You are on page 1of 11

Proceedings oI International ConIerence on Computing Sciences

WILKES100 ICCS 2013


ISBN: 978-93-3107-172-3
Fuzzy rule based expert system to automate university examination
grading
1
Vishu,
1
Prateek Agrawal*,
1
Sanjay Kumar Singh,
2
Leena Jain
1
Department of Computer Science and Engineering, Lovely Professional University, Punjab //
2
Department of Computer Application, Global Institute of Management and Emerging Technologies, Amritsar, Punjab
vishumadaan123gmail.com, prateek061186gmail.com, sanjayksingh.012gmail.com, leenajain79gmail.com
Abstract
This paper work is in regard to represent the application to automate examination grading based on Iuzzy logic approach. The
Proposed work explains a Iuzzy based expert system which would be designed to check students' perIormance and to award
examination grades oI students. Students' perIormance is calculated considering toughness / easiness oI paper level and course
level & the overall perIormance level oI section whereas criteria on the basis oI which grades would be awarded to students, is
determined through course level and paper level. This paper explains the clear objective oI proposed system, its implementation.
DiIIerent deIuzziIication techniques are applied to veriIy the accuracy and robustness oI the system. The proposed system is
implemented Ior evaluating the academic perIormances oI students oI M.Tech, Department oI Computer Science and Engineering
(CSE) oI Lovely ProIessional University, India. The method can also easily be implemented to any academic organization Ior the
perIormance evaluation.
2013 Elsevier Science. All rights reserved.
Keywords: Fuzzy expert system, Fuzzy logic, Academic perIormance evaluation, examination grades
1. Introduction
For years, researchers are Iocusing only on the ways to resolve tasks involving human eIIorts by replacing with
expert systems. Because it is a time consuming way to calculate examination grades oI students manually and it
also leads to inaccurate results most oI time. ThereIore there is a need to develop automated tools that provides
accurate results as here required in the Iorm oI grades oI individual student. Though some study on (Iuzzy system
Ior evaluating students' learning achievement) (Ibrahim Saleh*, 2009) and another study on (Evaluation oI student
perIormance in laboratory applications using Iuzzy logic) (Gokhan Gokmen a*, 2010) have been made, but those
are not strong enough to be eIIectively adopted by all kind oI organizations as there are no organized and
scientiIic academic perIormance evaluation methodology.
One time evaluation oI answer sheets can't judge student's overall learning achievement. There are number oI
parameters that eIIect student's perIormance. It may be the possibility that student appearing in exams oI previous
year is good but unable to attempt paper in a better way than the student appearing in exams oI current year, may
be due to diIIerence in diIIiculty level oI question paper, diIIerence in syllabus oI same course etc. ThereIore, we
can't compare perIormance oI these two students. This shows that not only student level but course level and
paper level also aIIects examination results. It leads to a need oI development oI such system that would be able to
do so automatically.
2. Design of Fuzzy Inference Process for Academic Performance Evaluation System
612 Elsevier Publications, 2013
The overall design oI proposed system is depicted through Iollowing Iigure
1. Corresponding author - Vishnu
Figure 1. Design oI Fuzzy InIerence Process Ior Academic PerIormance Evaluation
2.1 Identification of factors affecting student's performance
This is the initial step towards development i.e., to calculate Iactors inIluencing student's perIormance. There are
number oI Iactors that aIIect student academic perIormance, here we have considered 3 Iactors (Course Level, Paper
Level, Section Level) to check student's (students oI single section) perIormance and 2 Iactors (Course Level and
Paper Level) that are enough to determine the criteria on the basis oI which grades will be awarded to students. The
third level i.e., Section Level Iurther determines Faculty Level. It would be easy to judge about the Iaculty
involvement in improvement oI student's perIormance through section level. Here we have considered three section
oI previous year and Iour sections belong to current year, Ior which perIormance is calculated.
S
pi:
p stands Ior previous year, where i1 to 3.
S
ci:
c stands Ior current year, where i1 to 4.
The Iollowing Iactors determine the university perIormance oI a course based on the Course level, paper level
and section level considering data oI previous year as well as oI current year.
Course Level (CL)
It indicates the diIIiculty level oI subject. Course level is determined using average marks (Continuous Assessment
(CA) Mid Term Examination (MTE) End Term Examination (ETE)) oI students oI each course as shown in
table 1 (previous year (S
p1,
S
p2,
S
p3
), current year (S
c1,
S
c2,
S
c3,
S
c4
)).
The Iormula Ior evaluating course level is constructed as Iollows:


