publication 2 libre

International Journal of Green Computing, 1(1), 1-15, January-June 2010 1

Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global

is prohibited.

Keywords: Database, Neural Network, Record, Result Computation, Tertiary Institution

IntroductIon

Student enrolment in tertiary institutions is

increasing at a very alarming rate. The increase

in students’ population over the years has made

the work of administrative officer in charge of

processing students’ result a very tiresome ex-

ercise to deal with. In Ahmadu Bello University

which records large number of students turnout

and yet admit more on yearly basis, processing

of students’ academic record represents very

significant challenge as it tends to require a great

deal of human involvement thereby increasing

the cost and delay associated with it.

Students’ academic record refers to the vital

information relating to a student admission, and

design and Implementation

of Students’ Information

System for tertiary Institutions

using neural networks:An open Source Approach

Obiniyi Afolayan Ayodele, Ahmadu Bello University – Zaria, Nigeria

Ezugwu El-Shamir Absalom, Ahmadu Bello University – Zaria, Nigeria

AbStrAct

This paper identiies the causes associated with delays in processing and releasing results in tertiary institutions. An enhanced computer program for result computation integrated with a database for storage of processed results simpliies a university grading system and overcomes the short-comings of existing packages. The system takes interdepartmental collaboration and alliances into consideration, over a network that speeds up collection of processed results from designated departments through an improved centralized database system. An empirical evaluation of the system shows that it expedites processing of results and transcripts at various levels and management of and access to student results on-line. The technological approach for the implementation of the proposed system is based on open source solutions. Apache is used as Web server extended with PHP for server side processing. In recognition of the conidentiality of data contained in the system, communication networks are protected with open-ssl library for data encryption and role-based authentication. This system increases eficient service delivery and beneits both administration and students.

DOI: 10.4018/jgc.2010010101

2 International Journal of Green Computing, 1(1), 1-15, January-June 2010


is prohibited.

academic performance at the host university.

Students’ record can be described in terms of

their contents such as student bio-data which

includes full name, matriculation number, gen-

der, local government and state of origin. Then

on performance references we have the courses

registered, grade point average and the general

academic performance history of the student.

The academic records of students are

the property of the university, thus it is the

sole responsibility of the school to constitute

policies regarding consistency in the kind of

information collected and recorded. Students’

academic record usually provides both student

and staff with numerous services designed to

assist them in attaining their academic goals and

management objectives respectively. These may

include generation of individual students result,

transcripts and, publishing academic timetable

for each semester.

The proposed system for academic record

and grades processing is arranged and designed

to simplify the work in students’ academic

record system, due to its complexities. This

paper demonstrates a model and implementa-

tion of students’ information system for tertiary

institutions using neural network with open

source approach.

The paper is organized as follows: the

course evaluation is presented and the tech-

nological approach to the proposed system is

then discussed in detail. The implementation

case of the new system is also discussed, and

conclusions are presented.

EvAluAtIon of courSE GrAdInG SyStEm

The course credit system is a system in which

the syllabus of a subject in a degree program

is divided into courses arranged in progressive

order of difficulty or in levels of academic

progress that is 100 to 400 levels for faculty of

sciences. The course credit system is flexible

enough to accommodate both strong and weak

students. It minimizes duplication of courses as

it encourages inter-departmental collaboration

in curriculum planning, formulation and syl-

labus review. It is also possible for a student to

defer a semester or session for a genuine reason

either on medical or financial grounds. This is

because a credit earned is never lost.

In the ideal system, all courses are supposed

to be mounted every semester with the dual

purpose of allowing the exceptional students to

graduate before time and the weaker one prog-

ress at their own pace. This, however, may be

difficult here as a result of the staffing situation.

credit units and credit load

In the course credit system, courses are assigned

weights called credit units depending on how

many contact hours are required to complete

the course in a semester, for example, a one

credit course requires fifteen hours of lecture

per semester. A credit load is the total number

of credit units registered per semester by the

student. In faculty of science, the minimum

credit unit per semester is twelve and maximum

is twenty four.

categorization of courses

The courses within the faculty are categorized

as follows:

a. Core Courses: Courses that are funda-

mental to the degree in view and usually

offered by the department constitutes not

less than sixty percent of all credit units

that the student must earn in fulfillment

of the requirements for graduation. In

computer science section of Mathematics

Department, for example, these courses in-

clude among others COSC201, COSC203,

COSC207, COSC305, COSC401, and

COSC405.

b. Cognate Courses: Courses from related

fields that is necessary for an understand-

ing and appreciation of the student’s major

field. A computer science student, for

example, takes course like MATH205, is

a cognate course.



is prohibited.

c. General Study Courses: Courses that is

common and compulsory to every new

undergraduate student of the university.

