A NOVEL LEARNER SELF-ASSESSMENT APPROACH
Application to Pratical Works
Karima Boussaha
1
, Farid Mokhati
1
and Nora Taleb
2
1
Département d’informatique, Universite Larbi Ben m'hidi, Oum El Bouaghi, Algérie
2
Département d’informatique, Université Badji Mokhtar, Annaba, Algérie
Keywords: CEHL, e-Learning, Remote Practical Works, Assessment, Object-Oriented Programming, UML Diagrams.
Abstract: This paper presents a novel approach for learner self- Assessment. The aim of this approach is to offer to
learners an evaluation tool as a CEHL (Computing Environment for Human Learning) allowing comparing
their productions (C++ object-oriented programs) with those elaborated by teacher. Considered as part of
our platform E-TéléTP@AALP (Environnement des TéléTP Appliqués A l’Apprentissage des Languages
de Programmation), the tool we are developing allows essentially: (1) generating two UML diagrams (UML
class diagram and communication diagram) from the learner’s program, and (2) applying some matching
algorithms for measuring degrees of similarity and dissimilarity between learner’s diagrams and teacher’s
diagrams.
1 INTRODUCTION
Computing Environment for Human Learning
(CEHL) is an important research domain. Its
educational, social and economic importance has
attracted much interest from different research
communities.
Using novel technologies, the purpose of a
CEHL is to help and assist learners, teachers and
tutors. With the development of platforms, teachers
are required to make available to learners courses
and class rooms in various disciplines (Tchounikine,
2009).
Actually, the number of distance courses and
class rooms has dramatically increased.
Consequently, platforms’ designers are generally
more concerned about courses and class rooms,
practical works are completely omitted.
The teaching of practical works is revealed
essential in the scientific and technical trainings, in
presential as in distance, and answers a true need.
Often, learners are deprived from this essential
teaching and this is due to the problem of
availability of the assistants, the problem of
obstruction of learners, the material is expensive and
cannot be duplicated. In order to minimize these
remote problems, thus teaching must answer to these
needs (Guillaume, 2006).
Although it represents an important activity
during learner training, learner assessment has not
yet been adequately addressed. In this context, a
very few works have been emerged in the literature.
We quote among others the works of Auxepaules
(2009) and Tanana, Delestre, Pécuchet and
Bennouna (2009).
Auxepaules (2009) developed a CEHL called
“Diagram” dedicated to the training of OOM
(Object-Oriented Modelling) concepts starting from
a textual specification. It proposes an automatic
method for analyzing learner’s diagram by
comparing and matching class diagram components.
The proposed method is implemented as an
integrated component of “Diagram” called “ACDC”
(Automatic Class Diagram Comparator).
Tanana et al. (2009) proposed a formative
assessment method of learner’s skills using
algorithms of supervised classification. In order to
validate their method, they chosen numerical
electronics as application domain, essentially, the
assessment of numerical electronic diagrams
produced by learner during its practical works.
Of course, those works have considerably forwarded
the domain by proposing novel strategies for leaner
assessment. However, they don’t deal with learning
programming languages. Furthermore, they only use
matching methods which are based on simple
matchers.
We present, in this paper, a novel approach for
144
Boussaha K., Mokhati F. and Taleb N..
A NOVEL LEARNER SELF-ASSESSMENT APPROACH - Application to Pratical Works.
DOI: 10.5220/0003958501440148
In Proceedings of the 4th International Conference on Computer Supported Education (CSEDU-2012), pages 144-148
ISBN: 978-989-8565-07-5
Copyright
c
2012 SCITEPRESS (Science and Technology Publications, Lda.)
learner self-assessment which extends our platform
E-TéléTP@AALP introduced in a previous paper.
The main objective of our approach is to evaluating
learners’ practical works about development of C++
object-oriented programs. The proposed approach
consists of four main phases: (1) generating UML
diagrams (class and communication diagrams) form
learner’s and teacher’s programs, (2) decomposing
of the generated diagrams into subdiagrams, (3)
applying hybrid matching techniques for diagrams
(class and communication) mapping, and (4)
applying composite matching techniques for
combining the results of the third phase.
