A Decision Support System for Portuguese Higher Education Course

Selection – First Round

João Pedro Silva

, Filipe Portela

and Manuel Filipe Santos

Information System Department, University of Minho, Azurém, Guimarães, Portugal

Algoritmi Center, Information System Department, University of Minho, Azurém, Guimarães, Portugal

Keywords: University, Education, Decision Support System, Higher Education, University Courses, Portugal

Education.

Abstract: Application for higher education courses is a delicate and important process. In this phase students face a

great number of possibilities of choice. In order to help students in the decision making process a decision

support system called C.U.R.S.O was developed. Based on a structured questionnaire the system determines

de profile making use of a knowledge base and proposes a ranked list of the most suitable courses. Input

variables take into account not only social and economic aspects but also the aptitude for a particular area.

The first version of C.U.R.S.O was tested by hundreds of Portuguese students and schools. This paper pre-

sents and discusses the results attained during the first round.

1 INTRODUCTION

A Decision Support System (DSS), more precisely

the decision models associated with it, will never be

effective if the variables that affect the operating

model and the associated data are unframed from the

context in which it occurs. Any system of this na-

ture, treating data incorrectly will present to the user

misaligned answers from their expectations and

requirements.

The main objective of this project is the devel-

opment of decision models based on the comprehen-

sive study of the environment in which it is inserted,

more specifically, the variables that a system must

consider and their correspondence to the data model.

C.U.R.S.O. (Centre for Universal Gathering of

Oriented Suggestions) is a DSS designed for assist-

ing students who are in transition year between col-

lege and high school to choose the most suitable

course according to their profile.

This paper is divided into five sections. The first

section introduces the problem. Then, in the second

section a set of related concepts and some similar

tools are presented. The third section presents the

DSS and the knowledge phases. The fourth section

present the questionnaire results and makes a review

and an analysis of the questionnaire results. The

sixth section presents the system features and the

prototype deployed. Finally, some concluding

remarks are done and future work is outlined.

2 BACKGROUND

Decision Support Systems

2.1

In his work on data structures, Alfred V. Aho (Aho

et al., 2001), considers a decision model as a se-

quence of operations that branch execution based on

comparisons of data, referring to a generic model

simply making decision analyzing a large number of

variables regarding to a previously order established.

Decision is a choice among alternatives based on

estimated weights of these alternatives and the deci-

sion model will help to establish parameters for this

choice generating these estimates, changes or com-

parison and choice. A decision model is character-

ized by (Sprague and Carlson, 1982):

 Treat unstructured problems;

 Combining the use of modelling techniques, with

traditional functions of access and retrieval of in-

formation;

 Create a user-friendly interface that allows interac-

tion with the user;

 Emphasize the flexibility and adaptability for

monitoring changes, both the environment and the

different needs of use by the users.

360

Pedro Silva J., Portela F. and Filipe Santos M..

A Decision Support System for Portuguese Higher Education Course Selection – First Round.

DOI: 10.5220/0004545503600367

In Proceedings of the International Conference on Knowledge Discovery and Information Retrieval and the International Conference on Knowledge

Management and Information Sharing (KMIS-2013), pages 360-367

ISBN: 978-989-8565-75-4

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

The theoretical component of this project is

mainly based on two authors: Herbert A. Turban

Sharda and Simon (Simon, 1978; Turban, Sharda, &

Delen, 2011). Simon was the first to carry out exten-

sive studies in decision models and established a

methodology for their development. The model of

Simon is divided into four major steps (Simon,

1978). Turban (Turban, Sharda, & Delen, 2011)

added a fifth step that allows for monitoring the

system after its implementation:

1) Intelligence - Collection of information inherent

to the process;

2) Design - Design of decision model;

3) Choice - in this phase the alternatives previously

developed in the design phase are evaluated in

order to choose one of them;

4) Implementation - Implement the selected course

of action. Normally this phase includes an im-

plementation plan.

5) Monitoring - this phase evaluates the implemen-

tation and can contribute to improve the models,

being the results used in the intelligence phase of

next iterations.

Accessing to Higher Education

2.2

in Portugal

The National competition for accessing higher edu-

cation at public sector, Portuguese legislation (Ordi-

nance nº 195/2012 of the Ministry of Education and

Science) reports: "(...) To apply for courses in public

higher education institutions is done through a na-

tional competition organized by the General Direc-

tion of Higher Education, unless in the case of ex-

ceptions under the rule of law. " Briefly, the applica-

tion process in Portuguese public higher education is

accomplished through the seven steps.

