COURSE GRADE CALCULATION USING FUZZY LOGIC (CGCFL) NURRIDAYAH BINTI MUHAMAD ZAHARI

COURSE GRADE CALCULATION USING FUZZY LOGIC (CGCFL) NURRIDAYAH BINTI MUHAMAD ZAHARI A report submitted in partial fulfilment of the requirements for the award of the degree of Bachelor of Computer Science (Software Engineering) Faculty of Computer Systems & Software Engineering Universiti Malaysia Pahang MAY, 2011 FRPUSTAKAAN UNiVERSTL MALAYSIA PAHANG No. Peroe a Tarikh 20 JrYv 2b2 No. Pa Han '1 N Rc-

vi ABSTRACT Grade Calculation using Fuzzy Logic System is a system that developed to calculate the final grade of Data Structure and Algorithms student's more accurate and valuable result. Nowadays, the calculation of student grade for data structure subject based on three categories which are cognitive, psychomotor and affective. However, in the current scenario, the final grade not considered at all on the categories. Example of the current situation is if the students get low mark in psychomotor, however the others categories student get excellent, the final result still show that the student pass the course. To produce high quality of graduate, these three categories must be considered. Because of this situation, this system had been developed by implement new concept using fuzzy logic and use Mamdani inferences to calculate the final grade student accurately. After done the development and implementation process, the result show that the system can produces almost 90 % accurate value. This is because the new techniques of defuzzification had been implemented which are center of gravity (COG) and mean of maximum (MOM). However, the problem is there are some value that no really accurate. This happens because of more techniques need to be used. This system is better than manual system because it calculates all the value automatically and the three categories are being considered.

vii ABSTRAK 'Fuzzy Logic' merupakan salah satu teknik di dalam 'Artificial Intelligence' di mana ianya semakin meluas penggunaanya di dalam pembangunan sesuatu sistem hari mi. 'Course Grade Calculation using Fuzzy Logic System' adalah satu system yang dibangunkan untuk mengira gred markah bagi pelajar yang mengambil subjek 'Data Structure and Algorithm' untuk menghasilkan keputusan yang tepat dan logik Pada masa mi, pengiraan nilai pelajar untuk mata pelajaran 'Data STucture' berdasarkan tiga kategori iaitu kognitif, amali dan afektif. Namun, dalam senario mi, nilai akhir tidak dikira semua kategori. Contoh situasi yang dialami oleh UMP ketika mi adalah jika seseorang pelajar menerima markah yang rendah dalam amali, namun lain kategori pelajar itu mendapat markah yang tinggi, namun keputusan akhir masih mernmjukkan bahawa pelajar tersebut telah lulus didalam kursus mi. Pelajar yang cemerlang seharusnye mempunyai kelulusan yang tinggi di mana ketiga-tiga kategori harus dipertimbangkan. Setelah proses pembangunan dan pelaksanaan dilakukan, Keputusan kajian menunjukkan bahawa sistem mi mampu menghasilkan keputusan hampir 90% nilai yang tepat. Hal mi kerana teknik barn proses 'defuzzificaton' telah digunakan iaitu merupakan 'centre of gravity (COG)' dan 'mean of maximum (MOM)'. Namun begitu, masih lagi mempunyai masalah iaitu sesetengah nilai yang dihasikan adalah tidak benar-benar tepat. Hal mi terjadi kerana teknik yang lebih efektif perlu digunakan. Walaubagaimanapun, sistem mi lebih balk danipada sistem manual kerana mampu mengira nilai secara automatik dan tiga kategori di ambil kira.

vii' TABLE OF CONTENT CHAPTER TITLE PAGE TITLE PAGE SUPERVISOR'S DECLARATION STUDENT'S DECLARATION DEDICATION ACKNOWLEDGEMENT ABSTRACT ABSTRAK TABLE OF CONTENT LIST OF TABLES LIST OF FIGURES LIST OF SYMBOLS LIST OF APPENDICES i iv v vi vii ix xii xiv xvii Ax INTRODUCTION 1 1.1 Introduction 1 1.2 Problem Statement 2 1.3 Objectives 2 1.4 Scope 3 2 LITERATURE REVIEW 4 2.1 Overview of Outcomes-Based Education 4 2. 1.1 Theory of Outcomes-Based Education 5

