Predicting Student Performance Using Fuzzy Logic

ABSTRACT

This project aims at implementing fuzzy logic in predicting the end of first year performances of students using fuzzy logic based on previous or prior academic achievements presented during admission or registration process. The project involved modelling a fuzzy inference system (FIS) to predict the CGPA of students newly admitted into Computer Science department, Federal University of Technology, Minna. The predicted performances of each student were compared with their current GPA and 60% accuracy was achieved. The achieved result shows the capability of fuzzy logic to handle uncertainty, given that predicting or evaluating the academic performance of students involves handling uncertain or imprecise data, especially in decision making. Due to the competitive nature of securing admission into higher institutions of learning in Nigeria, I hereby recommend the use and implementation of the proposed modelled fuzzy inferential system in conducting admissions into Nigeria’s higher institutions in order to identify students that will perform well if granted admission.

TABLE OF CONTENTS

COVER PAGE

TITLE PAGE

APPROVAL PAGE

DEDICATION

ACKNOWLEDGEMENT

ABSTRACT

CHAPTER ONE

INTRODUCTION

1.1      BACKGROUND TO THE STUDY

• STATEMENT OF THE PROBLEM
• AIM AND OBJECTIVES
• SIGNIFICANCE OF THE STUDY
• SCOPE OF THE STUDY
• LIMITATIONS OF THE STUDY

CHAPTER TWO

LITERATURE REVIEW

• FUZZY SET THEORY
• WHAT IS FUZZY LOGIC?
• CHARACTERISTICS OF FUZZY LOGIC
• MEMBERSHIP FUNCTIONS
• LINGUISTIC VARIABLES
• AREAS OF APPLICATION OF FUZZY LOGIC
• REVIEW OF RELATED LITERATURE ON FUZZY LOGIC APPROACHES IN ACADEMIC PERFORMANCE PREDICTION
• SUMMARY OF RELATED LITERATURE

CHAPTER THREE

• FUZZY INFERENCE SYSTEM
• STRUCTURE OF A FUZZY INFERENCE SYSTEM
• RESEARCH METHODOLOGY
• INPUT-OUTPUT FUZZY INFERENCE SYSTEM
• MODELLING OF INPUT MEMBERSHIP FUNCTIONS
• MODELLING OF OUTPUT MEMBERSHIP FUNCTION
• DEVELOPMENT OF FUZZY RULES
• DATA COLLECTION
• SUMMARY

CHAPTER FOUR

• DISCUSSION OF RESULTS
• INTRODUCTION
• LOW STUDENT PERFORMANCE OUTPUT
• AVERAGE STUDENT PERFORMANCE OUTPUT
• HIGH STUDENT PERFORMANCE OUTPUT
• ENTRY MODE OUTPUT SURFACE PLOT
• UTME SCORE OUTPUT SURFACE PLOT
• SECONDARY SCHOOL RESULT OUTPUT SURFACE PLOT
• ENTRY AGE OUTPUT SURFACE PLOT
• DISCUSSION OF RESULTS

CHAPTER FIVE

• SUMMARY OF THE STUDY
• CONCLUSION

CHAPTER ONE

INTRODUCTION

1.0       BACKGROUND TO THE STUDY

The increasing demand for education in Nigeria especially tertiary education has been an issue in recent times. Government–owned tertiary institutions as well as those privately owned do not possess the capacity and manpower needed to accommodate the growing number of young school leavers hoping to secure admission into any of these tertiary institutions.

Huang (2013) posits that predicting and evaluating the academic performances of students has for a long time now been an interesting and important area of research in many academic disciplines. Evaluating students’ academic performances or achievement using the cumulative grade point average (CGPA) as a pointer is a typical practice in every tertiary academic environment. A high percentage of fresh students usually find themselves below the minimum grade point required at the end of their first year.

Lee (2008) pointed out that having a thorough understanding of a student’s prior achievement or knowledge is important and will go a long way in helping academic planners predict the future performance of such student.

Oladipupo (2012) stressed the importance of predicting and evaluating the academic performances of students. In his survey carried out in a privately owned university in Nigeria, the percentages of students below the 2.5 CGPA at the end of the 2007/2008, 2008/2009 and 2009/2010 academic sessions was alarming. This percentage gives a clear view of the declining nature of students’ academic performances in our tertiary institutions.

Thus predicting and evaluating students’ performances cannot be over-emphasised as it is a useful tool for academic planners, educators and students alike, as it gives an actual understanding of students’ weaknesses and proffer solutions.

Fuzzy logic is an extended branch of Boolean logic and a unique form of logic that deals with reasoning which is approximate other than fixed or exact and recognizes more than the simple true/false values. It is also a way of knowledge representation that is understandable to a computer. Fuzzy logic is a special type of logic in which a statement can be represented with degrees of truthfulness or falsehood. Zadeh (1996) describe fuzzy logic as computing with words in a way that words or linguistic variables are used to replace numeric values for reasoning and computing.

Zadeh (1996) also buttressed the importance of fuzzy logic and its usage when the data available is uncertain or imprecise to warrant the use of numbers or exact values and also when imprecision can be tolerated. Fuzzy logic due to its ability to handle imprecision and uncertainty has found its application in numerous fields of study.

