Fuzzy Logic for Academic Orientation and Its Impact on Success:

Content Analysis

Harraki Youssef

, Aniss Moumen

and Driss Gretete

Laboratory of Engineering Sciences, ENSA Kenitra, IbnTofail University. Morocco

Keywords: Education, Orientation, Fuzzy Logic, Data Analysis and Forecasting.

Abstract: This paper will present an exploratory literature review and content analysis of fuzzy set theory for risk

management in education, including the impact of academic orientation on success. The study is developed

using textual analysis techniques, to situate our topic about a similar project on a corpus of references

(reference library Analyzing with the Zotero tool) from the most available databases: Springer Link, Scopus,

IEEE Xplore, Web of Science, Jstor. The interest of this work is to analyze our library by type of reference

by year of publication and by topic, which offers an overview and a critical evaluation of a set of articles

related to our research problem with a degree of similarity. Finally, compare the different models used and

discuss the results.

1 INTRODUCTION

Real-life is too complicated for one to get precise

explanations; an approximation (or fuzzy) must

therefore be introduced to obtain a realistic but

traceable model. As we move into the information

age, human knowledge is becoming more and more

critical. We need a theory to formulate human

expertise and integrate it into engineering systems

and other information such as mathematical models

and sensory measurements.

In other words, the critical question is how to

transform a human knowledge base into a

mathematical formula. For this purpose, a fuzzy logic

system is used to model cause-and-effect

relationships, assess the degree of risk exposure and

consistently classify key risks, taking into account

available data and expert opinions.

https://orcid.org/0000-0003-2437-9695

https://orcid.org/0000-0001-5330-0136

https://orcid.org/0000-0001-6155-3969

2 THE FOUNDATIONS OF

FUZZY SET THEORY AND ITS

APPLICATION IN EDUCATION

Fuzzy logic is a type of modelling that focuses on

predicting a "subjective" categorical variable: it

depends on the observer. This framework goes

beyond the classical statistics in which the variable's

value can be objectivist (

J. Klir, 2006)

The application of fuzzy logic (

F. Martin and T.

Ellen, 1994)

amounts to Attempting to apply to reason

close to human thought; it, therefore, allows expert

systems to be integrated into automated processes.

The theory of fuzzy sets (

J. Klir, 2006)

Developed

in 1965 by Professor Lotfi Zadeh of the University of

Berkeley in a founding paper, which defines the

principles (ZADEH, 1965); it constitutes a

generalization of sets classics. It began to be used in

industry, medicine, the establishment of expert

systems in the mid-1970s and will see its widespread

use in the 1990s: autofocus, pressure cookers,

autonomous mobile systems, decision systems,

diagnosis, recognition, education, etc.

Its operation can be summarized in three main

steps (see

Figure 1

546

Youssef, H., Moumen, A. and Gretete, D.

Fuzzy Logic for Academic Orientation and Its Impact on Success: Content Analysis.

DOI: 10.5220/0010742200003101

In Proceedings of the 2nd International Conference on Big Data, Modelling and Machine Learning (BML 2021), pages 546-551

ISBN: 978-989-758-559-3

Figure 1: Mamdani fuzzy logic model.

Fuzzifier: this first step consists in transforming

the variables (input and output) into linguistic

variables, is mainly carried out based on statistical

observations (or by learning, supervised or not, to

group the values of a variable into homogeneous

categories) or expert opinion.

Inference Mechanism: construction of rules (and

results) based on linguistic variables, assigning a

truthfulness to each direction, and then aggregating

the regulations to obtain a result single (linguistic).

Defuzzifier: the last step of the fuzzy logic aims at

transforming the final activation curve obtained

during the aggregation step into an actual value.

2.1 Operation of Fuzzy Sets

As in classical set theory, fuzzy sets (

C. Servin,

2018)

have their operations, including union, intersection,

and complement. Unlike the process on classical sets,

the methods specific to fuzzy sets rely on the

membership function. (

Figure 2

) shows the operation

relating to conventional assemblies. (

Figure 3

)

indicates a possible type of operation dealing with

fuzzy sets.

Figure 2: Classical sets, 𝐴 ∪ 𝐵 = {𝑥, 𝑦, 𝑧}; 𝐴 ∩ 𝐵 =

{𝑦}; 𝐴̅ = {𝑧}

Figure 3: Fuzzy sets.

