1. Mining the Students’ Learning Interest in Browsing Web-Streaming Lectures Computational Intelligence and Data Mining (CIDM 2007)

2. outline Introduction Motivation Model Student Browsing Profile ANALYSIS OF MINING RESULTS ON TELE-TASK CONCLUSIONS

3. Introduction Web-Streaming lectures overcome the space and time barriers between learning and teaching, but bring higher requirements on the learning feedback of students when they browse lectures. The learning interests are expressed in six questions

4. Motivation Such web-based e-learning systems facilitate teachers and learners greatly. In classroom vs. Web

5. Model Student Browsing Profile

6. Model Student Browsing Profile(cont.)

7. 1) Are the online lectures welcomed by students? method:NA S,l and NP S,l compared to that on NC S,l shows if the lectures are welcomed by online learners or not. NA S,l and NP S,l can be directly computed from usage logs, and NC S,l can be gotten by asking the teachers or their teaching assistants.

8. Model Student Browsing Profile(cont.) 2) Is there any difference between viewing the live broadcasting lectures and browsing lectures after they are recorded and edited? method:NA S,l and NP S,l Lecture’s recording it is usually between 60 minutes and 90 minutes NA S,l always less than NP S,l

9. Model Student Browsing Profile(cont.) 3) Is there any preference on the different lectures in a course and preference on different pieces of one lecture method:Observe 觀察 ND s,l & NO s,l 是否大於一般值

10. Model Student Browsing Profile(cont.) Question 4: Did the students view other lectures when they accessed one lecture? method: the set including all the learning sessions are named as P From P, we try to mine the relations each of which is formed as Suppr is the number of sessions that viewed all the lectures in r (by frequent)

11. Model Student Browsing Profile(cont.) Question 5: Is there any relation between the exercise marks and the usage on lectures? method: where α + β + γ + δ + θ = 1 and the values of these five coefficients are assigned based on the statistical observations or the expert experiences. (US s,l :: usage score)

12. Model Student Browsing Profile(cont.) Question 6: For the same named courses supplied for different years, is there any changes on the students’ interest? method:

13. Model Student Browsing Profile(cont.) 1) From Course Level: where l 1i ∈ C1 and l 2j ∈ C2 2) From Lecture Level: K1 and K2 are the sets of the knowledge elements of l1 and l

14. Model Student Browsing Profile(cont.) 3) From Chapter Level:

15. ANALYSIS OF MINING RESULTS ON TELE-TASK

16. ANALYSIS OF MINING RESULTS ON TELE-TASK(cont.)

17. CONCLUSIONS The average time of a student spending on a lecture is about 10 minutes, while the normal length of a lecture is about 90 minutes