The remainder of the paper is organized as
follows. In section 2, we give a brief overview of the
platform we developed. Section 3 briefly presents
the kinds of matchers we use. Section 4 presents the
self-assessment approach we propose. Finally, we
give a conclusion and future work directions in
section 5.
2 OVERVIEW OF THE
PLATFORM E-TÉLÉTP@AALP
Realized in 2009, E-TéLéTP@AALP (figure 1) is a
platform for teaching and training of remote
practical works in programming in data processing.
Its architecture is organized into three levels (e-
learning, remote programming interfaces). The
platform is developed in ASP/MYSQL and it works
on any operating system. It offers the following
services:
For learner
It allows learner to carry out remote practical work,
to fit in the platform, to seek collaborators, to choose
a computer programming language, to download a
programming language and to consult the statements
of remote practical work.
For teacher, originator of remote practical
works
It allows teacher to register, to consult remote
practical work envisaged in the formation, to create
a new remote practical work and to contact other
teachers.
For tutors
It allows tutor to monitor learners and offers
multiple choices questionnaires in order to help
learners to remember the main concepts of
programming languages.
2.1 Architecture of E-TéLéTP@AALP
As quoted above, the proposed platform (figure 1) is
organized into three levels:
E-learning level: This level describe
interactions between human actors (learner,
tutor, teacher), interaction models in remote
practical works, all possible configurations
(various training situations between actors ) in
remote practical works on programming
languages, the teaching situations in remote
practical works, collaboration between actors,
etc.
Interfaces level: it plays the role of
intermediary between the e-learning and
remote programming levels.
Remote programming level: It presents the
activities related to remote programming
(code edition, compilation, error correction,
execution, etc).
Whether in class or within the CEHL, the
assessment is a key means of determining learner’s
forces and weaknesses during a training process.
Thus we extended the platform E-TéléTP@AALP
by integrating a new component (figure 1, in red)
into its architecture that allows learner to evaluate
his programming skills.
3 HYBRID AND COMPOSITE
MATCHERS
3.1 Hybrid Matchers
The hybrid matchers combine various matching
approaches to determine the candidates for matching
by focusing on criteria and multiple information
sources. They can provide better matching with a
better performance than executing in a separate way
several matchers (reducing the number of passes
made on the models) (Bunke, 2000).
3.2 Composite Matchers
The composite matchers combine the results
evaluated independently by several hybrid matchers.
This capacity to combine matchers is more flexible
than the “hard” combination of particular matching
techniques that are executed simultaneously or in a
fixed order (in hybrid matchers). In contrast, a
composite matcher selects from a set of modular
matchers those based, for example, on the
application domain or the formalism of the
considered models. Moreover, a composite matcher
may allow a flexible scheduling of matchers by
running them both simultaneously and sequentially.
ANOVELLEARNERSELF-ASSESSMENTAPPROACH-ApplicationtoPraticalWorks
145
In the latter case, the result of a first matcher is
consumed and extended by a second matcher to
perform iterative improvement of the result of
matching (Auxepaules, 2009).
4 THE PROPOSED SELF-
ASSESSMENT APPROACH
The proposed approach for leaner self-assessment
consists in four main phases (figure 2).
Phase 1: During this phase, learner and teacher
have to develop two C++ object-oriented
programs that undergo an analysis process for
generating UML diagrams (class and
communication diagrams). Class diagram
expresses the static structure of the program,
in terms of classes and relationships between
classes. Classes are essentially organized
through aggregation, inheritance or
association relationships. Communication
diagram describe how a set of objects
collaborate to accomplish a specific task. They
emphasize the dynamic interactions between
those objects (message exchanges) as well as
their synchronization.
Phase 2: In the second phase, the generated
diagrams must undergo a decomposition
process. It consists in breaking each diagram
into its component subdiagrams.