(1)
Where,
Ag
c
Average oI marks (Continuous Assessment (CA) Mid Term Examination (MTE) End Term
Examination (ETE)) oI all courses Ior particular programme (include current year batch).
Sg
c
Standard Deviation oI all the courses (Continuous Assessment (CA) Mid Term Examination (MTE)
End Term Examination (ETE)) Ior particular programme (include previous and current batches).
M
c
Average marks (Continuous Assessment (CA) Mid Term Examination (MTE) End Term
Examination (ETE)) oI particular batch student Ior particular course.
A
c
Average marks (Continuous Assessment (CA) Mid Term Examination (MTE) End Term Examination
(ETE)) oI all the student oI particular batch Ior all the courses.
IdentiIication oI Iactors aIIecting perIormance, Iuzzy linguistic
variables and their membership Iunctions
FuzziIication
Fuzzy Rule Construction
Fuzzy InIerence Generation
DeIuzziIication
Evaluating and Analyzing perIormance
Figure 1. Design oI Fuzzy InIerence Process Ior Academic PerIormance Evaluation
Award grades to students
613 Elsevier Publications, 2013
Vishu, Prateek Agrawal, Sanjay Kumar Singh, Leena jain
Fuzzy ule ased xpert ystem to utomate niversity xamination rading
S
c
Standard deviation oI all courses (include current batch).
Paper Level (PL)
It indicates the diIIiculty level oI question paper, so that we can answer a question that whether question papers oI
current year batch courses are diIIicult or easy Irom the question papers oI last year batch. Paper level is determined
using the average marks End Term Examination oI students oI each course as shown in table 2. Section S
p1,
S
p2,

and S
p3
belong to previous year and section S
c1,
S
c2,
S
c3,
and S
c4
are oI current year Ior which perIormance value
and grades are going to awarded.
The Iormula Ior evaluating course level is constructed as Iollows:



(2)
Where,
Ag
p
Average oI marks (End Term Examination (ETE)) oI all courses Ior particular programme (include
current batch). Sg
p
Standard Deviation oI all the courses (End Term Examination (ETE)) Ior particular
programme (include current batch). M
p
Average marks (End Term Examination) oI particular batch student
Ior particular course.
A
p
Average marks (End Term Examination) oI all the student oI particular batch Ior all the courses.
S
p
Standard deviation oI all courses (include current batch).
Section Level (SL)
It indicates level oI the considered section i.e., to know how much capable oI any section X is. Section level Iurther
explains the perIormance oI the respective Iaculty oI that section. Section level is determined using the average
marks (Continuous Assessment (CA) Mid Term Examination (MTE) End Term Examination (ETE)) all courses
oI each section.
The Iormula Ior evaluating course level is constructed as Iollows:



(3)
Where,
Ag
s
Average oI marks (Continuous Assessment (CA) Mid Term Examination (MTE) End Term
Examination (ETE)) oI all sections (include current and previous year sections) Ior all courses.
Sg
s
Standard Deviation oI all the marks (Continuous Assessment (CA) Mid Term Examination (MTE)
End Term Examination (ETE)) oI all sections (include current and previous year sections).
M
s
Average oI average marks (Continuous Assessment (CA) Mid Term Examination (MTE) End Term
Examination (ETE)) oI all courses Ior particular section.
A
s
Average oI average marks (Continuous Assessment (CA) Mid Term Examination (MTE) End Term
Examination (ETE)) oI all courses oI current year batch sections.
S
s
Standard deviation oI all sections.
Adjusting Factor (AF)
AF is indicating the adjusting Iactor that would be added in the actual marks scored by individual student in a
particular exam oI a particular course Ior the considered section to remove the eIIects oI diIIiculty/easiness oI that
course, paper and also the ability oI the students. AF is calculated by considering course level and paper level. The
value oI AF ranges Irom -10 to 10. Higher the value reIers to toughness oI course, paper and the poor level oI
student and vice versa. When student's marks are adjusted using above deIined adjusting Iactor (AF) then it is
checked that in which category oI grades student's marks Iall, the respective grade would be awarded to that student.
Figure 2 shows diagrammatic view oI overall system. It shows section level will act as an input Ior Iuzzy rule
base1 and other two critical Iactors i.e. course level and paper level will be input oI Iuzzy rule base 2 to Iind grades.
614 Elsevier Publications, 2013
Crisp World Fuzzy World
Input
Figure 2. Block diagram oI proposed system
Table 1. Sample record of marks (ETE+MTE+CA)
Table 2. Sample record of marks (ETE)
Section
S
p1
S
p2
S
p3
S
c1
S
c2
S
c3
S
c4