These courses cover such areas as National-

ism (GENS103), Moral Philosophy (GENS

201), Communication Skills (GENS

103) and History of Scientific Thinking

(GENS107).

d. Elective Courses: These are further clas-

sified as:

i. Courses from which a student is made

to select one or more from as the case

may be. These are normally decided by

the department concerned but should

not constitute more than fifteen percent

of all the credit units to be earned by

the students to qualify for graduation.

ii. Unrestricted elective courses are

chosen by a student according to his/

her wish. These can be taken from any

department within the university, but

should not constitute more than fifteen

percent of all credit units to be earned

by the student to qualify for graduation.

A special categorization of courses is

whether they are prerequisites or not. A prereq-

uisite course must be passed to qualify to take

some other course or courses in another level,

for example, in Mathematics Department; one

must pass MATH105 in 100 level in other to be

able to offer MATH201 in 200 level.

Grade Point System

This is the point scoring system in the course

credit system. Raw score are converted to let-

ter grades weighted accordingly as in Table 1.

The weight accorded each grade is called

credit point. For example, a student who scores

‘A’ in a three credit unit course has credit point

of fifteen in that course.

The following terms are prevalent in the

grading system:

a. Total Credit Unit Registered (TCUR):

Total credit registered in a semester or

session.

b. Total Credit Unit Earned (TCUE): Total

credit units earned in a semester or session.

In sciences, a student must earn at least

120 credit units to qualify for graduation.

The TCUE is also used in level placement

as in Table 2. Students in Faculty of Sci-

ence graduates in 400 level while those

of Faculty of Medicine and Vet Medicine

graduates in 600 level.

c. Total Credit Point (TCP): Total credit point,

that is sum of all credit points scored in a

semester or session.

d. Grade Point Average (GPA): The ration to

TCUR for the semester.

e. CGPA is the Cumulative Grade Point Aver-

age of the student courses.

Table 1. Point scoring system session (Faculty Orientation Committee, 1996)

Raw Score Letter Grade Weight

70-100 A 5

60-69 B 4

50-59 C 3

45-49 D 2

40-44 E 1

0-39 F 0



is prohibited.

review of the Existing System

The existing system is a file-based application

package written in Fortran 77 programming lan-

guage. It is not user friendly and has no end user

interactive forms; the operational environment

is command line like setting. The file-based

system has a number of inherent problems as-

sociated with it and does not meet the required

capability needed to run a large number of data,

as the case may be for a University operational

data bank. The existing system is best described

as system, in which data definition is embed-

ded into the application program, rather than

it being defined independently from the body

of the application.

modElInG wIth nEurAl nEtwork

Neural Networks are mathematical constructs

that emulate the processes people use to rec-

ognize patterns, learn tasks, and solve problem

(Schultz K, 2002). Neural Networks are usu-

ally characterized by the number and types of

connections between individual processing

elements, known as neurons, and the learning

rules used when data is presented to the net-

work. Generally, neural network is considered

to be a great data modeling tool, reason being

that it has the intricacy and capability of cap-

turing and representing complex input/output

relationships.

There are different numbers of neural

network model applicable in different areas

of scientific research, but the most commonly

used neural network model is the multilayer

perception. This model type is often referred

to as a supervised network, which requires

desired output for its learning. The pursuit

of this network model is to develop a model

which correctly maps input to the output using

historical data so that the model can then be

used to produce the output when the desired

output is unknown.

why use neural networks?

