model IJCSIS 1(1), May.
____ (2011) QoS-based objects replica placement among
peering Content Distribution Networks (CDNs). IJCSI 8(2),
March. ISSN (Online): 1694-0814 www.IJCSI.org pp 490-500.
Osofisan, A.O., Inanga, E.C. and Adeyemo, 0.0. (2004) Bank
asset portfolio management model using goal programming
techniques. ICMCS. 1, pg. 171-195.
Osofisan, A.O., Komolafe, O.P. & Akinola, S.O. (2005)
Discovering knowledge in road accident database. ICMCS 2.
pp 31-43,2005.
Osofisan, A., Ajayi, F., Adeoye, A., Akinola, 0., & Olarniti, A.
(2003) "Data Warehouse Application in Education Information
System: A Case Study of Junior Secondary Educational System
(Oyo State, Nigeria)" In Intelligence Economique: Recherches
& Applications Formation and Communication IERA '2003 a
l'INIST-CNRS, Nancy, France, Institut National De Recherche
en Informatique et en Automatique, INRIA, Lorraine. pp 79-
99.
58
Osunade, 0., Adeyemo, A. & Osofisan, A.O. (2001) "NETBILL:
An Ecommerce Solution for Sale of Information" In Computer
Association of Nigeria Conference Proceedings on Impact of e-
Commerce on National Economy and Development (E-
Commerce-ned). Eds. S. Juwe, H.O.D. Longe and A.D.
Akinde, Vol. 12 pp. 165 - 180.
Profgame Blogspot (2007) "Smile out of Poverty - Cyber Crime in
Nigeria - A Sociological Analysis:"
http://profgame.blogspot.com/2008/08/cyber-crime-in-nigeria-
sociological.html
Schwartz, M. (1977) Computer-communication network design &
analysis. Prentice-Hall Inc.
Speckt, D.F. (1991) A generalized regression neural network. IEEE
Transactions on Neural Networks 2(6), pp 568-576.
Tanenbaum, A.S. (1996) Computer networks. New Jersey:
Prentice-Hall International.
Wiederhold, G. (1981) Database design. Johannesburg: McGraw-
Hill International Book Company.
Wikipedia (2011) Computer network topologies.
http://www.en. wikipedia. org 6/8/2011.
Witten, 1. H. & Frank, E. (2000) Data mining: Practical machine
learning tools and techniques. New York: Morgan Kaufmann.
59
BIODATA OF PROFESSOR
ADENIKE OYINLOLA OSOFISAN
Adenike Oyinlola OSOFISAN was born on 11thMarch 1950
in Osogbo, to the late Chief Honourable Josiah Orisabinu
Adedipe, the Elemo of Akure and Mayegun of Osogbo and
late Chief (Mrs.) Felicia Fehintola Adedipe (nee Ogundare),
the Akuwajo of S1.Paul Anglican Church Emure Ekiti.
She started school at All Saints Primary School, Osogbo
(1956-59); and then transferred to Saint Steven's Primary
School, Akure (1960-61). Then, for her secondary education,
she went first to Fiwasaiye Girls' Grammar School in Akure
(1962-67) where she finished in First Division and as the first
student of the school to have a grade of distinction in
Mathematics. Then for her Higher School Certificate educa-
tion, she attended the Aiyetoro Comprehensive High School,
spending an extra year (1968-70) in order to convert from
Arts to Science.
Adenike was a pioneer student of the department of
Computer Science of University of Ile-Ife (now Obafemi
Awolowo University), and graduated with a 2nd Class Upper
Division Honours Degree in Computer SciencelEconomics.
She also has a Masters degree in Computer Science in 1979
from Georgia Institute of Technology in Atlanta and PhD
Computer Science in OAU to become the first Nigerian
woman to hold a PhD degree in Computer Science and the
first female to have a PhD degree from the Faculty of
Technology of Obafemi Awolowo University Ile-Ife. She also
has the MBA (Finance & Accounts) of University of Ibadan
in 1993, where she graduated with unprecedented nine
distinctions.
Prof. Osofisan worked at the University of Ibadan as a
Programmer between 1976 and 1978. She started her teaching
career as Lecturer II at The Polytechnic, Ibadan in 1979. She
was promoted Lecturer I in 1981; Senior Lecturer in 1984;
and was appointed, Head Department of Computer Science in
1986, a position she held till 1995. While in the saddle, she
60
was promoted to the position of Principal Lecturer in 1990.