C
1
37.53 32.28 36.1 62 64.69 61.85 63.63
C
2
57.42 53.34 57.12 59.8 64.43 60.51 63.96
C
3
56.3 49.26 55.2 58.83 59.09 55.89 57.44
C
4
48.21 47.06 53.56 63.03 67.18 65.2 63.48
C
5
56.81 56.44 58.05 46.45 50.39 46.77 49.71
C
6
69.45 67.65 70.72 53.07 59.68 53.62 57.17
C
7
53.43 52.38 54.88 65.83 66.9 65.58 64.96
C
8
68.29 66.18 70.17 68.21 68.21 68.21 68.21
Section
S
p1
S
p2
S
p3
S
c1
S
c2
S
c3
S
c4

C
1
51.275 47.15 49.285 57.9225 62.0175 59.1775 55.0525
C
2
64.025 57.89 60.6875 61.31 66.02 61.42 63.4825
C
3
60.4825 55.55 59.625 57.105 57.835 54.2325 55.735
C
4
53.8975 50.73 51.365 61.745 65.255 64.525 62.0075
C
5
58.705 53.6225 61.185 50.175 55.64 52.28 53.76
C
6
69.6875 66.9 68.8825 58.61 64.02 59.82 60.385
C
7
54.5525 54.2425 55.875 63.585 65.04 64.07 64.1225
C
8
70.3875 67.445 72.4975 54.9683 58.4431 54.7993 55.87625
C
9
56.9675 54.92 60.6875 39.1375 41.8075 38.0675 39.4825
C
10
55.435 54.54 59.5575 39.1375 54.318 47.443 54.775
C
11
53.805 49.41 49.3375 51.529 52.478 46.9875 49.96
C
12
66.7625 64.3375 70.4575 54.9683 58.4431 54.7993 55.87625
Course
Course
Grades
Paper level
PerIormance value Section level
AF(X)
Fuzzy Rule Base 2
Course level
Fuzzy Rule Base 1
615 Elsevier Publications, 2013
Vishu, Prateek Agrawal, Sanjay Kumar Singh, Leena jain
C
9
48 48.94 56.7 28.1 32.31 28.69 28.5
C
10
56 55.65 60.5 45.64 48.82 41.98 48.86
C
11
46.36 40.68 48.85 48.07 49.75 43.31 45.14
C
12
70.34 70.76 68.15 69.75 69.75 69.75 69.75
2.2 Fuzzification of Examination Results
When values oI three Iactors deIined above is calculated then there is a need to IuzziIy those input values.
FuzziIication oI Iactor's values was carried out by converting inputs into membership values oI the Iuzzy sets. Each
Iactor's value is input variables oI the Iuzzy logic based expert system. Course level, paper level and section level
would act as an input Ior Iuzzy system. Each input variable has three trapezoidal membership Iunctions. The Iuzzy
sets oI the input variables and their range are given in Table 3.
Table 3. Fuzzy linguistic variables and their fuzzy elements
2.3 Determination of Fuzzy Rules and Inference Method
The very next step aIter IuzziIication is to generate Iuzzy rules. Fuzzy rules are the knowledge base oI Iuzzy expert
system on the basis oI which expert system takes decision accordingly. The proposed system's creation depends on
appropriate choice oI Iuzzy sets and rules. It is a common situation that experts can only properly solve the problem
with Iorming oI Iuzzy sets and rules and hence there is a need to design proper set oI rules, rules may be designed
by some kind oI automatic means. Figure 3 shows appropriate rules determined to get accurate result.
Fuzzy linguistic input variables with
membership range(0-100)
Fuzzy Linguistic Output variable with
membership range(0-8)
T
o
u
g
h

M
o
d
e
r
a
t
e

E
a
s
y

P
e
r
f
o
r
m
a
n
c
e

L
o
w

A
b
o
v
e
A
v
g

A
v
g

B
e
l
o
w
A
v
g

H
i
g
h

Course Level -10-40 30-70 60-110
0-3 1-5 3-6 4-8 6-9
Paper Level -10-40 30- 70 60-110
Section Level P
o
o
r