Neural networks, with their remarkable ability to

derive meaning from complicated or imprecise

data, can be used to extract patterns and detect

trends that are too complex to be noticed by

either humans or other computer techniques. A

trained neural network can be thought of as an

“expert” in the category of information it has

been given to analyze. This expert can then be

used to provide projections given new situations

of interest and answer “what if” questions. Other

advantages include:

a. Adaptive learning: An ability to learn how

to do tasks based on the data given for

training or initial experience.

b. Self-Organisation: An Artificial Neural

Network (ANN) can create its own organi-

sation or representation of the information

it receives during learning time.

c. Real Time Operation: ANN computations

may be carried out in parallel, and special

hardware devices can also be designed and

manufactured which take advantage of this

capability.

d. Fault Tolerance via Redundant Information

Coding: Partial destruction of a network

leads to the corresponding degradation

of performance, however, some network

capabilities may be retained even with

major network damage.

Table 2. Expected total credit unit earned in a session (Faculty Orientation Committee, 1996)

1. 200 level: 24 credit units







is prohibited.

network layers

The commonest type of artificial neural network

consists of three groups, or layers, of units: a

layer of “input” units is connected to a layer of

“hidden” units, which is connected to a layer

of “output” units.

a. The activity of the input units represents

the raw information that is fed into the

network.

b. The activity of each hidden unit is deter-

mined by the activities of the input units

and the weights on the connections between

the input and the hidden units.

c. The behaviour of the output units depends

on the activity of the hidden units and the

weights between the hidden and output

units.

This simple type of network is interesting

because the hidden units are free to construct

their own representations of the input. The

weights between the input and hidden units

determine when each hidden unit is active, and

so by modifying these weights, a hidden unit can

choose what it represents (Gottfredson, 1999).

There and multi-layer architectures. The

single-layer organisation, in which all units

are connected to one another, constitutes the

most general case and is of more potential

computational power than hierarchically struc-

tured multi-layer organisations. In multi-layer

networks, units are often numbered by layer,

instead of following a global numbering.

Architecture of neural network for Student registration and result Processing

There are several task associated with the design

of a neural network based model among which

include selection of input variables, design of

network structures, training and validation

(Tecuenche & Uwadia, 2006).

The neural network model shown in Figure

1 is a representation of the stages in processing

activities at each layer. A set of mathematical

model is chosen to describe the various trans-

formations that take place at each stage of the

input layer, hidden layer and output layer.

Input layer representation

The input layer consists of sets of input variables

comprising of semester registration, course reg-

istration, and CA/Exam raw scores. The course

credit unit registered is represented as CC. First

and second semester courses are indexed with

Figure 1. Architecture of Neural Network for students’ registration and Result Processing



is prohibited.

odd numbers 1, 3, 5, 7, 9… and even numbers 0,

2, 4, 6, 8, … respectively. A student is expected

to register for a maximum of 48 credit unit in

100 level. Thus our first equation for the total

number of credit unit expected at the end of first

year is shown in equations 1, 2, and 3.

CCi

i

N

=

∑1

≤24 (1)

CCj

j

N

=

∑1

≤24 (2)

CCi

i

N

=

∑1

+ CCj

j

N

=

∑1

≤48 (3)

where N = total number of courses registered

for a particular semester

hidden layer

The input variable for the hidden layer is ob-

tained from the output of the first layer. The

function of this layer is to compute credit point

and grade point average.

TCP CG CCi i

i

N

==

∑ *1

(4)

GPA

CG CC

CC

i ii

N

ii

N= =

=

∑

∑

*1

1

(5)

where CCi is the Course’s Credits, N is the total

Number of courses, CGi is the Course Grade

Point of the course awarded to a student based

upon the course credit system evaluation policy.

output layer

The output layer produces the desired outputs

resulting from the computation and evaluation

process of the hidden layer. These outputs in-

clude student processed results and transcripts.

These results are generated automatically either

by group or individual based on the choice of

the examination officer.

using Artificial neural network to Predict Student Academic Performance

Every neural network possesses knowledge

which is contained in the values of the connec-

tions weights. Modifying the knowledge stored

in the network as a function of experience im-

plies a learning rule for changing the values of

the weights. Consider the information stored in a

weight matrix W of a neural network. Learning

is the determination of the weights, following

the way learning is performed; we adopted the

adaptive networks method of learning in neural

network which are able to change their weights,

that isdW/dt≠0 inpredictingtheacademicperformance of students.