This process starts after completion of internal

studies in the 12th year of schooling and finishes

after the application period by DGES with the selec-

tion of candidates. The candidates are ranked per

course. The process is repeated in a second round for

the open positions.

Similar Tools

2.3

After an exhaustive search, it was not possible to

found a system with similar characteristics.

In Portugal only are available two online services

to help users searching for a higher education

course: the site of the General Direction of Higher

Education (DGES) and the Office of Higher Educa-

tion. These services are based on pre-established

criteria in order to filter the questions. In both sys-

tems the user can search for courses filtering them

by location, type of institution, field of study and

specific ingression of the same. Similar systems can

found in other countries.

Table 1 presents some tools to help students and

if they is based in a search engine or in a question-

naire.

Table 1: Similar tools.

Institution Country Search Questions

Empresário México X

Univafu México X

Universidades.com Argentina X

UK Course Finder United Kingdom X

Go2Uni States X

Guia do Estudante Brazil X

Universities Guide Australia X

Guia da Carreira Brazil X

Gab. Ensino Supe-

rior

Portugal X

DGES Portugal X

Psychometric Tests

2.4

Psychometric tests have been used since the early

part of the 20th century and were originally devel-

oped for use in educational psychology

(Psychometric-success, 2012; Them, 2011). Psy-

chometric tests are very usual instruments to help in

the decision process. Helped by a psychologist

through a series of questions and mental exercises,

candidates identify a number of potential profes-

sions.

Psychometric tests aim to measure attributes like

intelligence, aptitude and personality (Psychometric-

success, 2012). However, this type of tests isn’t

direct competitor of the DSS in development. Alt-

hough it is for the same user segment, they are for

different situations. The DSS can be more general

and include the psychometric tests.

2.5 RIASEC (Holland Codes)

RIASEC is a test (system, 2012) that has the objec-

tive to answer to the question: “Which Career Path-

way is right for you?” This test is composed by a set

of different questions from different areas.

The final output of this approach consists in one

of six areas (Holland and Gottfredson, 1992): Real-

istic (Doers), Investigative (Thinkers), Artistic (Cre-

ators), Social (Helpers), Enterprising (Persuaders),

and Conventional (Organizers). With the result the

students get a set of areas to continue the academic

ADecisionSupportSystemforPortugueseHigherEducationCourseSelection-FirstRound

361

68%

29%

Yes

45%

53%

Yes

Iwillnotapply

75%

16%

Yes

Iwillnotattend

career according to the classification: R, I, S, A or C.

3 DECISION SUPPORT SYSTEM

DSS is a computer program that uses knowledge to

solve complex problems. In this case, the knowledge

is acquired in order to help the students to choose

the most suitable course. In the first round a very

simple decision model based in rules was imple-

mented. In order to develop a knowledge based

DSS, the following methodology was adopted

(Rezende et al., 2003)

 Planning (Identification of the knowledge domain;

Selection of the Team Development; Selection of

the Tool Development)

 Knowledge Acquisition (Identification; Conceptu-

alization; Formalization;

 Implementation (Representing the knowledge in

the tool; Implementing the Interface; Documenting

of the KBS;

 Test and refinement (Validation and Verification;

Refinement of SBC)

4 STUDENTS PROFILE

In order to better understand the target users of the

DSS, a study on Portuguese student population has

been carried out. Two different questionnaires were

deployed to different populations. The first was

designed to students that are:

 in the High Scholl (HS) (10

to 12

);

 studying in Portugal;

The objective of the first questionnaire was to

understand if the students have notion / idea about

which course to choose in the Higher Education

(HE) and if they know any tools to help them. The

second questionnaire was made to students that are:

 In the Higher Education (Lic or MsC);

 From the most important Portuguese’s HE Insti-

tutes;

The objective of the questionnaire made for HE

students was to analyze all Portuguese HE institutes

and had as main goal understand if the students

made the right decision and if they thought about

changing course. Both the questionnaires were made

using the internet and divulgated through email,

forums and directly in some institutions. As men-

tioned, in this study two different types of students

(profiles) were considered:

CS – High school students applying for a higher

education course; and

HS - Higher education students who intend to

change course;

In order to percept the requirements for the deci-

sion support system both types of students were

inquired. CS community has been confronted with

the following question: Do you know what area you

intend to choose?

By observing Figure 1, 29% of respondents said

they still have no idea of the area to which they will

Figure 1: Do you already know what area to choose in

HS?

apply, in other words, about five months to conduct

nominations, three in ten students do not know yet

how to fill out the application.

Figure 2: Do you already know what course to choose in

CS?

