ering for web usage mining with the proposed game usage mining approaches,
and finally a proposal for logging of player behavior in massively multiplayer on-
line games. The presented logging approach is partially implemented in paper C.
This paper tries to answer research questions RQ2.2-RQ2.4.
Support Vector Classifiers
fication using memoization to speed up classification with multiple classes. The
algorithm is empirically compared to an approach without the use of memoization
and empirically shown to be faster. The purpose of this classifier is to provide
scalable and memory efficient support prediction of action or clicks on the (mo-
bile) web (paper A and B) and massively multiplayer online games (paper C and
E). This paper tries to answer RQ3.2.
Support Vector Classification using Decay Coeffi-
cients
memory efficient decremental training of the incremental proximal SVM classifier.
This paper tries to answer research question RQ3.3
imal Support Vector Machine Classifier using a
Heap-based Tree Topology
is to handle even larger amounts of increasingly growing classification data (in
cyberspace services). It tries to go further than paper F in answering research
question RQ3.2