If A and B, then C. The maximum degree

of certainty of C is the lesser of the

degree of certainty of A «μA» and the

degree of certainty of B «μB»:

CA ∩ B  min 

B

(1)

If A or B, then C. The maximum degree

of certainty of C is the highest of the

degree of certainty of A and the degree

of certainty of B:

𝐴

∪𝐵 𝑚𝑎𝑥



𝜇𝐴, 𝜇𝐵



(2)

If it is not A, then C. The maximum

degree of certainty of C is that which is

deduced from the degree of certainty of

𝐴

̅  1  𝜇𝐴

)

2.2 Logic and Education

History teaches us that education systems have

changed a great deal over time, and we must now

manage these systems to ensure the success of as

many people as possible.

The evaluation of students' academic performance

is the most important for the higher education

institution to have a high reputation and ranking

(

Ishak, 2015)

A priority task is to make a model, by the expert

system more than an academic diagnosis, to provide

career guidance.

Therefore, it is necessary to evaluate different

psychological factors such as intelligence, patience

and perseverance, learning ability and speed of

problem-solving (

J. Sasi Bhanu, 2019)

The model where the traditional diagnosis we

obtained with the only parameter of the cut-off score

remains incomplete to determine the student's career

guidance. The fuzzy set theory illustrates its

advantages in this situation by adding human

parameters to the model.

3 METHODOLOGY

The working method is developed using textual

analysis techniques, to locate our topic about a similar

project on a corpus of references (run on Zotero

library), from the most known: Springer Link,

Scopus, IEEE Xplore, Web of Science, Jstor (see

Figure 3

The interest of this work is analyzed in our library

by type of reference by year of publication and by

theme (occurrence of words). Finally, compare the

different models used and discuss the results.

Fuzzifier

Fuzzy

Mechanism

Defuzzifier

Fuzz

rules

Input data

Output data

Fuzzy Logic for Academic Orientation and Its Impact on Success: Content Analysis

547

Figure 3: The working method.

After the analysis, most of the articles selected

from the corpus are under the theme of education and

vocational guidance (see Table1). We will descend

the problem of each work, the theoretical model used

with its different parameters; thus, the results

obtained.

Table 1: List articles on career guidance and career paths.

4 RESULTS AND DISCUSSIONS

The objective of the work (A. Tarasyev, 2018) is to

analyze a set of economic factors, which influence the

student's decision to change his educational path. To

estimate the possibility for each student to change

their educational path, they developed an approach

based on the fuzzy logic model of the Mamdani type

(see

Figure 4

Fuzzy logic model for Educational Path Change

Probability based on three variables in input:

 Direction: Variable that affects a student's

decision to change programs.

 Budget Support Possibility: Variable

represents the demand that these students

obtain budgetary support from the

government or the university to pay for their

studies.

 Expected Salary: Variable who estimate the

number of money students earns upon

graduation from their current program.

Figure 4: Fuzzy logic model of the Mamdani type to

estimate the possibility for each student to change his

educational path.

Career guidance requires taking into account country-

specific variables. Therefore, this study (M. Peker,

2017) aims to develop an automated system for career

guidance activities that can be applied to students in

the final year of secondary school. By introducing an

artificial intelligence approach, using a fuzzy logic

model of the Mamdani type (see

Figure 5

Fuzzy logic model for the prediction of vocational

guidance based on three variables in input:

 GPAs (MB, SSB): this is the average of the

annual average values of Grade 9 students in

classes based on mathematics (physics,

chemistry, biology and math) and social

studies (history, geography and language

arts).

 Teacher views (TV): the variable that shows

students' disposition towards specific careers.

The value assigned to the variable, obtained

through individual interviews with students

and by collecting the opinions of parents and

other teachers who knew the student.

 Interest values: Career guidance variable

estimated from a questionnaire with 150

questions whose answers are in the form of <<

yes, no, and sometimes >>.

year Author

2018 Tarasyev, Alexandr A.; Agarkov, Gavriil A.;

Ospina Acosta, Camilo A.; Koksharov,

Viktor A.

2017 Peker, Musa; Guruler, Huseyin; Sen, Baha;

Istanbullu, A

han

2019 Tajul Rosli Razak,Iman Hazwam Abd.

Halim,Muhammad Nabil Fikri

Jamaluddin,Mohammad Hafiz mypapit

Ismail.

2019 Sulaiman, M.S., Tamizi, A.A., Shamsudin,

M.R., Azmi, A.