ix 2.1.2 Level of Outcomes-Based Education 7 2.1.2.1 Program Educational Objectives (PEO) 9 for Bachelor of Computer Science (Software Engineering) 2.1.2.2 Program Outcomes (P0) for Bachelor of 9 Computer Science (Software Engineering) 2.1.2.2 Course Outcomes (P0) for 10 BCS1093:Data Structure & Algorithm (Software Engineering) 2.2 Fuzzy Logic 11 2.2.1 History of Fuzzy Logic 11 2.2.2 Fuzzy Theory 13 2.2.3 Fuzzy Inferences 14 2.2.3.1 Structure of Fuzzy Inference Engine 15 2.2.3.2 Fuzzy IF-THEN rules 16 2.2.3.3 Membership Functions 17 2.2.3.4 Classification of fuzzy inference methods 17 2.2.4 Mamdani Inferences 18-39 2.2.4.1 Project Staffing and Project Funding 19-26 System 2.2.4.2 Evaluation of Student Performance in 27-33 laboratory 2.2.4.1 Fuzzy Logic Traffic Light 34-39 3 METHODOLOGY 40 3.1 Introduction 40 3.2 Prototyping Model 41 3.2.1 Phase of Prototyping Model 42 3.2.2 Identify the problem 43 3.2.3 Develop the 2D prototype 44 3.2.3.1 Calculation of the final mark (100%) 45

X 3.2.3.2 Fuzzification process 46 3.2.3.3 Rule evaluation process 54 3.2.3.4 Aggregation 55 3.2.3.5 Defuzzification process 60 3.2.4 Implement and use prototype 63 3.2.4.1 Mean of Maximum (MOM) 64 3.2.4.2 Testing of 2D input 64 3.2.5 Revise and enhance prototype 66 3.2.5.1 Calculation of the final mark (100%) 66 3.2.5.2 Fuzzy Logic process 69 3.2.5.2.1 Fuzzification of the inputs 70 3.2.5.2.2 Rules evaluation 82 3.2.5.2.3 Aggregation process 85 3.2.5.2.4 Defuzzification process 86 3.2.6 Implement and use prototype 87 3.2.7 Convert to operational system 87 3.3 Development and deployment 87 3.3.1 Hardware and software requirements 88 4 IMPLEMENTATION 89 4.1 Introduction 89 4.2 System implementation environment 89 4.3 System implementation process 90 4.3.1 Main page of CGCFL 91 4.3.2 Input page of CGCFL 92 4.3.3 Fuzzy Logic process 94 4.3.3.1 Fuzzification process 95 4.3.3.2 Rules evaluation 100 4.3.3.3 Aggregation 103 4.3.3.4 Defuzzification 105 4.4 Testing Result 107

xi S RESULT AND DISCUSSION 108 5.1 Introduction 108 5.2 Analysis of result 109 5.2.1 Objectives achievement 109 5.3 Project Constraint 110 5.3.1 Development constraint 111 5.3.2 System constraint 111 5.4 Advantages and disadvantages of CGCFL 112 5.4.1 Advantages of the system 112 5.4.2 Disadvantages of the system 113 5.5 Suggestion and improvement 113 5.6 Assumption 114 6 CONCLUSION 115 REFERENCES 117 APPENDICES 120

xli LIST OF TABLES TABLE TITLE PAGE NO. 2.1 Fuzzy set of variables and its range 19 2.2 Rules of this application 20 2.3 Fuzzy set for each variables and its range 28 2.4 Fuzzy set of output variable. 29 2.5 25 rules for the variables 30 2.6 Exam scores and calculated performance 32 2.7 Fuzzy variables for input 35 3.1 Example of activities for cognitive and psychomotor domain 45 3.2 Formulae to convert the mark into 100% 45 3.3 Fuzzy set of cognitive domain 47 3.4 Fuzzy set of psychomotor domain 47 3.5 Table of equations for cognitive 50 3.6 Table of equations for psychomotor 51 3.7 Each fuzzy set have five rules 55 3.8 25 Rules in 2D Prototype 55 3.9 The grade for each mark 62 3.10 Percentages of sub categories of cognitive, psychomotor and 66 affective. 3.11 Formulae to convert the mark into 100% 67 3.12 Linguistic variable of cognitive input, x 70 3.13 Linguistic variable of psychomotor input, y 70

xlii 3.14 Linguistic variable of affective input, zi 71 3.15 List of the equations for fuzzy set of cognitive 75 3.16 List of the equations for fuzzy set of psychomotor 76 3.17 List of the equations for fuzzy set of affective 77 3.18 Expected output for fuzzification process 82 3.19 Hardware requirements in development of the system 88 3.20 Software requirements in development of the system 88