There are several contributing factors

which could affect students’ performance in

class. These may vary from financial constraint,

brain drain in the school system, inadequate

teaching and learning aids, poor class atten-

dance, low IQ and lack of seriousness on the part

of students. Among these several key options,

class attendance, class participation, number

of hours invested in extra studies, and level of

Table 3. Class attendance rating

Attendance Percentage Range Assigned Weight

75-100 4

50-74 3

25-49 2

0-24 1



is prohibited.

students IQ is considered to be the predicting

variables as shown in Figure 2.

Prediction variable for class Attendance

In every school system, student attendance is

usually considered very vital and taken as a

major criteria or prerequisite to which students

are expected to meet a certain percentage before

been allowed to seat for an examination. A

minimum of 75% attendance as mandated by the

university system was adopted for the purpose

of this research. Generally, students with very

low class attendance say between percentage

ranges of 0-70% should expect lower perfor-

mance compared to those with higher class

attendance rating ranging from 75-100%. Also

assigned to each of these percentage ranges is

a certain performance weight level for which

a student must meet in other to pass a certain

course as displayed in Table 3.

Prediction variables for class Participation

History has it that student who are usually very

active in class tends to also perform better in

their semester examination. Excellent class

contribution by students in terms of class ex-

ercise helps them further understand the basic

concepts and applications of such course and

at the end enhances them for good grades.

Some key standard yardsticks are applied for

measuring this variable as displayed in Table 4.

Figure 2. Neural network model for student academic performance prediction

Table 4. Class participation rating

Class participation Assigned Weight

Excellent 4

Very Good 3

Good 2

Poor 1



is prohibited.

Prediction variables for time Invested on Private Study

An average human brain is supposed to rest for

at most 8 hours per day, and the time chosen

for resting period now depends on individual

planning. Thus if there is proper time manage-

ment, students would have enough time left for

private study and extracurricular activities later

in the day. Table 5 shows the invested time of

study by the students.

Prediction variables for Student IQ level

Student with very high IQ (Intelligent Quotient)

tends to perform better than their colleagues

with lower IQ in brain storming exercise. The

average IQ rating by standard is 100. In most

cases, individual having IQ less that 100 would

not perform up to expectation, while individual

with IQ greater than 100 are expected to perform

intelligently.

Based on Table 6 of student IQ level rating,

it was possible to generate the following two

mathematical equations. Let n be the number of

predicting factors and numeric value 4 denote

the maximum assigned weight on the predicting

variables per course.

GPVP IQ PV

PV

kk

n

kk

n= =

−

=

−

∑

∑

241

1

1

1

* ( )* (6)

From equation 6, m is assumed to be the

number of registered courses by student in a

particular semester. Equation 7 gives the GPA.

GPAVP IQ V

PF

j

m

jkk

n

j

m

jkk

n=

+= =

−

= =

∑ ∑

∑ ∑

1 1

1

1 1

24

[ * ( ) ]*

(7)

where PV is the prediction value, IQ is the

students’ intelligent quotient and PF is the

prediction factor

tEchnoloGIcAl APProAch to thE ProPoSEd SyStEm dESIGn

In programming practices, it is not enough to

consider which programming language the pro-

grammer is familiar with, but which language

is most appropriate for the job at hand and has

Table 6. Student IQ level rating

IQ Weight

Very High 4

High 3

Average 2

Poor 1

Table 5. Invested time of study by students

Invested Study Hours Weight

12 hrs and above 4

8-11 3

5-7 2

0-4 1



is prohibited.

the capability and also the required facilities to

tackle such inherent problem.

The technological approach for the devel-

opment of the new system is based on WAMP

Server (Apache, MySQL, and PHP) open source

solution (Adewale, 2006). The web server has

what seems to be a simple straightforward job,

it sits there over a network, running on top

of the client’s machine listening to requests

that somebody on the web might make, and

it responds to those requests and serves out

the appropriate web pages. In reality, it is a bit

more complicated than that, and because of the

24/7 (twenty four hours seven days a week)

nature of the web, stability of the web server

is a major issue. PHP and MySQL are popular

pair for building dynamic web applications.