To further understand the problem, CS students were

asked if they knew the course in the concrete that

would apply. Figure 2 shows the number of students

who have not yet decided. More than half of stu-

dents in secondary education still don’t know what

course will choose to apply in higher education.

Figure 3 shows that 16% of students have al-

ready changed the course at least once time, and 9%,

although has not changed, ever thought about it.

Figure 3: In your academic career, did you ever changed

of course?

Based on these results, the HS students were asked,

if when applied to higher education, they previously

KMIS2013-InternationalConferenceonKnowledgeManagementandInformationSharing

362

knew what course would put in the application (fig-

ure 4).

Figure 4: Did you know that the way you'd apply at the

time of application?

Final results demonstrate that a DSS to support the

application process to Higher Education is very im-

portant for both profiles CS and HS.

Exploring Profiles

4.1

To gather information about potential users of the

DSS it was used an online survey to reach the largest

possible number of individuals. This phase of the

project was essential to:

 Understand the behaviour of those who will use

the system;

 Create a tool in order to identify the best course for

future students in Higher Education, and what kind

of variables that tool must consider in terms of de-

cision models.

Thereby, it was undertaken a nationwide survey

through an email message sent to all high schools

and higher education establishments of Portugal, in

order to communicate with all their students.

Tables 2 and 3 present the characteristics of the

students that answered the survey in terms of CS and

HS students. For example, the CS students have an

average of 17.2 years, are mostly female and are

studying in the 12

year. In this context only it

wasn’t possible obtain answers from two of twenty

districts.

Table 2: High school students.

Nº of respondents 381

Sex Male - 35% ; Female - 65%

Age average 17,2 years

Scholar year 10

- 7% ; 11

- 20%; 12

- 73%

Portugal Districts 18 of 20 (total) answered

Next steps involved direct inquire to both profiles

(CS and HS) in order to elicit the most important

variables/qualifiers to form de decision model.

Table 3: Higher Education students.

Nº of respondents 1699

Sex Male - 33% Female - 67%

Age average 23,03 years

Stage of studies 1º- 56%; 2º- 28%; 3º- 16%

Districts

Every single higher education

establishments of Portugal an-

swered the survey.

CS Profile

4.2

CS community have been asked on the following

points (Q1 to Q3). Q1 (Figure 5) was used to under-

stand if employment rate and institution reputation

were the most important for to the students decision.

Q2 and Q3 show that students are interested in a tool

that can help them defining a profile based in the

tests.

Q1: How much will these factors influence your

choice?

Figure 5: How much will this factors influence your

choice, when you attend to higher education?.

Q2: Will you consider useful a tool (web or

smartphone) that would help you (through a series of

questions) to choose the most suitable course for

you, or to ensure that the course you have chosen is

really the most adequate?

Table 4: Will the tool be usefull?.

YES 93% NO 7%

Q3: Considering that this system will scan your

profile, what aspects should be taken into account to

analyse which course is right for you?

82%

18%

Iknewwhatcourseto

choosepreviously

Ididn'tknowwhatto

chooseandimakemy

decisionaswhatmy

averageallowed

0,00%

5,00%

10,00%

15,00%

20,00%

25,00%

30,00%

35,00%

40,00%

45,00%

50,00%

Nothing Some Enough Alot

ADecisionSupportSystemforPortugueseHigherEducationCourseSelection-FirstRound

363

Figure 6: What aspects should be taken into account to

analyze which course is right for you?.

HS Profile

4.3

HS community have been asked on the following

points (Q4 to Q11). The next seven questions are

about the Higher Education and change of course.

For example, Q5 (table 6) shows that students

normally change course in the first year. Q6 (table7)

shows that almost 50% of the students when change

course also change area. Q8 (figure 7) shows the

main reasons for the change. Experience of family or

friends and the teaching staff are not the only rea-

sons for change, reputation of the institution and the

syllabus of the course are also refereed. Q9 (table 9)

shows the importance of having a tool to help defin-

ing the best profile and to choose the most suitable

course /area.

Q4: Have you ever changed course during the

academic journey?

Table 5: Changes of course.

Yes 16%

o 75%

o but I had alread

thou

ht about i

Q5: Which academic year did you changed the

course?

Table 6: Year of change.

69% 2

20% 3

1% 5

Q6: Why have you changed course?

Table 7: Reason of change.

I changed to the same course but in another

educational establishment

12%

Course in the same area but different studies 39%

Also changed area 49%

Q7: How do you characterize the information you

had about the course in which you entered at the

time you made your application to Higher Educa-

tion?