2019 J. Sasi Bhanu, V. Chandra Prakash, and J.

Sastry

2017 D. Calvo, L. Quesada, G. López, and L. A.

Guerrero

154

articles from

Most database

Six

articles

selected

Nvivo

Analyse

importer (*.RIS)

Zotero

bibliothèque

BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BML’21)

548

Figure 5: Model fuzzy logic, Mamdani type using WEB-

CGS, for the prediction of vocational guidance.

This document (R. Tajul Rosli, 2019) aims to explore

the HFS model (

Figure 6) for the career path

recommendation system (CPRS), particularly for a

recommendation of the web programmer career.

The CPRS is a recommendation system that

provides direction in choosing their career via skills

assessment based on the multiple-choice question

technique.

Questions Skills Scale (Input Variable)

 Q1 Design and develop a web base.

 Q2 Handle the whole web project from start

to rollout.

 Q3 Skill and knowledge in PHP, HTML,

CSS, Javascript and MySQL.

 Q4 Good in problem-solving,

communication, interpersonal and

organization skills.

 Q5 Up to date with the latest web technology

trends and programming techniques.

Figure 6: Career Path Recommendation System (CPRS) –

FLS topology.

The main objective of this study (M. S. Sulaiman,

2019) is to help students who often face a dilemma

and confusion in choosing the right path. It is

essential to choose the right course and the proper

education in the training years. Therefore, a course

recommendation system using Fuzzy Logic is to

make the right choice (see Fig9).

The model for recommending courses using fuzzy

sets is based on two characteristics, which are

competence and interest. Skills are contributed from

the eight courses based on the diploma results,

classified as programming, problem-solving, and

computer and data analysis skills. The exciting part is

based on a questionnaire with twelve questions which

can be classified according to Programming, Critical

thinking, Organizing information, Practical work,

Data processing, networking.

Figure 7: The model for recommending courses using fuzzy

sets.

A thorough reading of each work allowed us to

determine the problem, the model (sample, variable

and rules), the tool used, the results obtained and the

limits of each model (see table 2):

Table 2: Reading of each work.

Authors Method and

variable

implement Sample Limit

Tarasyev,

Alexandr

A.;

Agarkov,

Gavriil A.;

Ospina

Acosta,

Camilo A.;

Koksharov,

Viktor A.

(2018).

Mamdani

type Three

variable in

input and

one variable

in output

(see Fig5)

python 5300

students

Descriptive

model.

Peker,

Musa;

Guruler,

Huseyin;

Sen, Baha;

Istanbullu,

Ayhan

(2017).

Mamdani

type using

WEB-CGS:

Three

variables in

input and

three

variables in

output (see

Fig6)

Matlab Twenty

students to

predict and

300 students

for

modelling.

In this study,

the overall

results were

determined

just about

four

occupational

areas. It

would

therefore be

helpful to

carry out a

comprehensi

ve analysis

applied to all

occupational

areas.

Fuzzy Logic for Academic Orientation and Its Impact on Success: Content Analysis

549

Table 2: Reading of each work (cont.).

Authors Method and

variable

implement Sample Limit

Tajul Rosli

Razak,Iman

Hazwam

Abd. Halim,

Muhammad

Nabil Fikri

Jamaluddin,

Mohammad

Hafiz

mypapit

Ismail 2019.

Mamdani

type

Explore and

compare

HFS and

FLS for the

career path

recommend

ation system

(CPRS)

Matlab Comparativ

e model.

This

exploratory

study has

shown that

HFS is more

interpretable

than FLS.

Still, it does

not cover all

aspects of

interpretabil

ity, such as

the semantic

meaning of

fuzzy sets

and

intermediate

variables.

Sulaiman,

M.S.,

Tamizi,

A.A.,

Shamsudin,

M.R., Azmi,

Mamdani

type

Matlab A positive

achievemen

t among a

sample of

50 students

which are

ased on

vigour

J. Sasi

Bhanu, V.

Chandra

Prakash, and

J. Sastry

(2019)

TTT-GP-

CGS

Java Students of

our K L

University.

D. Calvo, L.

Quesada G.

López, and

L. A.

Guerrero

(2017)

System

Using,

Google

Home and

Telegram.

The proxy

service

(API.AI)

used for

communic

ation and

dialogue

computatio

n, and the

personality

evaluator

(IBM

Watson).

Freshmen.

One of the

most

disliked

attributes of

the system

was the

extent of the

conversation

, as the users

perceived it

as a long

process.

And

27.78% of

the

participants

expressed

that they did

not like the

personality

analysis.

The first model is a descriptive economic model

of the Mamdani type: modelled on a sample of 5300

students, it estimates the probability that a student can

change their careers.