PHP is the most widely supported and used web

scripting language. and an excellent tool for

building web database applications (Williams

and Lane, 2004). PHP and MySQL as a pair

have several advantages among which include

flexibility for integration with html, suitable in

accomplishing complex projects and their abil-

ity to communicate well with each other. The

security feature of MySQL database makes it

suitable for use as internet database. MySQL

is a client/server database that consists of a

multithreaded SQL server that supports differ-

ent back ends, several different client programs

and libraries, administrative tools, and a wide

range of programming interfaces (Connolly &

Begg, 2002). It can hold up to 60,000 database

tables with approximately 5 billion records).

The new system is built around a three-tier

architecture model. At the base of the application

is the database tier, consisting of the relational

database management system that manages

the data users create, query and delete. This

database tier is implemented using MySQL

database server. Built on top of the database tier

is the middle tier, which contains most of the

application logic that has been developed using

PHP as the scripting engine. This middleware

works closely with the web server to interpret

the request made from the World Wide Web,

process this requests, interact with other pro-

grams on the server to fulfill the request, and

then indicate to the web server exactly what

to serve to the client browser. At the user end

is the client tier, usually browser software (In-

ternet Explorer, Mozilla Firefox, or Netscape

Navigator) that interact with the application.

Figure 3 pictorially describes the architecture

model of a dynamic web based application as

used in the new system.

dAtAbASE dESIGn for thE ProPoSEd SyStEm

The objective of this task is to prepare technical

design specifications for a database that will be

adaptable to future requirements and expan-

sion. The deliverability of the task includes

the resulting database schemas as shown in

Figure 4. A database schema is the structural

model for a database. It is a picture or map of

the record and relationships to be implemented

by the database.

unified modeling language (uml) model of the Proposed System

The Unified Modeling Language (UML) activ-

ity diagram depicted in the Figure 5 shows the

functional activities in totality that takes place

in the proposed Student Information System.

The various system interactive scenarios

that take place upon successful system login

are presented in Figure 6 using case diagrams.

a. Student Registration: Figure 6 presents an

interactive user friendly form to the admin

officer in charge of student registration

in every semester to enter the necessary

student data into the system database.

b. Course Registration: Students are ex-

pected to carry out their registration process

online. They can login into the system

using their registration number and a pin

code given to them by the departmental

examination officer as pass mark. After a

successful login, they can select the appro-

priate course for that particular semester.

There are rooms for course registration

preview, but modifications of courses



is prohibited.

registered are not allowed not until after

the course registration weeks are over. This

activities are as shown in Figure 7.

c. Result Processing: In Figure 8 examina-

tion raw scores are entered by the examina-

tion officer after which the result processing

is performed automatically by the system

based on the set conditions inbuilt on the

package. Results are generated based on

request and students are only allowed to

view individual result as a semester result

print out.

ImPlEmEntAtIon of thE ProPoSEd SyStEm

The new system has been designed to evaluate

the course’s grades in Ahmadu Bello University

Zaria, which is based on the above formula and

Table 1. The system consists of the following

html and PHP modules: Student registration,

course registration, score processing, GPA/

CGPA calculator and result generation. Figure

9 shows the new system user login page and a

successful login welcome page. Figure 10 shows

the system home page with all the command

buttons required to assist the user navigate the

systems module by module.

Figure 3. Intranet/Internet section of the proposed system operation



is prohibited.

Register Students module: This is a PHP script

entry form that accepts students’ bio data.

It contains the students’ registration num-

ber, the student full name, entry mode,

session admitted, date of birth, gender,

local government and state of origin.

TheAdd/Drop Course module: This consists

of two sub-module, the add course mod-

ule, which accepts students’ course detail

such as course code, course title and the

course credit unit. The second module

is the Drop course page which retrieves

data from the database server based on

administrative request, it allows the user

to carry out data modification and then

update to the appropriate database server.