xlv LIST OF FIGURES FIGURE TITLE PAGE NO. 2.1 Different level of OBE 7 2.2 The percentage domains for Data structure & algorithm 8 course 2.3 Structure of fuzzy inferences 15 2.4 IF-THEN rules 16 2.5 Types of fuzzy inferences method 17 2.6 Step of fuzzification of crisp input 20 2.7 Formulae of OR operator 21 2.8 Formulae of OR operator 21 2.9 Rule evaluations 22 2.10 Aggregations of rule consequents 23 2.11 Formulae of COG 23 2.12 Calculate the COG by using the formulae 24 2.13 Graph of defuzzification 25 2.14 Interface of the application with the final input 26 2.15 Membership functions of Exam and Exam2 28 2.16 Membership functions of Exam and Exam2 29 2.17 Defuzzification functions for performance value 31 2.18 Membership functions for Queue 35 2.19 Membership functions for Arrival 36 2.20 Membership functions for Extension 36 2.21 Example rules for the application 37 2.22 Internal view of Mamdani inference 38

xv 2.23 Design of traffic light interface 39 3.1 Steps in Prototyping Model 42 3.2 Example of the interface and expected output of mark 46 calculation 3.3 Graph of degree of membership for cognitive domain 48 3.4 Graph of degree of membership for psychomotor domain 49 3.5 Formulae to find the gradient 49 3.6 Crisp input xl 52 3.7 Crisp input yl 53 3.8 Expected output for defuzzification 54 3.9 Rule evaluation for rule number 17 57 3.10 Rule evaluation for rule number 21 58 3.11 Aggregation of rules sequence 59 3.12 Expected output of aggregation process 59 3.13 Expected output for final grade by using formulae COG 60 3.14 Expected output on the defuzzification graph 61 3.15 Expected output on the interface of the system 62 3.16 Interface of the output when using MOM (weak) 64 3.17 Interface of the output when using MOM (excellent) 65 3.18 Interface of the output when using COG (average) 65 3.19 Example of the interface and calculation of student mark 68 3.20 Process to get the student mark based on fuzzy logic concept 69 3.21 Graph of degree of membership for cognitive input. 72 3.22 Graph of degree of membership for psychomotor input. 73 3.23 Graph of degree of membership for affective input 74 3.24 Graph of degree membership of cognitive with the value of 78 0.6 and 0.4 3.25 Graph of degree membership of psychomotor with the value 79 of 0.5 3.26 Graph of degree membership of affective 80

xvi 3.27 Expected output for fuzzification process that show in the 81 interface of application 3.28 Rule evaluation for rule number 57 83 3.29 Rule evaluation for rule number 62 84 3.30 Rule evaluation for rule number 62 84 3.31 Rule evaluation for rule number 87 85 3.32 Aggregation of rules sequence 85 3.33 Output from defuzzification process 4.1 Flow Chart for CGCFL 86 4.2 Home page of CGCFL 90 4.3 Input page of CGCFL 91 4.4 Source code of mark calculation and formula to convert in 92 100% 93 4.5 Interface of the result form where all the value are displayed 4.6 Interface of fuzzification process 94 4.7 Source code of fuzzification for cognitive input 95 4.8 Source code of fuzzification for psychomotor input 97 4.9 Source code of fuzzification for affective input. 98 4.10 Interface of evaluation rules show that the rules had been 99 fired are rules 57, 62, 82 and 87 100 4.11 Source code of rule 57 4.12 Source code of rule 62 101 4.13 Source code of rule 82 101 4.14 Source code of rule 87 102 4.15 Source code of aggregation of weak 102 4.16 Source code of aggregation of average 103 4.17 Source code of aggregation of good 104 4.18 Source code of COG function 104 41.9 Source code of condition to use the MOM function 105 4.20 Source code of condition to use the MOM1 function when 106 very weak area is largest. 106

xvii 4.21 Source code of condition to use the MOM2 function when excellent area is largest. 107

xviii LIST OF SYMBOLS OBE - COG - PEO - P0 - CO - CGCFL - MOM - Outcomes-Based Education Centre of Gravity Program Education Objectives Program Outcomes Course Outcomes Course Grade Calculation using Fuzzy Logic Mean of Maximum

xix LIST OF APPENDICES APPENDIX TITLE PAGE A Gantt Chart 120 B List of 125 rules 122 C Test result 128 D User manual 134

CHAPTER 1 INTRODUCTION 1.1 Introduction Nowadays, education in Universiti Malaysia Pahang (UMP) has been facing a problem which include measuring level of how educated theft student. The level of student is measured based on three categories which are cognitive, psychomotor and affective. Each subject has their own learning outcomes that apply those categories for final grade. However, in the current scenario, the student level in final grade does not comply with all those categories. Through the development of the project, the main objective of this project is to implement Artificial Intelligence (Al) concept by using fuzzy logic techniques approach to calculate the final course grade. The intelligence of the system can be show that the system will considered all the categories to measure the student's level. Then, the objective of university to produce not just excellent also can communicate well with others is achieved.