Through dint of hard work, Prof. Osofisan rose to the highest
position of Senior Principal Lecturer at the Polytechnic,
Ibadan in 1993. She was head of Computer Studies for nine
years (1986-1995) and Dean of Faculty of science for two
years (1995-97) before she transferred her services to ill as a
Senior Lecturer. She was promoted Reader in 2003 and
Professor in 2006 to become the first black African Female
Professor of Computer Science in Sub-Saharan Africa.
Professor Osofisan was the Pioneer President, Nigeria
Women in Information Technology (IT) and first female
President and Chairman of Council, Computer Professionals
Registration Council of Nigeria (CPN). Prof. Osofisan was
the longest serving female member on the National Executive
Council of Nigeria Computer Society.
She was a Council Member of the Nigeria Institute of
Management (NIM) for two years, Branch Chairman of NIM,
Ibadan chapter. Currently, she is a member of Governing
Council, National Mathematical Centre, Ahuja; Member of
the Governing Council, Achievers University Owo, Ondo
State; Treasurer Nigeria Association of Women in Science,
Mathematics, Engineering and Technology (NAWSTEM),
Oyo State Branch and Patroness of the Nigerian Association
of Computer Science Society (NACOSS). She belongs to
several professional bodies such as Board of Trustees of
Nigeria Internet Registration Association (NIRA). She is a
Board Member OMATEC Computers; Coordinator - Four
Nigeria Institutions of Higher LearninglNancy 2 University,
France Collaboration Programmes. She is also a Director of
Computerize Nigeria Project.
Prof. Osofisan has been the recipient of many national
and international prizes, scholarships and fellowship awards
such as, the International Council for Computer Communi-
cation Fellowship (ICCC) to New Delhi and Bombay (now
Mumbai) India, University Development Linkage programme
(UDLP), International Women in Science and Technology
(IWISE) Iowa USA, UNESCO, South Africa, United Nations
61
University (UNIST), and University of Technology
Melbourne, Australia among others. Prof. Adenike Osofisan
is a Fellow of the Nigeria Institute of Management (NIM)
chartered; Fellow Nigeria Computer Society (NCS); Life
Member, Nigerian Economic Society, Member, Computer
Society of IEEE; and Member Association of Computing
Machinery (ACM).
She has received numerous awards for the development of
Information Technology and Computer Science Education in
Nigeria, such as Meritorious Award (NACOSS) 1995,
Women of Achievement Award, (NACOSS), 1996; Digne
Ambassadeur, Computer Science Students Association
Obafemi Awolowo University, Ile-Ife 1997; Builder of Com-
munity Award, Akure 1998; Branch Merit Award NIM, 2000;
Award of Excellence, The Compronians, 2001; The Woman
of Merit Gold Award, 2002; Role Model Award, NAPTAN,
Ondo State Chapter, 2006; Woman of Distinction Merit
Award, 2006; Outstanding Dedication to the Growth and
Development of NACOSS, 2006; NCS Award towards
advancement of IT in Nigeria, 2007; the 2007 IT Woman
Personality of the Year; NASU University of Ibadan Award
towards the progress and development of Humanity, 2007;
ICT Female Legend, 2008; Distinguished National Award of
Excellence 2008; TITAN of Technology, Nigeria Award
2008; Information Technology Champion, ITAN 2009,
Fiwasaiye Old Girls Association (FOGA) Golden Jubilee
Merit Award for being the first Nigerian Professor of Science
and meritorious service to FOGA and Humanity February,
2010 and ICT for Africa 2010 Lifetime Achievement Award
in Recognition for Excellence in the Research and Practice of
ICT in Africa, March 27,2010.
Professor Osofisan has attended many conferences,
seminars both at national and international level presenting
papers of her various research works. She has over 70 publi-
cations authored in national and international journals and
conference proceedings.
62
Prof. Adenike Osofisan is a devout Christian, serving the
Christian Community in various positions such as Teaching
Leader, Ibadan Evening Women Bible Study Fellowship
International; Member Parish Church Council, Church of the
Transfiguration, Anglican Communion, Ikolaba Ibadan;
Bishop Nominee for the 2011-2013 Synod of The Ibadan
Diocese, the Church of Nigeria Anglican Communion. She
was the former Secretary and the former President of The
Resurrection Lilies Society, Chapel of The Resurrection,
University of Ibadan.
Prof. Adenike Osofisan is married to a literary giant, Prof.
Femi Osofisan of the Department of Theatres Arts, University
of Ibadan and former General Manager, National Theatre
Iganmu Lagos. The marriage is blessed with children, and
grandchildren.
63