Course Registration Module: This module

takes care of all the course registration

procedures. It comprises of a web entry

form which calls PHP file that validate

the students’ input, check the number of

credit units which should be within the

required range, check the lower level

courses not passed, check that the courses

have not been registered this particular

session before updating the record to the

database server.

Compute Result Module: the user is expected

to indicate the course, the number of test

Figure 4. System Database Schema



is prohibited.

Figure 5. Activity Diagram for the proposed Student Information System

Figure 6. Registration Functional Model



is prohibited.

Figure 7. Course Registration Functional Model

Figure 8. Result Processing Function Model

Figure 9. System user login page



is prohibited.

Figure 10. System Home Page with the transaction buttons

sat either one or two or an examination

and then a PHP file is called which load

the next web entry form that accepts the

students’ raw scores, validates and update

the database server.

Generate Report Module: In this module a user

is presented with web entry form where

a selection of the registration number of

which grade is to be computed is made.

The PHP file that is called for the process-

ing ensures that there is a raw score for

all the courses registered. This module

is also charged with the responsibility

of generating individual, group students’

result and transcript.

Check Result Module: A student is presented

with a web form where he/she is expected

to enter a valid registration number is ex-

pected to be keyed in and access is granted

to check for their semester result(s).

Admin Module: This is an html module that

enables the administrator to carry out

some administrative task such as record

modification, system update and some

basic system maintenance.

concluSIon And rEcommEndAtIonS

This paper presented a multitier academic man-

agement software models for managing students

academic records’ in tertiary institutions. The

application is designed using the technologi-

cal approach of WAMP (Web server, Apache,

MySQL and PHP) open source solution. A role-

based model access mechanism was built into

the new system to further boast up its security

stands because of the sensitivity of the system.

From our findings, we recommend that;

a. The existing system should be replaced in

order to enhance productivity of system

users and output performance.

b. Department Admin officer (examination

officer) and data entry operators needs to

be properly trained, and if need be sent on

short term computer courses so as to fully

exploit the advantages of the new system.

c. The university ICT department which

foresees the overall school networking

programme should be charged with the



is prohibited.

provision of efficient and effective in-

ternet services to different section of the

university.

d. There is need for the university authority

to have a robust academic databank that

will maintain both the university staff and

student record placed at maximum data

security.

rEfErEncES

Adewale, O. S. (2006). University Digital Libreary. In An Initiative to Improve Research, Teaching and Services (pp. 34-63). Akure, Nigeria: Adeyemo Publishing House.

Ashiby, W., & Kumar, L. (1996, August). A compari-son of neural networks and classical discrimination analysis in anticipating default among high yield bonds. In Proceedings of the Association for Informa-tion Systems 1996 American Conference (pp. 2-8).

Connolly, T. M., & Begg, C. E. (2002). Database System, A Practical Approach to Design, Implemen-tation, and Management (3rd ed.). Reading, MA: Addison Wesley.

Date, C. J. (2002). An Introduction to Database Sys-tem (3rd ed., Vol. 1). Reading, MA: Addison Wesley.

Faculty Orientation Committee. (1996). Faculty of Science Orientation Manual. Zaria, Nigeria: Ahmadu Bello University.

Gottfredson, L. (1999). The General Intelligent Factor. Retrieved January 5, 2009, from http://www.geocities.com/rnseitz/Definition_of_IQ.html

Jeffery, L., Whitten, D., & Bently, C. (1997). System Analysis and Design Methods (5th ed.).

Schultz, K. (2002). The Neural Network Teaching Computers to Think. Smart Computing, 2(1), 62–65.

Tecuenche, Y., & Uwadia, C. O. (2006). A Frame-work for a Decision Support System Using Artifical Neural Network. Journal of Computer Science and Its Applications, 12(1), 47–60.

Tyler, J. (2006). Class Participation Assessment Guidelines. Retrieved January 3, 2009, from http://www.brown.edu/departments/italian_Studies/dweb/pedagogy/partici-assessm.html

Williams, H. E., & Lane, D. (2004). Web Database Applications with PHP and MySQL (2nd ed.). Se-bastopol, CA: O’Reilly Media.

publication 2 libre

Documents