Phase 3: This phase consists in applying the
matching technique we adopted which offers
two kinds of hybrid matchers. The first
matcher is applied to class diagrams, while the
No
Yes
Remoteprogramming
Communication
Tools
ExecutionArea
Learner’sSpace
Tutor’sSpace
Inscription
Seekingcollaborators
Monitoringoflearner
Monitoringofgroup
Learnerassessment
Creationof
Practicalwork
Newsubjects
NewQCM
Consultingonlinehelp
ConsultingCollaboration
Appointment
LMSIHM
CompilerChoice
Sel
f
‐assessment
RemoteLoading
SharedEditor
ErrorsCorrections
Teacher’sSpace
E‐learning
Interface
Figure 1: The proposed platform’s Architecture.
CSEDU2012-4thInternationalConferenceonComputerSupportedEducation
146
Teacher’s C++
object-oriented
program
GenerationofUMLclassand
communicationdiagrams
Decompositionof
learner’sdiagrams
Learner’s C++
object-oriented
program
Learner’sUMLclassand
communicationdiagrams
GenerationofUMLclassand
communicationdiagrams
Learner’sUMLclassand
communicationdiagrams
Decompositionof
teacher’sdiagrams
Decomposedclassand
communicationdiagrams
Decomposedclassand
communicationdiagrams
Hybridmatching
Mappedcommunication
diagrams
Mappedclass
diagrams
Compositematching
Similarities and dissimilarities between
learner’s program and teacher’s program
Phase1
Phase2
Phase3
Phase4
Figure 2: Methodology of the approach.
second one is applied to communication
diagrams. In this phase a hybrid matcher is
applied between each pair of subdiagrams
(learner subdiagram, teacher subdiagram). As
output of this phase we quote the following
results:
ANOVELLEARNERSELF-ASSESSMENTAPPROACH-ApplicationtoPraticalWorks
147
- Mapped class diagrams (learner, teacher):
this result is obtained after applying the
first hybrid matcher to each pair of class
subdiagrams (learner’s class subdiagram,
teacher’s class subdiagram).
- Mapped communication diagrams (learner,
teacher): this result is obtained after
applying the second hybrid matcher to each
pair of communication subdiagrams
(learner’s communication subdiagram,
teacher’s communication subdiagram).
Phase 4: At the last phase, we apply a
composite matcher for combining the results
of hybrid matching of the third phase in order
to obtain a list of similarities and
dissimilarities between learner’s program and
teacher’s program.
It should be noted that we are currently working on
developing specific matching algorithms for our
approach and on its validation through a concrete
case study.
5 CONCLUSIONS AND FUTURE
WORK
Assessment activity plays a crucial role in improving
learners’ skills. Several approaches helping learners
during his training stage have been developed.
However, the majority of those approaches did not
consider learner assessment during its practical
works. In this paper, we proposed a novel approach
for learner self-assessment that allows comparing
two C++ object-oriented programs (learner’s
program and teacher’s program). The purpose of this
approach is to evaluate similarities and
dissimilarities between learner’s program and
teacher’s program through applying hybrid and
composite matching techniques to UML class and
communication diagrams that are generated from the
developed C++ programs. Our approach is however
limited to basic class and communication diagrams,
modelling only the most common features. As future
directions to this work, we plan on extending our
approach by introducing others UML diagrams in
order to capturing more programs’ aspects.
REFERENCES
Auxepaules, L. 2009. Analyse des diagrammes de
l’apprenant dans un EIAH de la modélisation orientée
objet. Thèse de doctorat, Université du Maine. Pages
226.
Bunke, H. 2000. Graph matching : Theoretical
foundations, algorithms, and applications. In Proc.
Vision Interface 2000, Montreal, pages 82–88.
Tanana, M., Delestre, N., Pécuchet, J.-P. et Bennouna, M.
2009. Génération d’exemples pour l’évaluation de
l’apprenant en électronique numérique à l’aide d’un
algorithme de classification. In Conférence EIAH
2009, pages 345–352.
Tchounikine, P. 2009. Précis de recherche en ingénierie
des eiah. (http://membresliglab. imag.fr/tchounikine/
Precis.html, dernière viste: juillet 2009).
Guillaume, D. 2006. Vers une scénarisation de
l’évaluation en EIAH, 13 avril 2006, université de
savoie RJC 2006.
CSEDU2012-4thInternationalConferenceonComputerSupportedEducation
148