Table 8: Quality of the Information.

Null 2,47% Good 25,31%

Weak 28,31% Excellent 3,59%

Enough 40,32%

Q8: How much the following factors will influence

your choice?

Figure 7: How much the following factors will influence

your choice?.

Q9: Have you ever undergone a psychometric test?

Table 9: Have you ever undergone a psychometric test?.

Yes and the result was that I followed in

academic caree

30%

Yes but the result turned out to not be

what I followed in m

academic caree

32%

I never attend one 38%

Q10: Did you consider useful a tool (web or

smartphone) that would help you (through a series of

questions) to choose the best suited course(s) to you

or to ensure that the course you have chosen is the

more indicated?

Table 10: The tool will be usefull?.

Yes 86% No 14%

Q11: Considering that this application would parse

your profile which aspects should be taken into ac-

count to analyse the most suitable course for you?

23,58%

44,07%

28,44%

3,91%

0,00% 20,00% 40,00% 60,00%

Other

Theareathat

you'restudyingin

highschool

Yourtastesfora

seriesofgeneric

areas(technology,

healthsciences...)

Willingnesstogo

toschooloutof

yourdistrict

0,00%

5,00%

10,00%

15,00%

20,00%

25,00%

30,00%

35,00%

40,00%

45,00%

Nothing Some Enough Alot

KMIS2013-InternationalConferenceonKnowledgeManagementandInformationSharing

364

Figure 8: What aspects should be taken into account to

analyze which course is right for you?.

Discussion

4.4

Due to the high number of responses and geograph-

ical diversity of the online survey, C.U.R.S.O was

revealed to be an excellent tool for analysing the

national panorama of those who applied recently, or

are about to apply in future for higher education

courses.

This instrument can be used to find evidences

that prove the existence of a high number of students

who are about to perform the application and still do

not know in concrete what course to choose.

For the respondents the most important indica-

tors to choose a course are very similar. High school

students select in the Q1 the employment rate (85%)

and the reputation of the course and institution

(78%) in the Q3 the possibility of experiencing ge-

neric areas (44%) was chose as the main variable to

decide. Higher education students gave more im-

portance to the reputation of the course and institu-

tion (Q8) (76%) and they decide based on generic

areas (Q11) (44%).

More specifically, students who are currently at-

tending high school, 3 in 10 (30%) have no idea of

the area that they will being study in higher educa-

tion, and of these 5 per 10 students (50%) do not

know the course in concrete that they will apply for.

Regardless to students who are currently attend-

ing higher education, about 2 out of 10 (20%) have

changed at least once the course, and from those 5

per 10 (50%) also changed the area. This means that

a high number of students undertake their applica-

tion for higher education wrongly.

From the above evidences we can foresee a huge

potential of C.U.R.S.O.

5 C.U.R.S.O.

First an expert system was implemented in format of

a tool called C.U.R.S.O. (Universal Collection Cen-

ter of Guided Suggestions). The main objective was

to assist individuals who are attending secondary

school in the process of Application to higher educa-

tion, particularly with regard to the choice of the

course they will put in the application.

Objectives

5.1

The creation of this tool has three goals based in the

characteristics of the prospective DSS:

a) Maintain Decision Support Models

The main objective of the system is to maintain

modes for assisting students find the most suita-

ble course;

b) Collecting Information

The information collected from the users can be

used to produce indicators very useful for specif-

ic entities, such as High Schools, Universities,

Ministry of Education and others.

c) Adaptability

The system should improve its performance year

after year using the collected information in or-

der to optimize the decision models.

DSS Design

5.2

5.2.1 Planning

In this phase it was understood the context of the

problem in a Higher Education approach and defined

the research team as also the development tool:

Corvid.

5.2.2 Knowledge Acquisition

and Representation

The main data source was the DGES website where

was queried data relevant to all courses of Portu-

guese Public Education. This knowledge was used in

the logic blocks. Two types of variables were con-

sidered:

Informative

 Higher Education Institution (HEI) (Name,

Ranking, District);

 Course (Name, Employability Rate, Average,

Scientific Area, prerequisites and specific sub-

jects).

Standings

 High school area;

 Students Desires.

25,79%

44,41%

23,86%

5,94%

0,00% 20,00% 40,00% 60,00%

Other

Theareayou're

studyinginhigh

school

Yourtastesfora

seriesofgeneric

areas(technology,

healthsciences...)

Availabilitytogo

toaschooloutof

yourdistrict

ADecisionSupportSystemforPortugueseHigherEducationCourseSelection-FirstRound

365

Figure 9: Inference Rules.