A
v
e
r
a
g
e

H
i
g
h

-10-40 30- 70 60-110
616 Elsevier Publications, 2013
Fuzzy ule ased xpert ystem to utomate niversity xamination rading
Figure 3. Fuzzy Rule Base view Ior evaluating academic perIormance
In this proposed system 27 rules are generated aIter consulting domain experts. These IF-THEN Iuzzy rules act as a
knowledge base Ior expert system. The Iuzzy rules are constructed by consulting experts oI Examination Department,
LPU.
Some sample IF-THEN rules Ior academic perIormance calculation Irom the rule base are given as below:
i)
IF (Courselevel is Tough) AND (Paperlevel is Tough) AND (Sectionlevel is Poor) THEN PerIormance is Low.
ii)
IF (Courselevel is Tough) AND (Paperlevel is Tough) AND (Sectionlevel is Average) THEN PerIormance is BelowAvg.
iii)
IF (Courselevel is Moderate) AND (Paperlevel is Tough) AND (Sectionlevel is Poor) THEN PerIormance is Low. Etc.
Similarly continuing this way all rules were generated with consideration oI all possible combinations oI Iactors.
respective SurIace view oI the rule base Ior evaluating academic perIormance oI students is generated and shown in Iigure 4.
Figure 4. SurIace view oI the rule base Ior evaluating academic perIormance oI students
617 Elsevier Publications, 2013
Vishu, Prateek Agrawal, Sanjay Kumar Singh, Leena jain
2.4 Fuzzy Inference Generation
To implement this proposed system standard Max-min (Mamdani) inIerence algorithm has been used in this
Iuzzy inIerence process, as it is considered to be most popularly used Iuzzy inIerence strategy (Ngai, 2003).
Mamdani inIerence method is intuitive and widely accepted. It is also suited to human input. (The mathworks,
2009). There are 27 rules deIined Ior this system. There may be a possibility that several rules are active Ior the
same output membership Iunction. It is mandatory that only one membership value is chosen. In this paper, we
used the method proposed by Mamdami, which is represented by Iollowing equation:






With the help oI above equation it is determined that what would be an output membership Iunction value Ior
e

ach active rule. When one rule is active, an AND operation is applied between inputs. The smaller input value is
chosen and its membership value is determined as membership value oI the output Ior that rule. This method is
repeated, so that output membership Iunctions are determined Ior each rule. At the end, 'min' operation is applied
between inputs and max operations are between outputs.
2.5 Defuzzification
When the inIerence process is complete, the resulting data Ior each output oI the Iuzzy classiIication system are
a collection oI Iuzzy sets or a single, aggregate Iuzzy set. Proposed system is deIuzziIied with help oI 'Centroid'
method as it produces crisp` output data by computing the horizontal-axis component oI the geometric centroid oI
the Iuzzy set.
2.6 Evaluating, Analyzing and Ranking performance
To assure robustness oI our proposed system it is necessary to analyze perIormance oI students using diIIerent
deIuzziIication methods. DiIIerent input values are entered into system and their corresponding rank is mentioned
in table 4.
Table 4. Ranking oI diIIerent perIormance obtained Irom diIIerent inputs aIter deIuzziIications
C
o
u
r
s
e

l
e
v
e
l

(
C
L
)

P
a
p
e
r

L
e
v
e
l

(
P
L
)

S
e
c
t
i
o
n

L
e
v
e
l

(
S
L
)

Defuzzified Performance and Rankings
L
O
M

R
a
n
k
s

M
O
M

R
a
n
k
s

B
I
S
E
C
T
O
R

R
a
n
k
s

C
E
N
T
R
O
I
D

R
a
n
k
s

S
O
M

R
a
n
k
s

20 30 32 1.76 2 1.52 1 1.76 2 2 2 1.28 2
25 25 50 3.04 3 3 3 2.96 4 3 3 2.96 6
45 29 20 1.52 1 1.52 1 1.52 1 1.5 1 1.52 3
26 22 89 4.48 6 4.48 5 4.48 6 4.5 7 4.48 4
50 19 56 3.04 3 3 3 2.96 4 3 3 2.96 6
27 56 22 3.04 3 3 3 2.96 4 3 3 2.96 6
76 22 15 1.52 1 1.52 1 1.52 1 1.5 1 1.52 3
56 15 87 4.48 6 4.48 5 4.48 6 4.5 7 4.48 8
24 54 46 4.48 6 4.48 5 4.48 6 4.5 7 4.48 8
88 23 57 4.48 6 4.48 5 4.48 6 4.5 7 4.48 8
38 34 55 3.76 4 3 3 3.36 5 3.44 5 2.24 5
62 31 35 4 5 2.4 2 2.88 3 3.04 4 0.8 1
15 28 65 5.2 8 3.6 4 3.6 6 3.56 6 2 4
22 76 33 4.88 7 4.48 5 4.48 7 4.5 7 4.08 7
In table 5, ranks are provided with respective to each input and it is Iound that all inputs (CL, PL, and SL) got almost
same rank in all deIuzziIication methods. It shows accuracy oI this model. It assures us to give eIIicient and reliable
results when would be applied in any educational organization.
(4)
618 Elsevier Publications, 2013
Fuzzy ule ased xpert ystem to utomate niversity xamination rading
Table 5. Performance value obtained using fuzzy logic
Course Name Course Level Paper Level Section Name Section Level Performance
C
1

