A novel social network incorporating intelligent techniques for personalized collaborative learning in adult education
Keywords
Adaptive tutoring ; Assessment strategies ; Badges system ; Digital learning ; Formative feedback ; Fuzzy logic ; Genetic algorithms ; Group formation ; Intelligent techniques ; Learning analytics ; Pedagogical methods ; Personalized learning ; Bloom taxonomy ; Social-extended student modeling ; Social networksAbstract
The proliferation of social networking sites and the general interest in student-centered pedagogies have attracted attention to the use of this popular technological advance to enhance distance education. Such potential can be effectively used in adult education, since it allows adult educators to prolong learning process beyond a classroom, and adult learners to learn without time and space restrictions. To this direction, this Ph.D. thesis presents a novel approach of an intelligent social network designed for adult learning that performs individualized instruction, adaptive feedback and project-based learning with the intention to provide an innovative learning environment employing state-of-art technology, artificial techniques and pedagogical methods. The developed system is called IAdA-SNL, an abbreviation of Intelligent Adaptive Social Network for Adult Learning.
From a technological perspective, IAdA-SNL is an integrated social network based on other well-known social networking sites standards; providing the capabilities of posting, commenting, annotating, reacting on posts and course objects, tagging keywords and peers, chatting, setting up profile, exploring peers’ profile and network feeds, and making friends. From pedagogical perspective, since IAdA-SNL is oriented to adult education, it employs a set of learning theories for effective instructional design, namely Connectivism, Revised and Digital Bloom Taxonomy, Behaviorism (including Social learning theory, Reinforcement theory of motivation, badges system), Malcolm Knowles’ Andragogy and Experiential learning.
IAdA-SNL extends traditional approaches of student modeling with characteristics pertained to learner’s social interactions with the network. The proposed dynamic student model is composed by profile information (name, age, gender, education, location), learning style (Honey and Mumford model), performance data (achievements in each cognitive skill based on RBT, total grades in chapters’ tests and performance progress), interactions with course (how many times s/he has visited a topic or learning resource, the reactions on course objects about his/her level of understanding, how many times the tests are given and his/her learning pace), social interactions (the number of posts, comments, chats, reactions etc.), and accomplishments (seven badge categories with three level badges). The social-extended student model is the core component based on which the other system components provide personalization and adaptivity.
To the direction of individualized instruction, IAdA-SNL adapts course content and learning object recommendation considering student learning style. The Honey and Mumford model used classifies students as activists, reflectors, theorists, and pragmatists. The reason why this model used is because it can promote better the adult learning since it is based on the way students approach new learning experiences. Moreover, a social annotation tool is provided to make notes at course material making it more interactive and engaging.
IAdA-SNL incorporates an integrated badges system and a component for the delivery of corresponding individual motivational messages, with the intention of encouraging learning, improving student engagement, acknowledging current accomplishments and incentivizing future ones. The badging model consists of seven badge categories including three levels of badges each one: Profile (Basic, Charm, Self-Introduced), Recognition (Get known, Well-known, VIP), Contributions (Writer, Blogger, Editor), User-Generated Content (Pupil, Assistant, Professor), Social Involvement (Attendee, Active, Outgoing), Cognitive level (Entry-level, Knowledgeable, Genius), and Global (Newbie, Competent, Well-rounded).
Moreover, IAdA-SNL provides actionable learning analytics in order to timely inform the student about critical issues concerning the learning process and encourage him/her to take the appropriate action for improving his/her learning outcomes. The proposed actionable learning analytics automatically deliver reports about student progress in curriculum, performance in curriculum in relation to the top 10 students and to the average grades of the learners’ community.
Regarding the assessment strategy, IAdA-SNL develops two innovative strategies. Firstly, an intelligent formative assessment (IFA) supporting feedback on student’s misconceptions is used as practice test for preparing the student for the chapter’s final test. The proposed IFA employs a fuzzy logic model that diagnoses the cause of misconceptions occurred at each chapter section considering the answer’s misconception degree, student’s misconception degree, answer time and question difficulty. The misconception diagnosis exports if student was imprudent, possibly imprudent, possibly unread or unread. After the completion of practice test, the IFA provides a cumulative feedback on the overall result of student evaluation, and suggestions for each section where a misconception occurred. Secondly, IAdA-SNL applies RBT to the development of chapter’s final tests in order to be composed of questions derived from the different RBT levels. RBT can be helpful for specifying correctly course objectives, preparing valid and efficient assessments by including activities related to different taxonomy’s levels and thus, improving students’ cognitive level. These tests not only provide the evaluative feedback on student’s performance, but also descriptive feedback on the performance at each cognitive level.
Concerning the collaborative learning, IAdA-SNL assigns a project-based activity to students. In order to form effective collaborative student groups, IAdA-SNL introduces a novel genetic algorithm. Its innovations pertain to the attributes used for the composition of groups and genetic operators applied. In particular, student attributes refer to the three main dimensions of learning in a SN-Learning environment, namely academic, cognitive, and social, and derive from the social-extended student model. Regarding genetic operators, the algorithm performs two crossover operators: a modification of 2-point crossover and a new approach, called 1-point per group crossover.
Finally, IAdA-SNL platform was used by higher education students and adult learners to learn system analysis and design using UML diagrams. For the evaluation of IAdA-SNL modules, an evaluation model is introduced based on the CIAO! Framework and the adjusted ISO-based model to SN-Learning environments. The results of the evaluation are very encouraging. They demonstrate that the system effectively adapts the learning process to the students’ needs, motivates them and suggests personalized learning strategies, and promotes collaboration and communication. Moreover, statistical hypothesis tests (t-tests) were employed for the evaluation of assessment and grouping modules. Evaluating the proposed assessment strategies, the results show that students benefit against the traditional assessment process, pointing the pedagogical affordance of constructive feedback provided in both strategies. Evaluating the proposed genetic algorithm performance, the results show its effectiveness and superiority over simple genetic algorithm approach. Moreover, from pedagogical perspective, the positive students’ attitude and high acceptance towards our group formation method is indicated.