The second model is a Mamdani type prediction

model (three input variables and three output

variables), modelled on a sample of 300 students. It is

used to predict the academic guidance of the 9th year

of vocational high school.

The third model is a model that compares two FLS

and HFS architectures of the Mamdani type for the

prediction of guidance to the web programmer

profile.

The fourth model helps students for an optimal

orientation, based on two axes: a recommendation of

the courses, more the competence and the student's

interests.

The outcome of this research work (

J. Sasi Bhanu,

2019)

is an expert system called Tic-Tac-Toe Game

Playing Career Guidance System (TTT- GP-CGS)

that is useful to assess the psychological factors of the

student through Tic-Tac-Toe Game Playing, build the

cognitive model of the student and predict the

appropriate career(s) for the student.

Finally, in this article (

D. Calvo, 2017) the model

presents a new and different approach to career

guidance systems. it uses Google home as a voice

interface and Telegram as a text interface to generate

conversation between users and a bot for career

guidance.

5 CONCLUSIONS

The fuzzy set theory aims at dealing with qualitative

situations and takes into account human variables.

However, it appears that this theory is the best

approach to realize a model for predicting academic

orientation. The model will help the student make the

right choice of direction and succeed in his career.

In the previous section, we could compare four

models of the Mamdani type, which were realized in

this way. This comparison aimed to situate our

problem in relation to the existing work and to choose

the model close to our case, on which it is based.

This comparison allowed us to see that we could

opt for a recommended course system and academic

path for the engineering cycle at the National School

of Applied Sciences (NSAS), Ibn Tofail University

(ITU) kenitra.

The system can be classified into three parts: the

competencies, the student interest, and the professors'

recommendations (See Figure 8).

BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BML’21)

550

Figure 8: System of recommended academic paths.

REFERENCES

L. A. Zadeh, 1965 “Fuzzy sets,” Information and Control,

vol. 8, no. 3, pp. 338–353, Jun, DOI: 10.1016/S0019-

9958(65)90241-X.

I. Ishak, 2015« Application of fuzzy logic to student

performance in calculation subjects ».

C. Servin, O. Kosheleva, and V. Kreinovich, 2018, “In

Fuzzy Decision Making, General Fuzzy Sets Can Be

Replaced by Fuzzy Numbers,” p. 7.

G. J. Klir, 2006. « Fuzzy Set Theory », in Uncertainty and

Information: Foundations of Generalized Information

Theory, IEEE,

M. F. Martin and T. Ellen, 1994 Fuzzy Logic A Practical

Approach. Elsevier. DOI: 10.1016/C2013-0-11164-6.

A. Tarasyev, G. A. Agarkov, C. A. Ospina Acosta, and V.

A. Koksharov, 2018: Fuzzy Logic and Optimization of

Educational Paths. In IFAC, vol. 51, no. 2, pp. 511–516,

Jan. DOI: 10.1016/j.ifacol.2018.03.086.

M. Peker, H. Guruler, B. Sen, and A. Istanbullu 2017: A

New Fuzzy Logic Based Career Guidance System:

Web-Cgs. In The. Vjesn. Vol. 24, no. 6, pp. 1863–1868,

Dec. DOI: 10.17559/TV-20151105201325.

R. Tajul Rosli and I, Hazwam Abd. Halim May 27, 2019,

“An Exploratory Study of Hierarchical Fuzzy Systems

Approach in A Recommendation System,”

M. S. Sulaiman, A. A. Tamizi, M. R. Shamsudin, and A.

Azmi, "Course recommendation system using the fuzzy

logic approach," Indonesian Journal of Electrical

Engineering and Computer Science, vol. 17, no. 1, pp.

365–371, 2019, DOI: 10.11591/ijeecs.v17.i1.pp365-

371.

J. Sasi Bhanu, V. Chandra Prakash, and J. Sastry,

“Assessing the intelligence of a student through TIC-

TAC-TOE game for career guidance,” International

Journal of Electrical and Computer Engineering

(IJECE), vol. 9, no. 6, Art. no. 6, Dec. 2019, doi:

10.11591/ijece.v9i6.pp5545-5551.

D. Calvo, L. Quesada, G. López, and L. A. Guerrero,

“Multiplatform Career Guidance System Using IBM

Watson, Google Home and Telegram,” in Ubiquitous

Computing and Ambient Intelligence, Cham, 2017, pp.

689–700. doi: 10.1007/978-3-319-67585-5_67.

Fuzzy Logic for Academic Orientation and Its Impact on Success: Content Analysis

551