49.03959 48.31345 S
c1
75.89105 8
49.03959 48.31345 S
c2
29.17163 3.04
49.03959 48.31345 S
c3
63.66274 5.04
49.03959 48.31345 S
c4
34.1833 3.76
C
2

52.95846 50.90173 S
c1
75.89105 8
52.95846 50.90173 S
c2
29.17163 3.04
52.95846 50.90173 S
c3
63.66274 5.04
52.95846 50.90173 S
c4
34.1833 3.76
C
3

49.11318 48.04503 S
c1
75.89105 8
49.11318 48.04503 S
c2
29.17163 3.04
49.11318 48.04503 S
c3
63.66274 5.04
49.11318 48.04503 S
c4
34.1833 3.76
C
4

51.73864 51.87953 S
c1
75.89105 8
51.73864 51.87953 S
c2
29.17163 3.04
51.73864 51.87953 S
c3
63.66274 5.04
51.73864 51.87953 S
c4
34.1833 3.76
C
5

47.34004 42.8714 S
c1
75.89105 8
47.34004 42.8714 S
c2
29.17163 3.04
47.34004 42.8714 S
c3
63.66274 5.04
47.34004 42.8714 S
c4
34.1833 3.76
C
6

53.3115 49.19686 S
c1
75.89105 8
53.3115 49.19686 S
c2
29.17163 3.04
53.3115 49.19686 S
c3
63.66274 5.04
53.3115 49.19686 S
c4
34.1833 3.76
C
7

52.79684 52.72017 S
c1
75.89105 8
52.79684 52.72017 S
c2
29.17163 3.04
52.79684 52.72017 S
c3
63.66274 5.04
52.79684 52.72017 S
c4
34.1833 3.76
C
8

51.33217 56.38654 S
c1
75.89105 8
51.33217 56.38654 S
c2
29.17163 3.04
51.33217 56.38654 S
c3
63.66274 5.04
51.33217 56.38654 S
c4
34.1833 3.76
C
9

39.9662 31.50506 S
c1
75.89105 4.72
39.9662 31.50506 S
c2
29.17163 1.68
39.9662 31.50506 S
c3
63.66274 3.68
39.9662 31.50506 S
c4
34.1833 2.08
C
10

43.83309 42.04994 S
c1
75.89105 8
43.83309 42.04994 S
c2
29.17163 3.04
43.83309 42.04994 S
c3
63.66274 5.04
43.83309 42.04994 S
c4
34.1833 3.76
C
11

43.80999 40.47484 S
c1
75.89105 8
43.80999 40.47484 S
c2
29.17163 3.04
43.80999 40.47484 S
c3
63.66274 5.04
43.80999 40.47484 S
c4
34.1833 3.76
C
12