2 1.2 Problem Statement Nowadays, the calculation of student grade for data structure subject based on three categories which are cognitive, psychomotor and affective. However, the final grade not considered all of the categories. The percentages of student grade are 50% cognitive, 45% psychomotor and 5% affective. Cognitive category is covered for final and quizzes. Psychomotor category is present as a technical skill such as hands on test (HOT). Then, for affective category consist of online participation. For example, for Data Structure & Algorithm course, let said that student get zero in effective, he or she still pass in the final course grade. This happens to all of the subjects in the program. As a conclusion, the course outcomes and program outcomes for all the evaluation are not achieved. The new concept here is to implement fuzzy logic so that all the learning outcomes must be achieved to determine the course grade of student by considered the entire categories which are cognitive, psychomotor and affective. 1.3 Objectives The objectives of the system are: (i) To develop a prototype to calculate the course grade by using fuzzy logic. (ii) To implement new concept of grade course calculation which integrate all the learning outcomes categories.

3 1.4 Scope The scopes of the system are: (i) The prototype will be test for BCS1093:Data Structure & Algorithm Course (ii) The prototype will be implement fuzzy logic concept. (iii)the program outcomes will be category into three domains which are cognitive, psychomotor and affective. (iv)input and output of fuzzy logic will be categories into five fuzzy set categories which are excellent, good, average, weak and very weak.

CHAPTER 2 LITERATURE REVIEW 2.1 Overview of Outcomes-Based Education (OBE) Outcomes-based education (OBE) is an educational process. Outcomes-based education (OBE) is one Of the methods of teaching which focuses on what student is can actually do after they are taught [1]. Directed and focussed at achieving certain specified outcomes in terms of individual student learning. Outcomes are key things student should understand and be able to do or qualities they should develop. All curriculum and teaching decisions are made based on how best to facilitate the desired outcome. The educational structures and curriculum in the Program Bachelor of Computer Science & Software Engineering are design to achieve those qualities and capabilities and they are regarded as means not end. Outcomes-Based Education (OBE) is important in education because the desired outcome is selected first and the curriculum, instructional materials and assessment are created to support the intended outcome. The outcomes-based education had been introduced because of the deficiencies of traditional education which are provides students with a learning environment with little attention to whether or not students ever learn the material Other than that, students are given grades and rankings compared to each other. The students become exam oriented or CGPA driven. Graduates are not

completely prepared for the workforce. Lack of emphasis on soft skills needed in jobs communication skills, interpersonal skills, analytical skills and others skills [8]. 5 2.1.1 Theory of Outcomes-Based Education (OBE) A recent definition of outcome-based education comes from James Towers. He wrote, "Education that is outcome-based is a learner-centered, results-oriented system founded on the belief that all individuals can learn"[l]. Towers (1996) listed four points to the OBE system that are necessary to make it work [2]:. What the student is to learn must be clearly identified. The student's progress is based on demonstrated achievement. Multiple instructional and assessment strategies need to be available to meet the needs of each student. Adequate time and assistance need to be provided so that each student can reach the maximum potential. William Spady (1994,) who is still widely regarded as OBE's leading advocate, explained outcomes-based education as focusing and organizing everything in the education system around what is essential for all students to be able to do successfully at the end of their learning experiences. This means starting with a clear picture of what is important for students to be able to do, then organizing the curriculum, instruction and assessment to make sure that learning ultimately happens [3]. William Spady and Kit Marshall wrote (1994), outcomes are clear, observable demonstrations of student learning that occur after a significant set of learning experiences. They are not values, attitudes, feelings, beliefs, activities, assignments,

