Other Applications: Sequential Dependence Modelling and Data Mining - On Statistical Pattern Recognition in Independent Component Analysis Mixture Modelling - page 152

Chapter 7
Other Applications: Sequential
Dependence Modelling and Data Mining
This chapter presents two diverse applications: diagnosis of sleep disorders
(apnea) and data mining in a web of a virtual campus. The first application presents
a procedure to extend ICA mixture models (ICAMM) to the case of having
sequential dependence in the feature observation record. We call it sequential
ICAMM (SICAMM). We present the algorithm, which is essentially a sequential
Bayes processor, which can be used to sequentially classify the input feature
vector among a given set of possible classes. Estimates of the class-transition
probabilities are used in conjunction with the classical ICAMM parameters:
mixture matrices, centroids, and source probability densities. These parameters
were estimated using the Mixca algorithm proposed in Chap. 3 . Some simulations
are presented to verify the improvement of SICAMM with respect to ICAMM.
Moreover, a real data case is considered: the computation of hypnograms to help in
the diagnosis of sleep disorders. Both simulated and real data analyses demonstrate
the potential interest of including sequential dependence in the implementation of
an ICAMM classifier.
In the second application, ICA is used as a data mining technique to discover
patterns in e-learning. An ICA model is proposed that defines the sources as
dimensions of the learning styles of the students. A novel non-parametric ICA and
standard ICA algorithms are applied to huge historical web log data from a virtual
campus in order to detect the relationship between web activities and learning
styles. The data are divided by the course types into graduate courses and regular
academic courses. Each of these divisions is separated into two subsets: cases with
grades and cases with no grades. Web activities include events such as course
access, email exchange, forum participation, news reading, chats, and achieve-
ments. Suitable learning styles of the students were positively detected for grad-
uate courses with grades using the non-parametric Mixca algorithm.