Lecture Notes in Computer Science- P10 pdf

e-Learning Issues under an Affective Perspective 35

The role of the Pedagogical Generator, however, is not restricted only to the reas-
surance of the appropriateness of the teaching method or the educational material. It is
concentrated also on providing the student with encouraging actions in order to pre-
serve his positive emotional state. The pedagogical actions which have been imple-
mented in the current version of our system are shown in Table 1.
The main concern of the Mentor Component, as it is already mentioned, is to en-
sure that the student’s mood is positive every time. This condition is very crucial in
order to involve the student efficiently in the process of learning [4], [9]. To achieve
this, the Mentor Component has to be aware of the student’s emotions. The input that
comes from the Emotional Component, which is in charge of the detection of the
student’s motivational state, is evaluated appropriately and thereafter the Mentor
Component adapts his reaction adequately to motivate the student either by encourag-
ing him or by praising him and in every case sustain his disposition flourishing. Once
the Mentor Component is aware of the student’s emotions, it can proceed into the
selection of the proper affective tactic.
Table 1. The pedagogical actions of the Pedagogical Generator
Ask for giving some help Explain the need for help
Give Help to student Reassure the appropriateness of help
Express satisfaction after a successful help Express unhappiness after an unsuccessful
help and ask for trying again
Give explanations in an appropriate way Express sympathy in case of fail
Encourage the student Congratulate the student
Praise the student Express admiration for the student
Reinforce student’s efforts Play a game with student
Give hope Open a dialogue with the student
Play a music video clip Present a part of a movie
Present a photo Tell a joke
Let us examine, for example, the case of a student whose personality belongs to the
Extraversion category, but his mood is recognized in the current session as negative.

For this type of student the Teaching Generator has already selected an exploratory
teaching method without examining his emotional state. Before the Mentor Compo-
nent applies this method, it interacts with the Pedagogical Generator. By analysing
furthermore why his mood is negative, it comes to light that the student is anxious for
some reason. The system takes upon making the student feel relaxed firstly by open-
ing a short dialogue with him. Thereafter, it presents to him either a joke or a funny
video clip, according to his preferences which are stored in the student model. Finally,
it motivates him either by encouraging him or by praising his abilities.
Another case is when a Conscientiousness student fails to accomplish a given task.
Then negative emotions such as sadness or disappointment can appear. He seems to
be less confident in the current session and there is the danger of giving up the trial.
He fears maybe that he has not got the ability to deal with a project that was assigned
to him and he will not live up to his teacher’s expectations. According to Table 1,
there are pedagogical actions which can be applied in order to eliminate the student’s
negative emotions. For instance, the system may praise him for his effort, give him
36 M. Leontidis, C. Halatsis, and M. Grigoriadou

help and encourage him to try again. Then the Mentor Component presents him an
easier problem to reinforce his confidence and to foster positive emotions. In this
way, the student has great chances to resolve the problem, so that his confidence
would be regained and positive emotions such as happiness or satisfaction can pre-
serve an upbeat to the student’s mood.

Fig. 2. The production rules of the Affective Tactics
Similar analyses have been made for the rest of the cases that have been imple-
mented in our system. At this moment in time, there are 20 affective tactics imple-
mented in our system in order to deal with 10 different cases respectively. In Figure 2
are exemplified the production rules of some of these cases.
We mentioned above some cases with the aim of showing how the Mentor Com-
ponent selects and suggests which affective tactic will be used. Our system is sched-

uled to deal with domain-independent educational environments. It would be used
therefore for teaching any domain of subject and this is the major point that our work
is diversified from the others.
4 Conclusions and Further Work
During the last years, the significance of the affective factors in human – computer
interaction has been established and great scientific efforts have been attempted towards
this direction. As emotions have long been a major concern, more and more computer
scientists have recently paid close attention to these factors in order to build their sys-
tems. The significant role of personality and the influence of emotions on memory,
thinking, reasoning and creativity, which are basic constitutes in the learning process,
have been taken into account in the integration of modern educational environments.
In this paper, we presented the MENTOR Affective Module which is responsible for
inferring students’ emotions and providing them with the appropriate affective tactic in
distance learning. The MENTOR is integrated in a Web Adaptive Educational System
e-Learning Issues under an Affective Perspective 37