goals, scores, grades, or averages, as many people believe. Typically, these demonstrations, or performances, reflect three things [1]: 6. What the student knows What the student can actually do with what he or she knows The student's confidence and motivation in carrying out the demonstration. Another theory of OBE system concept is in Wikipedia( the free encyclopedia), where it's explained that outcomes-based education (OBE) is a model of education that rejects the traditional focus on what the school provides to students, in favor of making students demonstrate that they "know and are able to do" whatever the required outcomes are. It reflects a belief that the best way for individuals and organizations to get where they're going is first to determine where they are and where they want to be then plan backwards to determine the best way to get from here to there. Outcomebased education (OBE) is a recurring education reform model. It is a student-centered learning philosophy that focuses on empirically measuring student performance, which is called outcomes [4]. Another set of outcome-based education theorists are Floyd Boschee and Mark Baron. They defined outcomes as future oriented, publicly defined, learner-centered, focused on life skills and contexts that characterized by high expectations of and for all learners, and sources from which all other educational decisions flow. Further they defined learning as facilitated carefully toward achievement of the outcomes, characterized by its appropriateness to each learner's development level, and active and experienced-based (Bosehee and Baron, 1994) [1].

7 2.1.2 Level of Outcomes-Based Education (OBE) There are three levels in OBE, which are program educational objectives (PEO), program outcomes (P0) and COUrSe outcomes (CO). The Figure 2.1 shows the level of OBE [5]. PROGRAM EDUCATIONAL OBJECTIVES (PEO) Few years after graduation Figure 2.1: Different level of OBE [5]. Program Educational Objectives (PEO) is statements that describe the expected accomplishments of graduates during the first several years following graduation. The concept that must be always remember to create the PEO: [6] Should they be practitioners in the profession of the discipline? Should they have entered the work force prepared for entry-level jobs? Should they be in a graduate or professional degree program? Should they have passed a licensure or certification exam in the field? Program Outcomes (P0) is focus on those abilities that are measurable at the SUCCeSSfIjI end of a student's academic program [5]. Other definitions of programme o utcomes are statements describing what students are expected to know and be able to perform or attain by the time of graduation, and shall [7]. These relate to the skills, Co gnitive, and behaviours that student acquire through the programme, and are linked to the P rogr amme educational outcomes.

8 Course outcomes (CO) describe what discipline faculty agrees are the most important abilities, cognitive, values, and attitudes that students should acquire as a result of successfully completing the course, regardless of where the course is taught, how it is delivered, or who teaches it. in developing course outcomes it often helps to think in terms of three domains which are [8]: Cognitive abilities - what does the student know? Psychomotor- what can the student do? Affective -what does the student care about? Figure 2.2 shows the domain of course outcomes (CO) and its type [9]. In Data Structure & Algorithm course also includes the three domains and each have their percentage to evaluate the performance student in different activities. We can conclude that the cognitive is a knowledge skill, psychomotor is a technical skill and affective is a soft skill. Course Outcomes (CO) (100%) Cognitive (50%) Psychomotor (45%) Affective (5%) - Cognitive domains - Involves cognitive and the devtlo5,ñeñt of intulleerual Skills Example of cognitive in Data Structure & Algorithms course: Final Exam (40%) Quizzes (10%) - Psychomotor domain - includes physical movement, coordination & use of the motor skill areas. Example of cognitive in Data Structure & Algorithms course: Hands on Test (45%) - Affective domain - includes manner we deal with things emotionally (eiuuplc feelin5s, interests, auirudes, appreciation, enthusiasms, motivations) - that might result from instruction).example of cognitive in Data Structure & Algorithms course: Online participation (5%) Figure 2.2: The percentages of domains for Data Structure & Algorithm course.

2.1.2.1 Program Educational Objectives (PEO) for Bachelor of Computer Science (Software Engineering) 9 Bachelor of Computer Science (Software Engineering) program has stated the program educational objectives that are consistent with the vision and mission of the Universiti Malaysia Pahang as follows [10]: PE01 801/o of our graduates should be employed in national, multinational or government organizations within 5 years in ICT related fields. PE02 60% of our graduates will be promoted to a senior level within 5 years. 2.1.2.2 Program Outcomes (P0) for Bachelor of Computer Science (Software Engineering) Performance skills and abilities are emphasized throughout the three year and half undergraduate program in order to prepare students to be successful engineers and to meet the university's program outcomes. Upon completion of the Bachelor of the Computer Science (Software Engineering) program at Universiti Malaysia Pahang, graduate will be able to [10]: P01 Able to demonstrate cognitive and understanding o the theory and principles of Computer Science specializing in Software Engineering P02 Able to apply appropriate techniques, skills and tools in Computer Science Practices specializing in Software Engineering P03 Able to identify problems and produce innovative solutions that comply with principles of Computer Science practices specializing in Software Engineering