The project TASK_Moniminer of the team of Prof. Dr. Christoph Meinel concerns on discovering the students learning interest from usage data in web-based learning environment, which is the further and auxiliary project for tele-TASK.
- Monitoring and mining the learning interest on multimedia lectures
TASK_Moniminer (version1.0) concerns on discovering the students learning interest from usage data in web-based learning environment, which is the further and auxiliary project for tele-TASK. The usage data were taken from kinds of server logs, which record the actions of visitors on real media lectures, podcast clips and text-based html pages. The browsing profile for one visitor in one visit (or one learning session) is built on usage data and represented as the sequence of learning objects, and each of learning objects is related with a bigram (Staying Time, Number of Clicks). The set of browsing profiles is the unified data source for discovering learning interest. The learning interest on multimedia lectures is expressed in many questions, such as "is there any preference on different lectures?", "did the students view other lectures when they access one lecture?" and etc. Different questions are induced into different mining models, which require different mining methods
- Query the learning interest in a semantic way
TASK_Moniminer (version 2.0) enriches version 1.0. The old version lacks the intuitive and expressive mechanism for querying the usage information on e-lectures. For example, when a teacher wants to query the usage on the lectures about "Weaknesses Windows", it is impossible in version 1.0, because in most cases the file name extracted from server logs does NOT show any semantic concepts of the content for one lecture. In version 2.0, such inconvenience is solved by using the concept information such as "title", "description" and "tags" stored in database for e-lectures. With this help, when a teacher queries the usage on ""Weaknesses Windows", he/she could get the right usage information not only on the course "Operation System", but "Internet Security". The query results are ranked in a list based on the similarity comparison in information theory.
- Elite multimedia lectures
There are already over 1000 recorded lectures in tele-TASK. Though the lectures can be archived in a bibliographic taxonomy, the visitors need to be supplied the "elite lectures", which show the most representative and popular e-lectures in our lecture base. This service is more than "top rated" in other tube web sites, because it compromises the "rating" on user site and the characteristics of lecture content. The "elite lectures" is dynamic with the increasing of numbers of visiting and lectures.
- Duplicate Lectures Detection and Deletion
- Long Wang, Christoph Meinel
Detecting the Changes of Web Students' Learning Interest
Proc. 6th IEEE/WIC/ACM WI, IEEE Press, pp. 816 - 819, Silicon Valley, USA, November 2007
- Long Wang, Christoph Meinel
"Mining the Students' Learning Interest in Browsing Web-Streaming Lectures"
In Proc. of IEEE CIDM'07, pp.194 - 201, Honolulu, Hawaii
Last updated: 18-04-2008