with the aim of providing personalized learning. The implementation of the MENTOR
has been achieved by using the PHP5 language for server-side scripting and the MySQL
for the data-base management, supported by Apache HTTP server 2.2. The recognition
of emotions is based on a formal representation of emotions using an appropriately
designed ontology which is implemented with the Protégé tool and is achieved by a
decision tree method. A DL-OWL inference engine has been used to make predictions
about the emotional state of the student.
The main purpose of the MENTOR, except from the recognition of emotions, is to
create and / or preserve a positive mood in the student, since this is a crucial factor for
the learning process. Moreover, it aims at providing the system with suitable informa-
tion about the personality and emotions of the student and also with appropriate peda-
gogical actions enhancing the student's motivation to “conquer” the intended knowl-
edge. At this time, we have implemented 20 affective tactics. The designation of these
tactics has taken into account the professional opinion of teachers and psychologists.

Furthermore, we are developing this component bearing in mind to be independent
from the specific domain model of educational systems, so that it has the capability to
be used by a wide range of them. In advance research, we intend to improve the accu-
racy of our system so that we are capable of recognizing more emotions and more
complicated emotional situations. We hope that in future versions the number of af-
fective tactics will be further evolved so as to include more cases. When the integra-
tion of the MENTOR will have been completed, we will be able to testify its reliabil-
ity conducting a web evaluation.
Acknowledgments. The authors would like to thank Hara Pantazopoulou for her
valuable support to the completion of this paper.
References
1. Ames, C.: Classroom goals, structures and student motivation. Journal of Educational Psy-
chology 84(3), 261–271 (1992)
2. Andre, E., Klesen, M., Gebhard, P., Allen, S., Rist, T.: Integrating models of personality
and emotions into lifelike characters. In: Proceedings International Workshop on Affect in
Interactions - Towards a New Generation of Interfaces, pp. 136–149 (1999)
3. Brusilovsky, P.: Adaptive Hypermedia, User Modeling and User-Adapted Interaction,
11(1/2), 111–127 (2001)
4. Coles, G.: Reading Lessons: The Debate over Literacy. Hill & Wang, New York (1998)
5. Conati, C., Zhou., X.: Modeling students’ emotions from Cognitive Appraisal in Educa-
tional Games. In: 6th International Conference on ITS, Biarritz, France (2002)
6. Costa, P.T., McCrae, R.R.: Four ways five factors are basic. Personality and Individual
Differences 1 13, 653–665 (1992)
7. Elliot, C., Rickel, J., Lester, J.: Lifelike Pedagogical Agents and Affective Computing: An
Exploratory Synthesis. In: Veloso, M.M., Wooldridge, M.J. (eds.) Artificial Intelligence
Today. LNCS (LNAI), vol. 1600, pp. 195–212. Springer, Heidelberg (1999)
8. Goldberg, L.R : International Personality Item Pool: A Scientific Collaboratory for the
Development of Advanced Measures of Personality and Other Individual differences
(1999),
38 M. Leontidis, C. Halatsis, and M. Grigoriadou