50.95211 56.99712 S
c1
75.89105 8
50.95211 56.99712 S
c2
29.17163 3.04
50.95211 56.99712 S
c3
63.66274 5.04
50.95211 56.99712 S
c4
34.1833 3.76
619 Elsevier Publications, 2013
Vishu, Prateek Agrawal, Sanjay Kumar Singh, Leena jain
Experimental results and Discussion
For experimentation we have used data collected Irom Lovely ProIessional University Phagwara (Punjab) India.
Above tables shows average marks oI all courses oI each section oI M.TECH Irom the Department oI Computer
Science and Engineering, Lovely ProIessional University. Formula deIined in equations 1, 2, 3 is applied to
calculate all Iactors. For previous years we have used data oI 2011 batch and grades are calculated oI 2012 batch
year students. On applying Fuzzy Logic approach on that data, we obtain a PerIormance value Ior each section
having same Course and Paper Level. Obtained PerIormance values oI all courses Ior all sections i.e., 48
combinations are listed in table 6.
2.7 Awarding Grades
DiIIerent universities have their own category oI grades and their range. A sample oI categories oI grade and
their range is as shown in table 5.
Table 5. Range oI grades
Grade Range
A Above 85
A 75-85
B 65-74
C 55-64
D 45-54
D- 35-44
E Below 35
II suppose student scored 54 in a course aIter adding AF marks become 62, would be awarded as grade 'C'. In this
case it may be the possibility that course level, paper level is Tough and student is poor in this course.
3. Conclusion
According to this proposed new method Ior students academic perIormance evaluation system obtained results
are very reliable and unbiased result comes out which proves a Iare way oI grading Ior all students. On the other
hand iI we talk about Traditional methods oI grading which are adhered to a constant mathematical rule, evaluation
with Iuzzy logic provides great Ilexibility and reliability.
In conclusion, this application is not only to check perIormance oI theoretical lessons oI engineering students
but can also be used Ior other educational domain like e-learning and distance education provided by Open
University system in India. To do proposed work successIully Iuzzy logic approach has been used to evaluate
student's perIormance, this system may be an application based on combination oI Neuro-Fuzzy technique in Iuture.
It may also be more justiIiable and accurate using one more Iactor that eIIects student's perIormance i.e., Faculty
Level.
References
|1| Biswas, R. An application oI Iuzzy sets in student's evaluation. Elsvier Science , pp. 18, (1995).
|2| C.C.Yee, Y. PerIormance Appraisal System using MultiIactorial Evaluation Model. International Journal of Human and Social Sciences , pp.
15, (2010).
|3| Chiung Moon, J. L. An implementation case Ior perIormance appraisal and promotion ranking. IEEE , pp. 17, (2007).
|4| Gokhan Gokmen a*, T. C. evaluation oI student perIormance in laboratory applications using Iuzzy logic. Elsevier Science Ltd , pp. 902909,
(2010).
|5| Gokhan Gokmen, T. C. Evaluation students learning achievment using Iuzzy logic. Elsevier Science Ltd , pp. 902909, (2010).
|6| Ibrahim Saleh, S.-i. K. A Iuzzy system Ior evaluating student's learning achievement. Esevier Science Ltd , pp. 62366243, (2009).
|7| K, D. The essence oI expert system. England: Prentice-Hall, (2000).
|8| L.A., Z. Fuzzy sets. InIormation and Control, pp. 338353, (1965).
|9| Nolan, J. R. An expert Iuzzy classiIication system Ior supporting the grading oI student , Expert Systems With Applications, pp. 5968, (1998).
|10| P, J. Introduction to expert systems. England: Addison Wesley, (1990).
620 Elsevier Publications, 2013
Fuzzy ule ased xpert ystem to utomate niversity xamination rading
|11| Ramjeet Singh Yadav, V. P. Modeling Academic PerIormance Evaluation Using SoIt Computing Techniques: A Fuzzy Logic Approach.
IJCSE , 676686, (2011).
|12| Shih-Ming Bai, S.-M. C. Evalating students learning achievemnt using Iuzzy membership Iunctions and Iuzzy rules. Elsevier Science , 399
410, (2008).
|13| Shilpa Ingoley. Use oI Fuzzy Logic in Evaluating Students` Learning Achievement.
pp. 4754, (2012).
|14| Sivanandam, S. a. Introduction to fuzzy logic using Matlab, (2006).
|15| http://www.mathworks.com/access/helpdesk/help/pdf_doc/fuzzy/fuzzy.pdf.
|16| Zadeh, L. A. Making computers think like people. IEEE. Spectrum, pp. 2632, (1984).
|17| Zimmermann, H. J. Fuzzy Set Theory and its Applications. Boston, MA: Kluwer. (1991).
621 Elsevier Publications, 2013
Vishu, Prateek Agrawal, Sanjay Kumar Singh, Leena jain
Index

A
Academic performance evaluation system, 612617

D
Defuzzification, 618

F
Fuzzy expert system
academic performance evaluation system, 612617
Fuzzy inference generation, 618
Fuzzy inference process, 612
adjusting factor (AF), 614
analyzing and ranking performance, 618
awarding grades, 620
course level (CL), 613614
defuzzification, 618
evaluating, 618
of examination results, 616
paper level (PL), 614
rules and inference method, determination process, 616617
section level (SL), 614