Abstract

Blended learning is usually known as a didactically valuable curriculum covering phases of group presence and computer supported individual work in virtual learning environments (VLE). Virtual classrooms fill one important gap in between as the most powerful method of computer supported collaborative learning (CSCL), which is applied increasingly in enterprise further education lectures. However, Web 2.0 trends are often completely ignored in this case. Trainees, especially if working in parallel, are looking for much more flexible granularity offered as learning offers in accordance to their daily life habits including: short term learning pieces, choice from multi-medial options, mobile device support, multi-channel intercommunication options with their mates and tutors. Life-long learning is not longer dedicated to certain time slots and methods. It becomes a kind of game often played in parallel to other activities of both, business and private life. On the one hand, there are more knowledge sources than can be discovered by Google, YouTube and Wikipedia. Thousands of multi-medial contents are appropriate to enrich web-based self-controlled learning on fixed (e-learning) and mobile (m-learning) devices. Short breaks can be filled smoothly by consumption of undiscovered video clips, fitting to a personal curriculum, covering multiple subjects. On the other hand, always experienced people, willing to help, are online. Search for documents moves to search for appropriate experts and their recommendations. The full paper will address a new learning approach based on personalization and social context. Recommendation approaches are compared and an expert finding system called Spree, based on personal interest and expertise profiles will be introduced. Spree does not stop with expert seeking, but seamless integrates communication and community components for instant and off-line multi-user chat, presence, blogging, rating and scoring. Simple questions, written in native language are used as triggers to establish new contacts. Technically, language parsing including stemming, n-gram-analysis and metric’s definition and application leads to initial classification according to a domain dependent, learning field specific taxonomy. A learner’s profile is represented as an instance of this taxonomy. Multi-classification is applied. Results known from our open source system askspree.de and fast track trials led the re-implementation of a more powerful solution for Deutsche Telekom regarding the matching core and usability (GUI) to enhance social networking.

@inproceedings{...,
author = {Eichler, Gerald and Lommatzsch, Andreas},
title = {Profile-based Recommendation of and by Domain Experts supporting web 2.0 Community Paradigms},
booktitle = {Proccedings of the International Conference on Education and New Learning Technologies},
year = {2009},
isbn = {978-84-612-9802-0},
location = {Barcelona},
publisher = {...},
address = {...},
}