The application of fuzzy logic in predicting and evaluating students’ academic performances gives effective and attainable solutions as the knowledge of fuzzy logic is suitably required when human evaluation is needed. Furthermore, researches show that fuzzy logic is a more essential technique to handle imprecision and uncertainty given that evaluating the prior knowledge or past achievements of students with the aim of discovering the risk of students’ failure involves dealing with uncertain and imprecise data.

Therefore, this project aims at implementing fuzzy logic in predicting students academic performances as the fuzzy approach is likely to offer an alternative way of handling imprecise data, especially in decision making.

• STATEMENT OF THE PROBLEM

The inability of Nigerian graduates to cope well in any working environment they find themselves has for long now raised a lot of concerns. This has been attributed to the inadequacies and flaws in the universities admission process. The criteria for students intake is no longer clearly defined hence the aim of admitting suitably qualified candidates who are likely to perform better both in school and in the real world after school has been defeated.

Obielumani (2008) pointed out that due to the multi-ethnic nature of the country and the increasing demand for quality education, the Nigerian federal government introduced the quota system to enhance the increasing access to admission opportunities for candidates who are from disadvantaged states in the country.

The quota system has been defeated as lobbying, bribery and the likes have created a flaw in the university admission process and the criteria for student selection is not specified and thus the rightfully qualified students are neglected.

In essence, evaluating students’ academic performances is based on the CGPA and is a regular practice in every tertiary institution. Most institutions consider a CGPA of 3.5 and above as an indicator of good performance. Sansgiry (2006) stated that the CGPA remains the most well-known factor or variable used by academic planners to assess progression in an academic environment. Various factors that have a likelihood of affecting students performances have been identified but despite its limitations in providing a clear view of the state of a student’s performance, the CGPA is just about the main instrument or factor that academicians use in evaluating students academic performances.

This project suggests that there is a need to combine other factors to predict the future performances of students, the end of first year performances to be precise, in order to enhance the evaluation of students admitted into tertiary institutions.

• AIM AND OBJECTIVES

This project aims at predicting the end of first year performances of students using Fuzzy Logic based on previous or prior academic achievements presented during admission process.

The objectives of the study are;

1. To model a Fuzzy Inference System (FIS).
2. To predict the CGPA of students using the Fuzzy Inference System (FIS).
3. To use the predicted CGPA as a yardstick for student performance evaluation.
   • SIGNIFICANCE OF THE STUDY

With the increasing competition among higher institutions of learning to admit outstanding students, and with the limited space but highly competitive nature of admission process, more attention is on how to grant admission to only the best students and increase student retention rates and academic performance as well as producing quality graduates that can cope in any working environment and also the number of degree completion.

Also, there is a growing interest and concern about the problem of poor performance and school failure and the determination of its main contributing factors. So far, there is limited number of researches which examine the influence of these contributing factors on first year academic performances of students in tertiary institutions.

This project focuses on designing a predictive system framework using fuzzy logic to serve as a basis for policy planners to predict and therefore evaluate the end of 100 level performances of students admitted into tertiary institutions based on their previous academic achievements.

• SCOPE OF THE STUDY

This project is covers the use of fuzzy logic in predicting student academic performance. That is, developing a model or framework of predictive system by identifying the most suitable factors to predict end of first year performances of prospective students which will then be used as a yardstick for student performance evaluation, since reports show that Fuzzy Logic technique is an important technique in handling imprecision and uncertainty.

1.5       LIMITATIONS OF THE STUDY

This work is intended to tackle and solve problems related to predicting end of first year academic performance using fuzzy logic based on previous or prior knowledge or academic achievements presented during admission or registration process. It is limited to considering four factors only in modelling a fuzzy inference system for predicting CGPA of students. These factors considered are student’s mode of entry, UTME score, secondary school result and students age at entry.

CHAPTER FIVE

SUMMARY, CONCLUSION AND RECOMMENDATION

5.1 SUMMARY OF THE STUDY

The main aim of this project is to predict the end of first year performances of students using fuzzy logic based on previous or prior academic achievements presented during admission process. The previous academic achievements were used as factors to predict the CGPA of students. The technique involved modelling a predictive framework known as a fuzzy inference system (FIS). The FIS was modelled using the MATLAB fuzzy logic toolbox. The model was then used to predict the CGPA of students newly admitted into Computer Science department, Federal University of Technology, Minna.

The predicted performances of each student were compared with their current GPA and 60% accuracy was achieved. The achieved result shows the capability of fuzzy logic in handling uncertainty, given that predicting or evaluating the academic performances of students based on their previous achievements involves handling uncertain or imprecise data.

5.2 CONCLUSION

In this project work, a rule based fuzzy inference system was successfully modelled for predicting the end of first year performances of students using fuzzy logic, based on their previous academic achievements which were used as predictive factors.

It can be concluded or deduced that the modelled fuzzy inference system in this project is an educational technique in predicting the academic performances of students based on the considered factors or input variables.

5.3 RECOMMENDATION  

Due to the competitive nature of securing admission into higher institutions of learning in Nigeria, I hereby recommend the use and implementation of the proposed modelled fuzzy inference system in conducting admissions into Nigeria’s higher institutions in order to identify students that will perform well if granted admission. This inference system was used to predict the end of first year performances of students admitted into Computer Science department, Federal University of Technology, Minna only. It is also recommended for use across other departments as well as other schools in and around the country.

Given that predictions in this study were based on only four factors which were considered as input variables, it is furthermore recommended that more factors be considered in future work in order to improve the accuracy of the model.