9. Izard, C.E.: Emotion Cognition relationships and human development. Emotions, cogni-
tion and behaviour, pp. 59–67. Oxford University Press, Oxford (1984)
10. Keller, J.M.: Using the ARCS Motivational Pro-cess in Computer-Based Instruction and
Distance Education. In: Theall, M. (ed.) Motivation in Teaching and Learning: New Direc-
tions for Teaching and Learning, Jossey-Bass, San Francisco (1999)
11. Leontidis, M., Halatsis, C.: An affective way to enrich learning. In: Proceedings of the
IADIS International Conference on e-Learning, Lisbon, Portugal, pp. 32–36 (2007)
12. Lepper, M.R., Woolverton, M., Mumme, D., Gurtner, J.: Motivational techniques of expert
human tutors: Lessons for the design of computer-based tutors. In: Lajoie, S., Derry, S.
(eds.) Computers as Cognitive Tools, pp. 75–105. Lawrence Erlbaum, Mahwah (1993)
13. Martinho, C., Machado, I., Paiva, A.: A Cognitive Approach to Affective User Modeling.
In: Paiva, A. (ed.) IWAI 1999. LNCS, vol. 1814, pp. 64–75. Springer, Heidelberg (2000)
14. McCrae, R.R., John, O.P.: An introduction to the five factor model and its applications.
Special Issue: The five factor model: Issues and applications. Journal of Personality 60,
175–215 (1992)
15. Oren, T.I., Ghasem-Aghaee, N.: Personality Representation Processable in Fuzzy Logic
for Human Behavior Simulation. In: Proceedings of the 2003 Summer Computer Simula-
tion Conference, Montreal, PQ, Canada, July 20-24, pp. 11–18 (2003)
16. Ortony, A., Clore, G.L., Collins, A.: The Cognitive Structure of Emotions. Cambridge
University Press, Cambridge (1988)
17. Jiang, P., Xiang, H., Ren, F., Kuroiwa, S.: An Advanced Mental State Transition Network
& Psychological Experiments. In: Yang, L.T., Amamiya, M., Liu, Z., Guo, M., Rammig,
F.J. (eds.) EUC 2005. LNCS, vol. 3824, pp. 1026–1035. Springer, Heidelberg (2005)
18. Picard, R.W.: Affective Computing. MIT Press, Cambridge (1997)
19. Scherer, K.: Psychological models of emotion. In: Borod, J. (ed.) The neuropsychology of
emotion, pp. 137–162. Oxford University Press, New York (2000)
20. Wasson, B.: Determining the Focus of Instruction: Content Planning for ITS. PhD thesis,
Department of Computational Science, University of Saskatchewan, Canada (1990)
21. Aroyo, L., Dicheva, D., Cristea, A.: Ontological Support for Web Courseware Authoring,

Int. Conf. on Intelligent Tutoring Systems (ITS 2002), France, pp. 270-280 (2002)
22. Quinlan, J.R.: Improved use of continuous attributes in c4.5. Journal of Artificial Intelli-
gence Research (4), 77–90 (1996)
23. Bourdeau, J., Mizoguchi, R.: Ontological Engineering of Instruction: A Perspective. In:
Proceedings of AIED 1999, pp. 620–623 (1999)

F. Li et al. (Eds.): ICWL 2008, LNCS 5145, pp. 39–48, 2008.
© Springer-Verlag Berlin Heidelberg 2008
Recommendation in Education Portal by Relation Based
Importance Ranking
Xin Wang, Fang Yuan, and Li Qi

Computer and Information Management Center, Tsinghua University
Beijing, China
{wxin,yf,qili}@cic.tsinghua.edu.cn
Abstract. Recommendation in education portal is helpful for students to know
the important learning resources in schools. Currently, previous methods which
have been proposed to solve this problem mainly focus on page view counts. A
learning resource is important just because many students have viewed it. How-
ever, as the metadata in a resource is becoming available, the relations among
the resources and other entities in real world are becoming more and more. Un-
fortunately, how to use such relations to make better recommendations has not
been well studied. In this paper, we present a complementary study to this prob-
lem. Specially, we focus on a general education portal, which consists of differ-
ent typed objects, including resource, category, tag, user and department. The
recommendation object is resource. However, we have found that a resource’s
importance rank can be affected by its relations to other typed objects. Thus, we
formalize the resource recommendation as a ranking problem by considering its
relations to other typed objects. A random walk algorithm to estimate the im-
portance of each object in the education portal is proposed. Finally, the experi-

mental result is evaluated in a real world data set.
Keywords: Recommendation, Importance Ranking, Random Walk.
1 Introduction
Recommendation has been proven to be a useful approach for reducing users’ efforts
to find information which may be interesting. It has been applied to many popular
commercial web based applications, e.g., www.amazon.com, www.facebook.com,
imdb.com and so on. In education portals, it’s also important to recommend resources
to users, especially, recommend learning resources to students. It can be very helpful
to benefit the students to get more interesting topics and affect the efficiency of learn-
ing. For example, the MIT’s open course ware is helpful for students in the whole
world.
Recently, many methods have been proposed for the recommendation in academic
and industry, for example, content based filtering [1], clustering model [2], associa-
tion rule based approach [3], and graph model [4]. The proposed methods are mainly
applied to the open web applications, which mean that the application’s end users are
normal users in the internet. Different from the open web, the education portal is
aimed at providing services to students and the resources in education portal have