Figure 10: Inference Rules.

Figures 9 and 10 show an examples of the inference

rules (if then else). First a set of rules is used to

define the student profile then, a scoring criteria

was used to present the final suggestion.

5.2.3 Implementation

At this stage a web based prototype was deployed

(Figure 11) and a set of tests were done in order to

evaluate the tool. A prototype of C.U.R.S.O. was

made available online at

http://ulife.webnetpt.com/curso/index.php. From

July to September 2012 the site received 9714 visits.

The system was developed in the free version of

Exsys Corvid (Vahidov and Ji, 2005) platform. To

give support on the use of the system, the users

could use email and a Facebook page to keep in

touch with the developers and to send feedback on

system usage.

5.2.4 Tests and Refining

Released the prototype it was necessary to be aware

that the performance has to be monitored. In this

phase also was carried out some tests and refinement

Figure 11: Prototype Interface.

to the system in order to find which model is the

most suitable to carry on. To this phase it was

planned a series of tasks to accomplish: creating an

email to which users can send instances of problems,

suggestions or comments, and monthly monitoring

of the number of site visits. Subsequently a plan was

developed to further system improvements.

Main Features

5.3

The system presents as main features:

 The use of logical blocks;

 Knowledge Base;

 User-friendly interface (web);

 Informative and Standings questions;

 Feedback form;

 Capability to be improved.

6 CONCLUSIONS

The need for a DSS like C.U.R.S.O is notorious to

help the students in the process of accessing to high-

er education. The process of choosing a course is

very complex and should be done in a very short

period of time.

This compromises the future / motivation of the

students at academic and professional levels. 93% of

students thought that C.U.R.S.O. is very important

to help in the decision process and 30% considered

the available information insufficient to support the

choice.

C.U.R.S.O can, in the future, reduce the number

of students that swap over courses during their aca-

demic life (estimated at 25%).

7 FUTURE WORK

Further work includes the research on other scien-

tific areas like psychology to develop better decision

models. The interoperation with governmental data-

KMIS2013-InternationalConferenceonKnowledgeManagementandInformationSharing

366

bases will be also explored in order to assure the

appropriated information about the courses offered.

A second round is being prepared in order to test the

models in a broader environment.

ACKNOWLEDGEMENTS

This work is supported by FEDER through Opera-

tional Program for Competitiveness Factors –

COMPETE and by national funds though FCT –

Fundação para a Ciência e Tecnologia in the scope

of the project: FCOMP-01-0124-FEDER-022674.

The work of Filipe Portela was supported by the

grant SFRH/BD/70156/2010 from FCT.

REFERENCES

Holland, J. L., & Gottfredson, G. D. (1992). Studies of the

hexagonal model: An evaluation (or, the perils of

stalking the perfect hexagon). Journal of vocational

behavior, 40(2), 158-170.

Psychometric-success. (2012). Psychometric tests.

Retrieved 08-Mai-2013, 2013, from http://www.

psychometric-success.com/psychometric-tests/psycho

metric-tests-introduction.htm

Rezende, S. E. A., Pugliesi, J. B., & Varejão, F. M.

(2003). Sistemas Baseados em Conhecimento.

Sistemas Inteligentes: fundamentos e aplicações.

Editora Manole. Barueri, SP.

Sprague Jr, R. H., & Carlson, E. D. (1982). Building

Effective Decision Support Systems: Prentice Hall

Professional Technical Reference.

system, U. o. H. (2012). RIASEC Test.

Them, W. E. U. (2011). Psychometric tests.

Vahidov, R., & Ji, F. (2005). A diversity-based method for

infrequent purchase decision support in e-commerce.

Electronic Commerce Research and Applications,

4(2), 143-158.

Simon, H. A. (1978). Decision Support Systems, 50(3),

Turban, E., Sharda, R., & Delen, D. (2011). Decision

Support and Business Intelligence Systems. (S. Yagan

& E. Svendsen, Eds.) (p. 715). Prentice Hall. Aho, A.

V, Laboratories, B., Hill, M., Hopcroft, J. E., York, N.,

& Ullman, J. D. (2001). Data Structures and

Algorithms Use of the Book.

Simon, H. A. (1978). Rational Decision-Making in

Business.

Turban, E., Sharda, R., & Delen, D. (2011). Decision

Support and Business Intelligence Systems. (S. Yagan

& E. Svendsen, Eds.) (p. 715). Prentice Hall.

ADecisionSupportSystemforPortugueseHigherEducationCourseSelection-FirstRound

367