Information Technology Reference
In-Depth Information
Epoch number
2
(a)
1.5
1
50
100
150
200
250
300
350
400
450
2
(b)
1.5
1
50
100
150
200
250
300
350
400
450
(c)
2
1.5
1
50
100
150
200
250
300
350
400
450
Fig. 7.4 Hypnograms corresponding to patient 2. a Expert hypnogram. b SICAMM hypnogram.
c ICAMM hypnogram
parametric model of the source pdf; however there are recent contributions based
on non-parametric density estimation of the sources (see for instance [ 22 , 23 ]).
In this section, we provide an application of non-parametric ICA to detect
learning styles in e-learning. This was carried out on data of graduate and
undergraduate courses at the Universidad Politécnica Abierta (UPA) site. The
UPA is a virtual campus at the Universidad Politécnica de Valencia, which at the
time the data was collected in 2005 had more than 6,000 students registered in
about 230 courses. Figure 7.5 shows a general schema of the facilities at the virtual
campus learning environment. The e-learning event activities at the campus web
were analyzed to recognize patterns on learning styles of the students.
Data from the use of the UPA web facilities included the following information
about e-learning event activities: 1(course access), 2(agenda using), 3(news
reading), 4(content consulting), 5(email exchange), 6(chats), 7(workgroup docu-
ment), 8(exercise practice), 9(course achievement), and 10(forum participation).
The date and time for each event were also available. Besides the information on
the web activity, the exercises performed and the grades obtained by the UPA
students were also available. The data were collected from the virtual campus web
in the period from January 2002 to March 2005, totalling 2,391,003 records.
A learning-style model classifies students according to where they fit on a
number of scales corresponding to the ways in which they receive and process
information. One of the most accepted learning style taxonomies for engineering
students is Felder's model [ 24 ]; see Table 7.3 (one learning style is formed by the
combination of one feature in each dimension, for instance, intuitive-visual-
deductive-active-global). This model was used in the present work.
Search WWH ::




Custom Search