Department of Informatics University of Rijeka Radmile Mateji

  • Slides: 21
Download presentation
Department of Informatics, University of Rijeka Radmile Matejčić 2, 51000 Rijeka, Hrvatska Tel. :

Department of Informatics, University of Rijeka Radmile Matejčić 2, 51000 Rijeka, Hrvatska Tel. : + 385 51 584 700 Fax: + 385 51 584 789 http: //www. inf. uniri. hr E-learning Activities Recommender System (ELARS) Nataša Hoić-Božić, Martina Holenko Dlab Division of multimedia systems and e-learning

Outline • Introduction • ELARS recommender system – System structure • Experimental results and

Outline • Introduction • ELARS recommender system – System structure • Experimental results and Conclusions • Future plans - "E-learning Recommender System" project 2 DAAD WS 2014 26/8/2014

Introduction • Collaborative learning activities (e-tivities) • Web 2. 0 tools interactivity • Recommender

Introduction • Collaborative learning activities (e-tivities) • Web 2. 0 tools interactivity • Recommender systems personalization • Ph. D thesis: Recommender system for activities in computer-supported collaborative learning (defended in July 2014) – Supervisors: Professor Vedran Mornar and Associate Professor Nataša Hoić-Božić 3 DAAD WS 2014 26/8/2014

ELARS - E-Learning Activities Recommender System • Personalization of collaborative e-learning activities performed using

ELARS - E-Learning Activities Recommender System • Personalization of collaborative e-learning activities performed using Web 2. 0 tools − recommendations for students and groups before and during e-tivities • Activity level estimation quantity and continuity of student’s (group’s) contributions – enable recommendations generation – support teachers in evaluation of students’ work 4 DAAD WS 2014 26/8/2014

ELARS system structure ELARS 5 DAAD WS 2014 26/8/2014

ELARS system structure ELARS 5 DAAD WS 2014 26/8/2014

Activity model • Course activities: – classified (6 different types) – grouped in learning

Activity model • Course activities: – classified (6 different types) – grouped in learning modules • Items for recommendations: Mind mapping – Optional e-tivities – Web 2. 0 tools – Collaborators – Advice Your activity level is not satisfying so it is highly recommended that you participate to a greater extent. 6 DAAD WS 2014 26/8/2014

Activities workflow example 7 DAAD WS 2014 26/8/2014

Activities workflow example 7 DAAD WS 2014 26/8/2014

Student and group models • Student’s characteristics – Learning styles (VARK model) – Web

Student and group models • Student’s characteristics – Learning styles (VARK model) – Web 2. 0 tools preferences – Knowledge level – Activity level • Group’s characteristics – Activity level 8 DAAD WS 2014 26/8/2014

Subsystem for generating recommendations • Ranking items according to usefulness – support students in

Subsystem for generating recommendations • Ranking items according to usefulness – support students in decision • Filter appropriate advice from pre-defined set • Techniques adapted to the educational domain – include pedagogical rules – teachers can modify the recommendation criteria 9 DAAD WS 2014 26/8/2014

Recommending offered optional e-tivities • Usefulness: similarity student ↔ e-tivity (teachers criteria) • Technique:

Recommending offered optional e-tivities • Usefulness: similarity student ↔ e-tivity (teachers criteria) • Technique: content-based 0. 80 0. 67 0. 50 student 10 group DAAD WS 2014 26/8/2014

Recommending possible collaborators • Usefulness: similarity (teachers criteria) • Technique: content-based 0. 90 0.

Recommending possible collaborators • Usefulness: similarity (teachers criteria) • Technique: content-based 0. 90 0. 33 0. 87 0. 45 0. 86 0. 67 0. 82 0. 90 Homogeneous 11 Heterogeneous DAAD WS 2014 26/8/2014

Recommending Web 2. 0 tools offered for certain e-tivity • Usefulness: student’s preference of

Recommending Web 2. 0 tools offered for certain e-tivity • Usefulness: student’s preference of the target tool (prediction of missing values) • Technique: hybrid (collaborative filtering + content-based) 0. 90 0. 32 -0. 5 student 12 DAAD WS 2014 group 26/8/2014

Providing advice • Technique: knowledge-based – Useful/not useful 13 DAAD WS 2014 26/8/2014

Providing advice • Technique: knowledge-based – Useful/not useful 13 DAAD WS 2014 26/8/2014

ELARS web application • http: //161. 53. 18. 114/elars 14 DAAD WS 2014 26/8/2014

ELARS web application • http: //161. 53. 18. 114/elars 14 DAAD WS 2014 26/8/2014

ELARS demo • http: //161. 53. 18. 114/elarsdemo 15 DAAD WS 2014 26/8/2014

ELARS demo • http: //161. 53. 18. 114/elarsdemo 15 DAAD WS 2014 26/8/2014

Learning modules and activities 16 DAAD WS 2014 26/8/2014

Learning modules and activities 16 DAAD WS 2014 26/8/2014

Decision activities 17 DAAD WS 2014 26/8/2014

Decision activities 17 DAAD WS 2014 26/8/2014

Activity levels and advice 18 DAAD WS 2014 26/8/2014

Activity levels and advice 18 DAAD WS 2014 26/8/2014

Experimental results and Conclusion • Evaluation focused on pedagogical aspects – two e-courses (University

Experimental results and Conclusion • Evaluation focused on pedagogical aspects – two e-courses (University of Rijeka, Croatia) – control and experimental group: • better results (points) for e-tivities – survey: • students are satisfied with received recommendations and find the system useful • Good results on experimental courses and students’ satisfaction provide motivation for further work and improvements. 19 DAAD WS 2014 26/8/2014

Future plans - "E-learning Recommender System" project • "E-learning Recommender System" project supported by

Future plans - "E-learning Recommender System" project • "E-learning Recommender System" project supported by University of Rijeka • Future plans – recommendations algorithms improvements – improvements of ELARS authoring component for teachers – creating and evaluating didactical models for the use of Web 2. 0 based e-tivities in different types of blended and online higher education e-courses 20 DAAD WS 2014 26/8/2014

Thank you!

Thank you!