Back to search

University of Piraeus (UNIPI)

User modelling and adaptivity in computer supported collaborative learning environments for object oriented software development methodologies

Abstract

dc:description

The aim of this thesis is to introduce a new approach to intelligent and affective recommendations offered by a Computer Supported Collaborative Learning (CSCL) system, called AUTO-COLLEAGUE (AUTOmated COLLaborativE leArninG Uml Environment). It is an environment for training users in UML, but it can also be used for other popular domains, such as programming languages. It may be used both in educational institutes and software houses. AUTO-COLLEAGUE traces every action and performance indication in order to provide useful recommendations to the trainees, as well as the trainer who is often neglected in similar environments. The contributions of the system presented in this thesis concern CSCL systems and group formation tools. In specific, the novel approaches implemented in AUTOCOLLEAGUE are: • The personality-related characteristics included in the student models and, especially, the way they are automatically traced and evaluated, • The effectivity implemented to recommend optimum groups of trainees and • The use of a leadership theory, and specifically the Hersey and Blanchard Situational Leadership Theory, for adapting intelligent recommendations in a learning environment. The recommendations to the trainees concern (a) the next UML topics they should study, (b) the appropriate colleagues with whom they should collaborate and (c) supportive/encouraging messages that would increase their performance. The recommendations to the trainer include (a) historical/statistical reports and charts of showing the progress of the trainees and (b) a group formation tool that proposes optimum organization of the trainees into groups. The recommender system is built using both the content-based and the collaborative filtering methods. The intelligence of the recommendations is based on the individual student models of the trainees. The system uses a hybrid student modelling technique, combining perturbation (buggy) and stereotype-based student models to describe the trainees. The student model contains both domain dependent and domain independent data. The domain dependent data represent the level of expertise referring to the domain knowledge description. It is structured in topics, each of which bears a degree of knowledge. The domain independent data refer to (a) the overall emotional state (positive/negative) of the trainee and (b) a set of personality characteristics that affect the trainee’s learning and collaboration processes. The overall emotional state of the trainee is inferred using the OCC theory. OCC is a cognitive theory of emotions that suggests a computational model for emotion recognition and modelling. It is the most popular theory of emotions adapted in affective systems. The emotional state of trainee is used as a criterion for the group formation recommendations and the supportive messages offered to the trainee. The personality characteristics are structured in eight stereotypes: self-confident, diligent, participative, willing to help, sceptical, hurried, unconcentrated and efficient. In order for the system to update the student model, it monitors specific attributes that are relevant to the included stereotypes and concern the trainee’s use of the environment and performance on the domain knowledge. The way the system manipulates the values of these attributes to make inferences is instructed by the results of an empirical study conducted amongst experienced trainers. The contributions of this thesis are associated mainly to the kinds of personality characteristics evaluated and the criteria of the group formation algorithm. Most of the CSCL systems are based on the abilities and knowledge of the trainees to provide intelligent and adaptive advice. Others use learning styles and team roles, but the vast majority does not automatically traces their values observing the student. Instead, they use scientific instruments, such as questionnaires. None of these systems use personality characteristics similar to ours. In addition, considering experimental studies on group formation tools, limited work is done despite the recognized importance of effective group formation in collaborative learning tasks. Indeed, there are interesting approaches of group formation, but most of them attempt to create heterogeneous or homogeneous groups considering the knowledge, learning styles or team roles. On the contrary, the group formation algorithm proposed in this thesis is based on desired and undesired combinations of personality attributes and the emotional influence between the trainees during the collaborative learning activities. Furthermore, the group formation tool is implemented using the Simulated Annealing algorithm, which has never been used in similar processes and environments. The Simulated Annealing algorithm is a genetic algorithm that serves as a general optimization technique for solving combinatorial optimization problems. It uses a control parameter (called temperature) to avoid getting trapped in poor local optima. Another contribution of our study lies in the use of the Hersey and Blanchard Situational Leadership Theory for adapting the environment and recommendations to the trainee and for proposing to the trainee the most appropriate leadership style to follow per trainee and per learning task. We were inspired by educational studies that emphasized on the importance of the teacher to take a leadership role in classroom. The Hersey and Blanchard Leadership Theory is based on the principle that successful leaders should be flexible enough to adapt their leadership style according to the maturity of the followers for each assigned task. No other learning system has yet used a leadership theory. The effectiveness of the provided recommendations has been demonstrated by the positive results of two evaluation experiments conducted in real time with real trainees. An extended part of this thesis has been dedicated to describing the importance of intelligent Computer Supported Collaborative Learning systems and analyzing the origins of collaborative learning in the psychology and educational literature.

Degree

thesis:*
Grantor dc:publisher
University of Piraeus (UNIPI)
Year dc:date
2011

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Tourtoglou, Kalliopi
  • Τουρτόγλου, Καλλιόπη

Subjects

dc:subject × 20

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/28797

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Tourtoglou, Kalliopi; Τουρτόγλου, Καλλιόπη. User modelling and adaptivity in computer supported collaborative learning environments for object oriented software development methodologies. University of Piraeus (UNIPI), 2011. http://hdl.handle.net/10442/hedi/28797